MODULE OUTLINE
6.1 MEASURES OF ASSOCIATION and EFFECT
6.1.1 General Concepts
6.1.2 Tests of Association
6.1.5 Meta Analysis
6.2.5 Survey Error and Sampling Bias
6.3 HEALTH STATUS
6.3.3 Disease Registries with Cancer as an Example
6.4 HEALTH SERVICES
6.4.1 Health Economics
6.5 READING AND WRITING SCIENTIFIC LITERATURE
6.5.1 Literature Search
6.5.2 Critical Reading of a Journal Article
6.5.3 Abuse or Misuse of Statistics
6.5.4 Scientific Writing
6.5.5 Scientific Publishing
UNIT 6.1
MEASURES OF ASSOCIATION and EFFECT
Learning Objectives:
· Tests/measures of association for continuous data: t test, F test, regression
· Tests/measures of association for discrete data: chi-square
· Measures of effect: rate ratio, risk ratio, odds ratio, attributable rate
· Validity and precision of effect measures
· Meta analysis
· Association, measures/tests of association
· Chi-square, Mantel-Haenszel chi-square
· Chi-square, Pearson chi-square
· Effect modification
· Effect, measures of effect
· Interaction
· Measures of trend
· Meta-analysis
· Odds ratio
· Odds ratio, Mantel-Haenszel odds ratio
· Precision
· Meta analysis
· Rate difference
· Risk ratio
· Validity
UNIT OUTLINE
6.1.1 GENERAL CONCEPTS
A. Analytic Epidemiology
B. Hypothesis Testing
C. Preliminaries to Data Analysis
D. Procedures Used:
6.1.2 TESTS OF ASSOCIATION
A. Tests of Association on Means:
B. Tests of Association on Proportions: Single 2x2 Contingency Table
C. Tests of Association on Proportions: Single: 2 X K Contingency Table:
D. Tests of Association on Proportions: Stratified 2 X 2 Contingency Tables
A. Validity
B. Internal Validity
C. External Validity
D. Precision
6.1.5 META ANALYSIS
A. Definition and Historical Background
B. Advantages of Meta Analysis
C. Steps in Meta Analysis
D. Difficulties of Meta Analysis
A. ANALYTIC EPIDEMIOLOGY
Analytic epidemiology is very important in public health because of its major role in the planning and evaluation of public health interventions. Intervention is the backbone of public health. Data analysis must be taken very seriously because of its involvement in practical field decisions and programs that affect the individual, the community, and the eco-system. Wrong analysis will lead to wrong conclusions that will have deleterious effects.
B. HYPOTHESIS TESTING
Tests for association, effect, or trend involve construction of hypotheses and testing them. Hypothesis testing is involved in all 3 major types of study design: cross-sectional, case-control, and follow-up. The discussion below is for a case-control study comparing proportions. Similar formulations can be made for other types of studies and other measures such as means. A decision also must be made whether a 2-tail of 1-tail test is being used. The 2-sided test covers the joint testing of two inequalities between proportions, p1>p2 and p2>p1. The 1-sided test covers the testing of only one inequality, p1 > p2 or P2 > p1. The 2-sided test is preferentially used because it is more conservative. The null hypothesis for a 2-sided test states that there is no association between the exposure and the disease outcome which also implies an odds ratio of unity, OR=1. The alternative hypothesis for a 2-sided test states that there is association between the exposure and the disease outcome; the association may be positive with OR>1.0 or negative with OR <1.0. The null hypothesis for a 1-sided test states that there is either a negative or no association between the exposure and disease; OR=1.0 or OR<1.0. The alternative hypothesis for a 1-sided test states that there is a positive association between the exposure and disease outcome; OR>1.0.
C. PRELIMINARIES TO DATA ANALYSIS
Simple manual inspection of the data is needed before applying the tests above. Indiscriminate application of the tests to data leads to wrong or misleading conclusions. Acquiring familiarity with the data by simple manual inspection can help identify outliers, assess the normality of data distribution, and identify commonsense relationships among variables that could alert the investigator to errors in computer analysis. It is also most important that the data model be selected properly to suit the data at hand. The data models for continuous data can be straight line regression, non-linear regression, or show trends. The trends may be parallel or non parallel. Special models are used for repeated (paired) observations. More data models are used in the analysis of categorical data. The maximum likelihood model assumes a binomial distribution and derives the maximum likelihood estimate (MLE) which is the value of the parameter that maximizes the data function. The logistic model allows use of proportions to compare two groups. The chi square is used to compare 2 proportions where no raw data is available. If there are more than 2 outcome categories, cells of the tables can be collapsed to produce a 2 x 2 table. If this is not possible, the log linear model is used.
D. PROCEDURES USED:
Two procedures are employed in analytic epidemiology. The test for association is done first. The assessment of the effect measures is done after finding an association. Effect measures are useless in situations in which tests for association are negative. The tests for association commonly employed are: t-test, chi-square, the linear correlation coefficient, and the linear regression coefficient. The effect measures commonly employed are: Odds Ratio, Risk Ratio, Rate difference. Measures of trend can discover relationships that are not picked up by association and effect measures.
6.1.2 TESTS OF ASSOCIATION
A. TESTS OF ASSOCIATION ON MEANS:
The tests described below are used for continuous measurement data. Their details are described in elementary books of statistics. The Student t-test is used for two independent sample means. The Student paired t-test is used for two paired sample means. Analysis of variance, ANOVA (F test) is used for more than 2 sample means. Multiple analysis of variance, MANOVA, is used to test for more than one factor. Linear regression is used in conjunction with the t test for data that requires modeling. Dummy variables in the regression model can be used to control for confounding factors like age and sex.
B. TESTS OF ASSOCIATION ON PROPORTIONS: Single 2x2 Contingency Table
The tests for association described below can be applied to discrete data generated in 4 types of study design: cross-sectional, case-control, follow-up, and clinical trials. For two independent proportions, the chi-square test for independent samples is used (for large samples) and the Fischer's exact test is used (for small samples). For two paired proportions, the MacNemar chi-square test for paired samples is used for adequate samples and the Fischer exact test is used for small samples.
C. TESTS OF ASSOCIATION ON PROPORTIONS: Single 2 x k Contingency Table:
The table could be ordered qualitatively or quantitatively. The global chi-square test is used to determine if there is any associations in the table. More specific associations can then be studied by partitioning the table and obtaining partial chi-squares. The results from the partition analyses can be used to decide on how to collapse some cells into one another for further analysis. In the end the aim should be to collapse the complex table into a 2 x2 contingency table and then apply the methods described above. The M-H chi-square test for linear trend could alternatively be used in a 2 x k table. If the data is scanty with few cell counts that make the chisquare test invalid, the exact test can be employed. With extremely sparse data some form of modeling will yield better results.
D. TESTS OF ASSOCIATION ON PROPORTIONS:Stratified 2x2 Contingency Tables
A stratified design gives rise to several 2 x 2 tables, one table for each stratum. The Mantel-Haenszel chi-square statistic is used. It is a weighted average of separate chi-squares across the strata. In order for this test to be valid, the chi-square of each separate table must be homogenous across all strata. There are special tests of homogeneity that must be applied before the Mantel-Haenszel test is applied. The homogeneity test essentially indicates whether the separate chi-square test statistics are of the same order of magnitude and can therefore be combined in the M-H procedure. The M-H statistic is based on the hypergeometric distribution and follows a chi-square distribution with 1 degree of freedom. The following are wrong methods of combining data from several tables: summing chi-squares across tables, computing the chi-square of the combined (total) table, and computing chi-square from sum of 'O' and 'E' across groups. The M-H procedure breaks down in cases of too many strata and the multiple logistic regression procedure will have to be used in such cases.
There are two types of chi square: the Pearson and the Mantel Haenszel chi squares. They are defined as shown in the table below:
Exposure + | Exposure - | Total | |
Disease + | A | b | m1 |
Disease - | C | d | M0 |
n1 | n0 | N |
The Pearson chi square (cp)2 is defined as the summation over all cells of the table of Ã¥(O-E)/E = {n(ad-bc)2}/ {n1 n2 m1 m0}. The Mantel-Haenszel chi square is defined based on the ‘a’ cell only as (cMH)2 = {O(a) – E(a)} / Var(a) = {(n-1)(ad-bc)2} / { n1 n2 m1 m0}. The difference between (cp)2 and (cMH)2 is negligible when n is moderately large. Both statistics are reasonable approximations when cell frequencies exceed 5. (cMH)2 is preferred for stratified analysis and for computation of test-based confidence intervals.
The chi-square statistic has 2 components. The total chi-square is the sum of chi-square due to homogeneity and the chi-square due to association. Four assumptions must be fulfilled for the chisquare test to give valid results: the sample size must be big enough, the data must have been obtained by random sampling, observations must be independent of one another, and data must be normally distributed. Validity of the statistic is affected by the overall sample size but also by the cell numbers. According to Cochran, the statistic is valid if at least 80% of cells have more than 5 observed, at least 80% of cells have more than 1.0 expected, and at least 5 observed in 80% of cells. If the observations are not independent of one another as in paired or matched studies, the McNemar chisquare test is used instead of the usual Pearson chisquare test. The chisquare works best for approximately Gaussian distributions. The chi-square is a continuous distribution used for discrete data. This discrepancy calls forth the use of a continuity correction that is not agreed unanimously among statisticians. The Yates correction is used to correct for the fact that chi-square distribution is continuous but is used for discrete date. In the special case when degrees of freedom = 1, Yate’s correction to derive the correct formula for the chisquare as shown here: c2 = Ã¥ {(|obs – exp| - 0.05)2 / exp}. The shape of the chi-square distribution varies by the degrees of freedom. The chi-square statistic can not be negative. It can be zero if the expected is equal to the observed.. A distinction must be made between the significance and strength of the association. The chi-square statistic is a measure of significance of association and not degree of association. The coefficient of correlation, phi, measures the degree of association. The coefficient varies between case control, follow-up, and cross-sectional studies which makes it less useful that the odds ratio that is invariant across case control, follow-up, and cross-sectional studies. The phi coefficient is used to adjust a computed chisquare for sample size. Phi = {c2/ N}1/2 . Phi is considered the correlation coefficient for data in 2 x 2 tables. Cramer’s V is the equivalent of the phi coefficient in the r x c table.
USES OF THE
CHI SQUARE
CHI SQUARE
The chi-square statistic is very versatile and is widely used and misused. It is used to test for independence of 2 variables by comparing observed with what would be expected unfder the null hypothesis assumptions. It is used in to test for goodness of fit by comparing observed values of a distribution with those expected under the binomial, poisson, or normal distributions. It is also used to test for homogeneity among stratified 2 x 2 tables. It is also used to test for trend in 2 x k and r x c tables. The p-value obtained is only approximate.
INTERPRETATION OF THE REGRESSION MODEL
Trends in individuals like salt intake and BP
6.1.4 VALIDITY and PRECISION
A. MEASUREMENT AND ITS ERRORS IN EPIDEMIOLOGY
MEASUREMENT IN EPIDEMIOLOGY
An epidemiological study should be considered as a sort of measurement with parameters for validity and precision. Validity is an expression of the degree to which a measurement measures what it purports to measure. For example FEV1 (forced expiratory volume in one second) is a valid measure of respiratory function but skin-fold thickness is not on its own a valid measure of obesity ((page 95 John M Last: Public Health and Human Ecology. 2nd edition. Prentice Hall International, Inc ? year). Accuracy is the extent to which a measurement conforms to or agrees with the true value (page 95 John M Last: Public Health and Human Ecology. 2nd edition. Prentice Hall International, Inc ? year). Validity can be classified as internal validity and external validity. External validity is also called generalizability. Reliability is reproducibility i.e. does the instrument of measurement produce the same result under the same conditions all the time?
SOURCE OF ERROR
Error can arise as instrument error, digit preference, observer variation, variations in individual response, true biological variations, bias including confounding. Observer variation can arise in 2 ways. Within-observer variation is largely random. Random subject variation on repeat measurement regresses to the mean. Between-observer variation is usually systematic or biased (systematic) subject variation. Bias is defined technically as the situation in which the expectation of the parameter is not zero i.e. E(q-hat) is not 0. The following types of bias are explained in the next unit: Misclassification bias, Selection bias, and confounding bias. Bias may move the effect parameter away from the null value or toward the null value. In negative bias the parameter estimate is below the true parameter. In positive bias the parameter estimate is above the true parameter. A study is not valid if it is biased.
TYPES OF ERROR
Errors can be systematic or non-systematic (random). Systematic errors lead to bias and therefore invalid parameter estimates. Systematic or biased errors are known as dirty dirt because they bias conclusions and are therefore epidemiologically fatal. They are not decreased by increase of sample size. They are difficult to recognize nd hard to quantify. It is therefore difficult to compensate for them in the analysis. Random error or non-systematic errors lead to imprecise parameter estimates. Random error leads to misclassification. It is however not serious because it affects both comparison groups about equally and epidemiological study is concerned with making comparisons. Random errors lead to large standard errors in parameter estimates and can be controlled by increasing the sample size. However random errors will cloud association and underestimate true correlations in within-subject studies. It is possible to assess the magnitude of random errors by including measures of variability in the study. If the size of the random error is known and is small, the error can be tolerated as clean dirt.
B. INTERNAL VALIDITY
DEFINITION
Internal validity is concerned with the results of each individual study. Internal validity is impaired by study bias.
TYPES
Validity can also be described as selection validity, comparison validity, follow up validity, specification validity, measurement validity, and control selection validity. Validity is improved if the index group is similar to the comparison group in everything except the exposure of interest. Decreasing selection bias improves validity.
METHODS OF ASSURING VALIDITY
Selection validity is avoiding selection bias in case and control selection. Comparison validity is assured by removal of confounding bias by adjustment procedures or by covariate selection. In cohort studies the exposure groups must be comparable in everything except the exposure. In case control studies, the risk of disease in the controls must be the same as the cases had they got the exposure of interest. Random sampling reduces confounding because it theoretically makes the distribution of the confounders in the sample the same as in the original population. Follow-up validity requires that the two populations being compared be followed up in the same way. It is relevant to cohort studies and irrelevant to case control studies. Specification validity requires that the right sampling and analytic methods be used. Measurement validity is an issue in both cohort and case control studies. It requires that measurement errors be removed. The errors may be mistakes in the procedure, errors in the selection and use of proxy variables, or construct errors due to ambiguities in variable definitions. Errors can be classified as non differential and differential. Non differential are random errors whereas differential are systematic errors. Errors may be independent of one another or may be dependent on one another. The situation of dependent errors is worse because the effect of an error is compounded by its dependence on other errors.
C. EXTERNAL VALIDITY
In external validity, inference is pertinent to the general population. Traditionally results are generalized if the sample is representative of the population. In practice generalizability is achieved by looking at results of several studies each of which is individually internally valid. It is therefore not the objective of each individual study to be generalizable because that would require assembling a representative sample.
D. PRECISION
DEFINITION
Precision is a measure for lack of random error. An effect measure with a narrow confidence interval is said to be precise. An effect measure with a wide confidence interval in imprecise. Precision is increased in three ways: increasing the study size, increasing study efficiency, and care taken in measurement of variables to decrease mistakes.
STUDY SIZE
The appropriate study size is based on consideration of both the study precision and study cost. Larger and hence more precise studies are costly. Striking due balance between these two contradictory objectives is difficult. Usually the epidemiologist has to work amidst constraints of finding enough study subjects and finding enough funding.
STUDY EFFICIENCY
Study efficiency depends on study design. An efficient sampling design will decrease sampling variation (sampling error). Study efficiency is assessed as the amount of information obtained per subject or the amount of information obtained per unit of cost. An efficient study design requires a balanced consideration of the following sometimes opposing parameters: the proportion of subjects exposed, the proportion of the exposed who will develop disease, and distribution of study subjects according to potential confounders.
IMPROVING STUDY PRECISION
A. DEFINITION and HISTORICAL BACKGROUND
DEFINITION
HISTORICAL BACKGROUND
Review articles written by discipline leaders were the most popular method of combining findings from various studies. They however had two serious draw-backs. They were subjective and therefore prone to error and bias. The reviewer was free to make decisions on what data and conclusions to emphasize. Secondly they dealt with summarization of conclusions without looking at the data on which the conclusions were based. Meta analyis was first developed for randomized clinical trials. Its use in observational studies is more problematic mainly because of the difficulty of knowing and adjusting for confounders.
B. ADVANTAGES OF META ANALYSIS
Meta-analysis has become popular with the proliferation of epidemiological studies on particular subjects. Writers of review articles and practicing epidemiologists would like to have some form of consensus or summary of the findings of various studies. Meta analysis enables computation of an effect estimate for a larger number of study subjects thus enabling picking up statistical significance that would be missed if analysis was based on small individual studies. Many clinical trials especially with invasive intervention can not recruit enough patients in one center to reach statistical significance. Meta analysis also enables study of variation across several population subgroups since it involves several individual studies carried out in various countries and populations. Meta analysis makes the process of reviewing several studies with view to reaching a general conclusion very transparent because it is based on quantitative assessments.
C. STEPS IN META ANALYSIS
The first step is to identify the variables of interest such as outcome, exposure, confounder, intermediate, and effect modifying variables. The variables are used together with other relevant key words to identify relevant articles from the data bases of such as medline or from other unpublished sources. Criteria must be set for what articles to include or exclude. These include definition of what effect measures are used and the populations covered. It is best if there is as much homogeneity in the population as is practical. Data collection is carried out by abstracting information from the articles on a standardized data abstract form with standard outcome, exposure, cofounder, or effect modifying variables. The ideal is to abstract raw data and reanalyze it in a standard way to enable combination of several studies to get one summary effect measure. In practice the raw data is not available and only effect measures are available. The first step is to display the effect measures with their 95% confidence limits to get a general idea of their distribution before proceeding to compute summary measures. The summary effect measure, OR or b, is computed from the effect measures of individual studies using weighted logistic regression or computing a MH weighted average in which the weight of each measure is the inverse of its precision ie 1/(se)2. In both the logistic or MH procedures, each study is treated as a stratum. The combined effect measure is then statistically adjusted for confounding, selection, and misclassification biases. Tests of homogeneity can be carried out before computing the summary effect measure. Sensitivity analysis is undertaken to test the robustness of the combined effect measure.
D. DIFFICULTIES OF META ANALYSIS
E. READING RESULTS OF META ANALYSIS
The following questions should be considered when reading review or meta-analysis articles: are the methods clearly stated (b) was the search for articles comprehensive enough?, were the criteria for selecting articles for review stated, were the criteria objective and were they adhered to?, was there a possibility of bias in the selection of those articles?, was the methodologic quality of each article assessed?, were differences between studies explained or were they just glossed over?, was the combination of results from the primary studies appropriate?, and were the conclusions of the reviewer supported by data?
UNIT 6.2
SOURCES and TREATMENT OF BIAS
Learning Objectives:
· Misclassification bias: information bias & detection bias
· Selection bias: exposure bias, referral bias, non-response bias, & follow-up bias
· Prevention of confounding bias at the design stage by stratification & matching
· Treatment of confounding bias at the analysis stage by standardization, stratified analysis, and multivariate adjustment
· Sampling bias
· Bias, confounding bias
· Bias, detection bias
· Bias, exposure bias
· Bias, follow-up bias
· Bias, information bias
· Bias, misclassification bias
· Bias, non-response bias
· Bias, referral bias
· Bias, sampling
· Bias, selection bias
· Confounding characteristic
· Confounding factor
· Confounding variable
· Matching
· Multivariate adjustment
· Standardization
· Stratification
· Stratified analysis
UNIT OUTLINE
A. Wrong statistical model
6.2.5 SURVEY ERROR and SAMPLING BIAS
A. Types of Sampling Errors
B. Sampling Bias
C. Assessment of Validity and Precision
CONFOUNDING AND EFFECT MODIFICATION
6.2.5 SURVEY ERROR and SAMPLING BIAS
A. TYPES OF SAMPLING ERRORS
Total survey error is the sum of the sampling error and three non-sampling errors (measurement error, non-response error, and coverage error). Sampling errors are defined as….. Measurement errors arise due to wrong data collection. Non-response error arises when subjects selected into the sample refuse to provide information. Coverage errors are due to failure to cover the whole anticipated sampling frame. Sampling errors are easier to estimate than non-sampling errors. Sophisticated mathematical techniques have been developed to measure their magnitude. Sampling error decreases with increasing sample size. Non-sampling errors may be systematic like non-coverage of the whole sample or they may be non-systematic. Non-systematic errors cause severe bias.
B. SAMPLING BIAS
Sampling bias, positive or negative, arises when results from the sample are consistently wrong (biased) away from the true population parameter. For example bias is said to arise if the sample studied is not from the target population or is not representative of the target population. In such a case we say that the sample represents only the population of inference that is a small segment of the study population. Inferences from a biased sample can not be applied to the target population. There are several sources of bias: (a) incomplete or inappropriate sampling frame such as use of the telephone directory when not all people have telephones (b) Use of a wrong sampling unit for example sampling on the basis of the event of hospitalization when the proper way would have been on the basis of persons since one person can be hospitalized more than once (c) Non-response bias (d) measurement bias for example Interviewer bias (b) coverage bias when data is collected from only part of the sample (f) sampling bias. Non-response becomes a problem only if it violates the randomness of the sample. Non-response bias is increased if there is a relationship between the target variable and response behavior. For example if all people who are drug addicts consistently refuse to answer the study questionnaire, the results of the study will be severely baised. Correction for non-response bias is not easy. It requires using information from a proxy or auxiliary variable whose population distribution is known. Such variables are not easy to find and even if found more work may be needed to collect the necessary information. In the few cases in which full information from a proxy variable is available, we may use regression estimators to correct for the bias due to non-response bias.
C. ASSESSMENT OF VALIDITY and PRECISION
Variability in statistics computed on the basis of sample data may be due to variability in measurement, measurement error, or may be due to sampling error. Both variance and covariance are measures of the efficiency of the sampling.
D. SENSITIVITY ANALYSIS
Sensitivity analysis can be carried out for the major types of bias. In analysis of confounding, If we have no information on confounders we can assume or simulate data and see the effects of adjustment on the effect estimates. In analysis of misclassification, data simulation is carried out by assuming various misclassification probabilities and seeing the effect on the effect measure. The probabilities used can be obtained from a special validation sub-study. In analysis of selection bias, the OR can be adjusted by using the probability of case or non case selections.
UNIT 6.3
HEALTH STATUS
Learning Objectives
- Hospital data: sources, types, uses in public health
· Definition, classification, and uses of various rates and proportions: vital, demographic, and morbidity
- Crude rates, specific rates, and standardized/adjusted rates.
- Population pyramid: age and sex-structure of different populations
Key Words and Terms
· Catchment area
· Demographic, life table
· Demographic, rates
· Demographic, shift
· Epidemiological, transition
· Health status indicators
· Hospital, morbidity
· Hospital, bed capacity
· Hospital, information system
· Hospital, mortality
· Hospital, planning
· Hospital, rates
· Hospital, records
· Life, expectancy
· Life, table (abridged)
· Life, table (cohort or generational
· Life, table (current)
· Life, table (followup)
· Patient discharge
· Population, control
· Population, density
· Population, dynamics
· Population, growth
· Population, life tables
· Population, pyramid
· Quality assurance
· Rate, adjusted rate
· Rate, birth rate
· Rate, crude rate
· Rate, death rate
· Rate, rate density
· Rate, growth rate
· Rate, hazard rate
· Rate, incidence rate
· Rate, morbidity rate
· Rate, mortality rate
· Rate, pregnancy rate
· Rate, specific rate
· Rate, standardized rate
· Rate, vital rate
· Ratio, mortality ratio
· Ratio, odds ratio
· Ratio, risk ratio
· Ratio, vital ratio
· Record linkage
· Records, medical record
· Registration, cancer registration
· Statistics, vital statistics
· Survival rate
A. Hospital Information Systems
B. Medical Records Department
C. The Medical Record
D. Performance Indicators: Process and Outcomes
E. Hospital Administration
6.3.3 DISEASE REGISTRIES WITH CANCER AS AN EXAMPLE
A. Definition and Uses
B. Methodology of Data Collection and Validation
C. Incidence and Prevalence Analysis
D. Etiological Studies
E. Other Disease Registries
A. Introduction
B. Birth Rates, Ratios, and Proportions
C. Death (Mortality) Rates, Ratios, and Proportions.
D. Marriage and Divorce
E. Morbidity Rates
A. Definition of Basic Rates
B. Population Pyramid
C. Population Projections
D. Demographic Shift/Transition
A. HOSPITAL INFORMATION SYSTEMS
B. MEDICAL RECORDS DEPARTMENT
The Medical records department is staffed by trained medical record administrators. Its roles are to collect, store, and access information. The information is needed for health care, financial or administrative reasons. The efficiency of the medical records department impacts on clinical care and administration. Quality control is needed to eliminate inaccuracy and inconsistencies. Regular chart review is part of hospital quality assurance. Quality involves both completeness of recording and accuracy. Ethical and legal issues arise in medical records departments. The computer revolution enables collection, storage, and retrieval of a lot of information about the patient. In order for this information to be used, it must be readily available to all care-givers. With so many people accessing information the issues of privacy arise in a new form that was not known before the computer age. The problem becomes worse if the data is available on local, regional, or national data net-works.
C. THE MEDICAL RECORD
Medical forms are used to record all medical information and activities: Traditionally medical information was written by hand on various forms each serving a particular purpose. Birth and death certificates are legal documents. The patient's chart consists of admission record, physician notes, psychology notes, social welfare records, laboratory and pathology records, nursing notes, pharmacy records, diet record, dental record, discharge notes, and autopsy records. The forms used are updated regularly to keep up with developments in medical technology and to be user-friendly.
The Problem oriented medical record was developed to make decision making and follow-up easier. One of the most popular problem-oriented record systems is SOAPIE. SOAPIE is an acronym for Subjective complaints, Objective complaints, Assessment, Plan, Intervention, and Evaluation.
The electronic medical record (EMR) is becoming popular in many hospitals. Some written records are being scanned to save the time of key-board entry. Physician notes, unlike other types of records, are more difficult to capture in a coded form. This will however be gradually overcome when physicians learn to use check-lists more effectively. EMR enables use of clinical data net-works at the hospital level, the local level (LAN), the national, and the international levels.
Clinical work station: many hospitals realizing the importance of EMR are experimenting with direct on-line data entry. Computer terminals on wards, laboratories, pharmacies, and other services are used to input data directly into the EMR. For some types of laboratory assays the data is automatically read into the data-base without the necessity of any additional data-input.
Record linkage: The use of computers, networks and sophisticated software can now enable linkage of many types of records belonging to the same person. The following are examples of linkable records: birth, death, physician, nursing, pharmacy, laboratory, socio-demographic information, and health insurance data. It is conceivable that in the future some personal records like credit-card based purchases of food can also be linked to provide an overall profile of the patient. Some consensus is needed on uniformity of data recording so that record linkage will be easier.
Integrated medical record: Computer technology makes it possible to develop integrated records. The integration must be patient-centered to be useful clinically and for epidemiological purposes. The integrated record should consist of the following types of records: hospital information system, radiological information system including archiving and retrieval of images, laboratory, blood grouping and cross matching, pharmacy, anesthesia, and surgery.
D. PERFORMANCE INDICATORS: PROCESS AND OUTCOMES
Clinicians, administrators, and other stake-holders are interested in performance indicators to evaluate health care delivery and make necessary strategic decisions. Two types of indicators are used: process indicators and outcome indicators. Epidemiological methodology is used in the computation and interpretations of various rates of these indicators. Process indicators describe the daily routines, work out-put or productivity. These include admission rates, discharge rates, average stay in bed-days, average bed occupancy, number of medical/surgical procedures, prescriptions filled, laboratory assays done, and radiological examinations undertaken. Administrators are interested in process indicators for planning, staffing, and cost-control purposes. Outcome indicators describe the end-results of medical intervention. Mortality, morbidity, and chronic disability are commonly-used negative outcome indicators. Positive outcomes such as cure rates can also be used. There is still need for developing consensus on outcome classification to enable uniform reporting. The discharge abstract is the source of the routinely used outcome information. Discharge data can be used to estimate morbidity and mortality due to specific conditions. It can also be used to compare outcomes for various medical or surgical procedures and to compare outcomes across various hospitals or institutions. The clinical data-base and patient profiles can be used to predict the rates and risks of various outcomes. Outcome prediction is more important for critical illnesses. It is interesting to compare the predicted and the actual observed statistics. The data-base can also be used to profile patients and the behavior of caregivers relating them to outcome. The data can also be used to explain the factors leading to various outcomes. Clinical, pathological, and radiological information readily available on line can be used to select predictive factors of certain outcomes.
E. HOSPITAL ADMINISTRATION
Planning and projections: Hospital managers as decision-makers analyze hospital data for use in planning, projections, cost analysis, and assessing accessibility and affordability of the services provided. There is concern sometimes about excess bed capacity or inadequate bed capacity. Decisions may have to be made about hospital closure or hospital mergers. The total staffing as well as the structuring of jobs may have to be changed with changing patient and disease profiles. Both epidemiological and clinical data are used to guide planning and projections. Epidemiological disease rates can enable good estimate of disease burden in the community. They also can be used to detect trends and changes in disease incidence that require changing the total number of beds provided or their mix. Data on discharge diagnoses from the hospital data-base can provide information about trends that are used in planning. Trends in disease occurrence can be identified earlier from hospital than epidemiological data.
Cost analysis: Hospital administrators are interested in both the total cost of health care services as well as the cost of individual diagnostic categories. They also want to make cost-effectiveness and cost-benefit analyses. This data enables them to cut costs where necessary and also to set reasonable rates for their services. Objective financial decisions are based on high quality data. Cost analyses also affect the behavior of care-givers. Physicians for example would be more restrained in ordering some tests of they knew their true costs. Regression techniques can be used to estimate rates for procedures and services. Epidemiological methodology is used in definition of disease categories and also computing rates for cost analysis. Special epidemiological studies may have to be set up to answer specific questions.
Access and affordability: Discharge abstract can be used to give information about community utilization of services. For better interpretation financial and demographic information is also used.
Surveillance: A systematic review of the hospital clinical data-base can reveal conditions that may be missed. It can also catch early trends of changing disease risk. Case-finding can detect undiagnosed or under-treated cases.
EPIDEMIOLOGIC STUDIES
The availability of a lot of computerized data on individual members of the community has opened up a new area of epidemiological investigation. Healthy individuals undergo clinical, laboratory, and radiological assessment on a regular basis as part of annual physical fitness exercises or as part of disease screening or for purposes of employment, life, and health insurance. Socio-demographic data is collected at the same time. When they fall sick they go to hospitals where their clinical data is recorded. The availability of record linkage will bring together all these records so that etiological relationships can be explored. It is possible to design case control and follow up studies.
The following methods are used for public health decision making: operation systems, decision trees, and game theory. Operational research reduces systems and problems to mathematical relations that aid decisions. Statistical modeling can either be based on memoryless distributions such as the Binomial and the poisson or memory distributions such as the Markov process. When the system is too complex to model resort is made to Monte Carlo techniques. Operations research is used in a wide variety of public health applications: health facility management, outpatient appointments, inpatient admission/discharge scheduling, facility sizing, staffing, support services, regulatory health planning, ambulance systems, epidemic control, clinical decision making, program evaluation and policy analysis. Decision making by decision trees uses sequence trees. Game theory is analysis of a decision against the expected. Zero sum games involve 2 persons. The games may be cooperative or non zero sum games.
PROBLEM SOLVING IN PUBLIC HEALTH
The problem must be defined and its magnitude is determined. The determined are identified. Strategies and priorities are then considered before deciding on a solution that is then implemented. Evaluation must be carried out at the end.
PUBLIC HEALTH PLANNING
6.3.3 DISEASE REGISTRIES WITH CANCER AS AN EXAMPLE
A. DEFINITION and USES
Cancer registration is continuing, systematic collection of data on reportable neoplasms. The data is comprehensive including socio-demographic, clinical, laboratory, radiological, and treatment variables. The cancer registry helps physician in follow-up of patients because it may be the most integrated data source available. It also contributes data to community, regional, or national cancer registries. Data collected within the hospital can be used for hospital-based epidemiological studies.
B. METHODOLOGY OF DATA COLLECTION and VALIDATION
The registry obtains information from various sources: pathology logs, discharge summary, death certificates, and clinical diagnoses. When a case is first identified using one data source, the other sources are also examined to validate and complete the picture. The problem of duplication of records may arise because cancer patients are treated and may go into remission. On recurrence of cancer at the same site or at another site, the patient mat be admitted again and may be treated as a new case. The classification of cancer sites is not easy especially where multiple primaries are involved and where secondary cancers occur following immunosuppressive treatment.
C. INCIDENCE AND PREVALENCE ANALYSIS
Hospital registry data does not give true incidence rates because one hospital cannot capture all cases of cancer arising from the community. The incidence rates computed from hospital cancer registries are considered minimal rates. If the hospital is a major referral center, it receives cases from far away places which may inflate the incidence rates for the immediate community unless care is taken to eliminate them from the analysis. The following types of analyses can be made: analysis for trends of incidence, analysis for immediate causes of death, analysis of survival, and analysis of treatment trends.
D. ETIOLOGICAL STUDIES
The hospital cancer registry is a good source of controls for case control studies of cancer etiology
E. OTHER DISEASE REGISTRIES
Disease registries started with cancer. At the moment registries are being set up for other non-communicable diseases.
A. INTRODUCTION
Most of vital statistics are from mandatory reporting. As required by law, reports are made on a prescribed form in the prescribed manner. The event must be recorded when and where it occurs. Other sources of vital information are questionnaires, interviews, observations (informal, critical, and controlled), census, and special studies (case histories, surveys). Hypotheses can be generated from analysis of vital data for example a cancer mortality map may suggest hypotheses on environmental causes of cancer. Vital data can be analyzed in conjunction with ecologic or environmental data. Vital data is used in cohort analysis to test hypotheses on the effects of exposures early in life. Cohort effects and period effects can be identified from the data.
B. BIRTH RATES, RATIOS, and PROPORTIONS
Birth rates are affected by the level of socio-economic development. Industrialized countries have lower birth rates than developing countries. The birth rate is reflected in the population structure. Countries with high birth rates have relatively more youthful populations. Ascertaining births in developing countries is not complete because of logistic difficulties. There is confusion between rates and proportions in birth statistics. The crude birth rates are true rates. The still-birth, premature, low-birth weight rates are actually ratios or proportions and not true rates although they are customarily referred to as rates. Crude Birth Rate is the total number of births per 100,000 of mid-year population per year. Crude Live Birth Rate is the number of live births per 1000 of mid-year population per year. Still-birth (fetal death) Rate is the number of fetal deaths above 20 weeks per 1000 births (Live + still births) per year. A still-birth is defined as a fetus delivered weighing at least 500 grams that showed no signs of life during expulsion or extraction from the mother. The corresponding gestational age is 22 weeks and the corresponding body length is 25 cm crown to heel. Premature Birth rate is the number of births at gestation age 28-48 weeks per 1000 live births per year. Low birth weight rate is the number of births with weight less than 2500 grams per 1000 live births per year. Very low birth weight rate is the number of births with weight less than 1500 grams per 1000 live births per year
C. DEATH (MORTALITY) RATES, RATIOS, and PROPORTIONS.
OVER-VIEW
Mortality is highest at extremes of life, infancy and old age. At all ages male mortality is higher than female mortality. Mortality of married persons is lower than that of the unmarried. Environmental factors like climate, seasonal variation, urban or rural residence and SES indicators explain some of the variation in mortality rates. There are problems of certification of cause of death. It may not be possible to distinguish immediate from underlying cause of death. Only one main cause is written but we know that several other causes contribute. Some physicians are not serious in certifying death. Final diagnosis from autopsy comes too late for death certification; rarely do death certificates get corrected. Changes in disease classification ICD revised every 10 years make it difficult to compare and study trends. Differences in diagnostic terminology and development of new diagnostic categories make comparisons difficult. There is confusion in mortality statistics between rates and proportions or ratios. Many proportions are loosely called rates. In deference to common usage we have not made corrections to the terminology but the discerning student should recognize that some of the rates are actually proportions. The general formula for death rates is: # deaths in a year / # of population in that category at mid-year. The formula for proportions does not prescribe any time period.
CRUDE AND SPECIFIC RATES/PROPORTIONS
The Crude Death Rate is the number of deaths in a year per 100,000 of mid-year population. Crude rates are not very useful. Specific rates are more meaningful. Specific rates are computed by age, gender, and race for purposes of public health analysis, 4 types of rates are used: standardized or adjusted, and specific. Standardized/Adjusted Death Rates take into account the population age and sex structure such that they are comparable with other populations. Age-Specific Death Rates refer to deaths at certain ages that are of public health importance for example infant mortality rate, neonatal mortality rate, post-neonatal mortality rate, and peri-natal mortality rate. The Cause-specific Rate is the number of deaths from a specified cause per 100,000 mid-year population. The cause-specific death ratio is the percentage of deaths from a specific cause used as a comparative measure of death. Autopsy studies have diagnostic accuracy for cause of death but suffer from selection bias Additional special rates are also used. The proportional mortality ratio is akin to the cause-specific rate. It is the number of deaths of a specified kind for example above 50 years of age or due to malaria as a proportion of the total number of deaths in the year. It is used to compare mortality experience across communities.
SPECIAL RATES
Case-fatality ratio is the proportion of deaths from persons with a specified disease condition.
Fetal deaths: The fetal death ratio is defined as the total fetal deaths in a year as a proportion of the total live births in a year. It is specifically defined as deaths at or below 20 weeks of gestation per 1000 live births thus still birth rate = (number of stillbirths in a year) / (no of live births + number of stillbirths) x 1000. The abortion ratio is the number of induced abortions per 1000 live births.
Infant mortality: IMR is the most important indicator of community health. The Infant Mortality Rate is the number of deaths at ages 0-12 months per 1000 live births per year. Neonatal mortality: Neonatal mortality is due to pre-maturity, congenital anomalies, peri and neonatal care, trauma. The Total Neonatal Mortality Rate is the number of deaths of liveborn infants weighing at least 500 grams (corresponding to 22 weeks of gestation or 25 cm crown to heel length) within 1-27 days of birth per 1000 live births. Thus neonatal mortality = (number of neonatal deaths in a year) / (number of live-borns in a year) x 1000. The neonatal mortality rate has two components, early and late. The Early Neonatal Mortality Rate is the number of deaths within 7 days of birth per 1000 live births per year. The Late Neonatal Mortality Rate is the number of deaths at age 7-27 days of birth per 1000 live births per year. Post Neonatal Mortality Rate is the number of deaths aged 28days -1 year of birth per 1000 live births per year. It reflects the impact of nutrition, sanitation, SES, medical care, and infections. Peri-natal Mortality Rate is the number of deaths from gestation age above 28 weeks up to 7 days of birth per 1000 total births (live and still-born) per year thus perinatal mortality rate = (number of stillbirths in a year + early neonatal deaths in a year) / (number of live births in a year + number of stillbirths in a year) x 1000 . The perinatal mortality ratio is the number of fetal deaths >= 28 weeks + deaths within one week of birth per 1000 live births.
Maternal Mortality Rate is the number of deaths in pregnancy or within 42 days of delivery per 100,000 births. The ideal denominator should be the number of women who were pregnant in the year but this is not customarily done. The following problems arise when live birth is used as the denominator: (a) fetal deaths are excluded thus inflating the maternal mortality rate (b) maternal deaths are counted only once for twin pregnancies. This inflates the MMR. (c) Live births are under-registered in developing countries where the problem of MMR is highest. This serves to inflate MMR (d) maternal death may occur in the year following that of birth and may not be counted.
INTERPRETATION OF TRENDS IN RATES
The determinants of mortality are: (a) socio economic: age, sex, marital status, family size, income, education (b) health-related behavior: smoking, alcohol, drugs (c) maternal factors: maternal age, parity, child spacing (d) environmental (e) nutrition (f) injuries (g) medical care. Artifactual causes of changing disease trends may be due to (a) misdiagnosis of related diseases for example lung cancer may be misdiagnosed as tuberculosis (b) wrong census denominator. Age-specific death rates follow a U-shape being high in infancy and old age. Neonatal deaths are due to congenital anomalies or delivery complications. Post-neonatal deaths are due to social and environmental factors.
D. MARRIAGE & DIVORCE
The marriage rate is number of marriages in a year per 1000 of population. The divorce rate is defined as number of divorces in a year per 1000 or population or per 1000 marriages. It is wrong to relate divorce rates in a year to marriage rates of the same year because those married do not necessarily divorce in the same year. The age at marriage is measured as mean or median age at marriage. Age at remarriage can be described for the 1st, 2nd, 3rd remarriage.
The rates of marriage and divorce are affected by war, economic depression, recession, and social attitudes. The total marriage dissolution rate includes both separation and divorce. Separation is usually not recorded in vital events. The divorce rate can be computed by age, order of marriage, and by marriage duration. The median age at marriage or divorce differs by gender. It is also possible to compute the median duration of marriage. Multi-state life-tables can be used to study marriage and divorce. Marriage leads to lower mortality with males benefiting more than females. This is explained by two factors. The healthier and stable persons are more likely to be selected for marriage. Marriage also ensures psychological and economic support.
E. MORIBIDITY RATES
OVERVIEW
Mandatory reporting of specified diseases has been legislated into law in many countries. Compulsory reporting of infectious disease started in one town in UK in 1876. It had become national by 1889. The purposes of infectious disease notification are: (a) access to treatment (b) local administrative action (c) epidemic control (d) research (e) aid diagnosis. The sources of data are: Compulsory notification data, diseases registries (cancer, substance abuse, birth defects, mental, congenital anomalies), Hospital discharge data, health service utilization indices, health status indicators, Ministry reports, and Health, Nutrition and Morbidity Surveys. Both seasonal and cyclic trends in disease rates must be studied. Data from medical records gives information on clinical, demographic, sociologic, economic, administrative, and behavioral variables.
NOTIFIABLE DISEASES
Infections/communicable diseases: The following are examples of reportable conditions: (a) GIT infections: hepatitis, cholera ,typhoid & paratyphoid, amebic dysentery & bacillary dysentery (b) food poisoning (c) respiratory infections: tuberculosis, diphtheria (d) parasites: malaria (e) sexually transmitted td: syphilis, HIV (f) viral: dengue
OTHER DISEASES
Accident statistics: (a) industrial accidents (b) non-industrial accidents (road traffic accidents, sports and recreation accidents, home accidents)
CHILD HEALTH
Child growth & health: The following parameters are used to assess child growth and development in the community: (a) Nutritional status: weight for height, BMI (b) Low birth weight rate (c) Mean length 0-1 year; mean height 1-18 years (d) Mean weight 0-1 year, 1-18 years (e) Chest Circumference (f) % immunized fully. School health: vision defects, hearing defects, and dental defects. Food intake: The following are used to assess food intake: (a) Energy kcal/day (b) Protein g/day (c) Fat g/day (d) Minerals (Ca, Fe)
HEALTH CARE DELIVERY SYSTEM
Medical facilities & personnel: (a) # hospital beds per 10,000 population (b) Hospital stay: patien-days/100,000 population (c) Bed occupancy: # bed-days per year (d) Admissions and discharges per 100,000 of population (e) # outpatient visits per year (f) Physicians, dentists, pharmacists, midwives, & nurses per 100,000 population
Hospital statistics: admissions, discharges, diagnoses, and procedures
A. DEFINITION OF BASIC RATES
INTRODUCTION
Interpretation of demographic rates requires knowledge of the following data: Per capita income/ per capita GDP, Adult literacy, Women of child-bearing age as percentage of all women, Contraceptive use rate, Age and sex structure of population, Urban-rural distribution, Dependent vs economically active population. The factors of fertility are: age and sex distribution of the population, socioeconomic indicators (education, occupation), cultural and religious attitudes, marital status, and duration of marriage. Other factors are: migration, mortality, marriage, and desired family size.
FERTILITY (Measures of fertility p830 Vol 2 Textbook of Public Health)
The Total Fertility Rate is the number of births per year 1000 women in mid-year population aged 15-44. Age-specific, standardized, and differential fertility rates could be computed. The gross reproductive rate is reproductive rate computed for girls only. The net reproductive rate is the proportion of girls surviving to the reproductive age out of 1000 live births. Determinants of fertility: exposure to sexual intercourse, exposure to contraception, and gestation & successful parturition. The replacement level is TFR of 2.1. The contraceptive failure rate can be measured using life table methods. Multiple decrement life tables are used to account for competing causes of contraceptive failure. Adolescent pregnancy in developed countries is unplanned. It is planned in developed countries. Factors of adolescent pregnancy: (a) socio-economic and demographic: age, race, ethnicity, religion, education, parent income. (b) psychological: self esteem. (c) family influence. (d) Peer influence. (e) School performance. (f) The mass media. (g) Risk behavior: drugs, alcohol. (h) Physiology: level of hormones, onset of menarche. The consequences of adolescent pregnancy are: dystocia, economic and educational disadvantages, lower chances of marriage, and increased mental instability
MIGRATION
Migration may be internal or international. Migration involves pull factors such as economic opportunities and push factors such as war and persecution. Migrants have particular personal, social, or economic characteristics. Interpretation of migratory studies requires knowledge of the following: (a) pre-migratory environment (b) age at migration (c) selection factors for migration. Migratory studies can be designed to compare migrants with siblings who stayed in the mother country
B. POPULATION PYRAMID
Population composition is important in public health because it can indicate disease risk, risk-related behavior. Table 833-837 in Textbook of Public Health Vol 2. Data used for demographic analysis is the national census, vital statistics, sample surveys, and qualitative surveys. Demographers describe a population using characteristics that do not change: age, sex, race/ethnicity, marital status, education, and occupation.
The population pyramid of a country reflects both birth and death rates. The birth rate has a bigger impact on the pyramid that the death rate. The pyramid shows the population structure by age, gender and other variables that may be chosen. The age structure is affected by : fertility, mortality, migration. Developed industrial countries have population pyramids with a narrow base and a wide top. Developing countries have a wide base and a narrow top. Atypical pyramids: Major sudden dislocations like war or genocide could produce atypical pyramids by absence of certain age or sex-groups. Migration of young unmarried laborers produced a bulge in the middle of the pyramid the receiving country while leaving a corresponding crest in the originating country. Migration of families with young children causes widening of the base in the receiving and its narrowing in the sending countries.
The size and structure of a population is determined by the birth rate, the death rate, and migration. Fertility has a bigger impact on population age structure than mortality. A high proportion of elderly persons with a deficiency of males indicates higher female longevity. A higher proportion of the elderly will lead to increase of the death rate. Cohort analysis of mortality can indicate long term changes in health.
C. POPULATION PROJECTIONS
Population projections are based on fertility, mortality, and migration data. The rate of natural population increase, NI, is the difference between the crude birth rate (CBR) and the crude death rate (CDR). Negative population growth occurs when CDR > CBR. Positive population growth occurs when CBR > CDR. Population replacement requires that each couple, on the average, produces 2 offspring. The rate may be lower than this in situations of immigration. Population projections can not be too precise. Estimates of population increase can be made in 6 different ways: (a) arithmetic method: this assumes an increase of the population by a constant amount each year. This could happen in a situation in which the native population registered neither increase nor decrease (zero population growth) and there is regulated immigration of a fixed number every year. (b) Geometric method: this assumes increase of the population by a constant rate every year. (c) Graphic extrapolation: Extrapolations of graphs beyond the area covered by available data can be used to estimate future population assuming continuation of current trends (d) National vital statistics: use of the previous census together with birth, death, and migration data to estimate the population (e) use growth curve formulas.
D. DEMOGRAPHIC SHIFT/TRANSITION
This term is generally used to refer to changes in population structure that occur with socio-economic development. They are a consequence of falling birth rates and falling death rates. The proportion of children decreases while that of the elderly increases.
E. DEMOGRAPHIC POPULATION LIFE-TABLE
DEFINITION
The British astronomer Edmund Halley first developed a life-table to describe longevity of 17th century residents of Breslau . In 1815 Joshua Milne published the first exact life table based on mortality in a city in northern England . The population life table describes the current mortality experience of a given population. Its concept can be used in survival analysis studies as we shall see later.
CLASSIFICATION
Life tables can be classified as current or cohort (generational) life tables. The current life tables are the ones most often used. Current life tables are constructed by applying current population death rates to a hypothetical population of 100,000. They change every year with the publication of new current death rates. Cohort life tables are constructed by following a given cohort through its life. Death rates appropriate to each age of the cohort are used. Life tables can also be classified according to the factors of attrition. Ordinary life tables show attrition of a cohort from a single factor like death. Multiple decrement tables show attrition from 2 or more attrition factors. Multi-state life-tables describe changing states like marriage, divorce, and how they impact on death. Life tables can be general or can be constructed for different gender, race, and place groups. A complete life table is computed for each single year of age. An abridged table groups the ages instead of using individual years. The life table can be constructed at birth or starting at the exposure to risk. The cause-elimination life table gives life expectancy after removing competing causes of death
USES
Actuarial, pension computation & annuities, assess/compare health services. Four types of information can be obtained from a life table. It shows life expectancy at birth and at any future age. It shows survival probability which is the probability of survival for a person at a certain age to a given age in the future. Death rates are derived from life tables and are used to compare populations since the rates are inherently age-standardized. The potential years of life lost can be computed as the difference between life expectancy and age at death.
ADVANTAGES
The life table has several advantages. It enables conversion of age-specific death rates to life expectancies. It is easier to interpret life expectancies than death rates. Life-table death rate is the reciprocal of life expectancy at birth. The life table also enables comparison of different populations independent of the population age distribution. It is therefore possible to compare mortality experiences without going through the troubles of standardization.
CONSTRUCTION
Construction of the table: We start with the assumption that a hypothetical population of 100,000 were born. The life table has 7 columns as explained below. Column #1 is the year of age, x, listed from 0 to 100 (it is possible to list above 100 if higher life expectancy is likely). Column #2 is the age-specific death rate for each year of age, q, obtained from the most recent vital statistics. It is the probability of death in the year interval. Column #4 shows the number of survivors at the start of the year, l. For the zero year, l is set equal to 100,000. The value of l for subsequent years is obtained by subtraction of the number of deaths in the previous year. Column #4 shows the number of persons dying in the year. This is computed by applying the age-specific death rate, q, to the number surviving at the start of the interval. Column #5 shows the total number of years that the 100,000 persons lived in that year, L. It is computed by subtracting one half of deaths in the year, d , from the number surviving at the start of the year, l. Column #6 shows the total number of years that the 100,000 persons lived in this year and subsequent years, L. Its computation is more complicated because it starts at the bottom of the table (last row of the table) and you work your way to the top (the row of year of age 0). The value of T for the last row is set at 100,000 since all the group will in the end have lived through all the years. Working bottom up, the subsequent values of T are computed by subtracting L from the value of T in the row immediately below. Column #7 shows the life expectancy at that year, e. It is the average number of years of life remaining at the start of the interval. It is equal to the result of dividing T by l.
INTERPRETATION OF LIFE EXPECTANCY AT BIRTH
Life expectancy at year of age zero has a special significance. It is called the expectation of life at birth. Life expectancy at birth is a sensitive indicator of the life of the community. It is the single indicator of current death rates. Life expectancy at birth is higher for females than males. IMR is the single most important determinant of life expectancy at birth. Rich industrialized countries have higher life expectancies at birth because of lower IMR. Life expectancy at birth is lower than that at age 1 because of the heavy toll due to IMR in the first year of life. It is possible to compute the conditional probability of survival from one year of age to another by merely dividing the number of survivors at the start of the year of the former by the number of survivors at the start of the year of the latter.
Learning Objectives
· Health economics
· Public Health: Policy, Planning, Finance, and Delivery
Key Words and Terms
· Analysis, cost benefit analysis
· Analysis, cost effectiveness analysis
· Analysis, cost utility analysis
· Care, long-term care
· Care, managed care
· Community diagnosis
· Cos t, cost of event averted
· Cost, opportunity cost
· Cos t, out of pocket costs
· Cos t, total cost
· Cos t, variable cost
· Costs, fixed costs
· Costs, ind irect costs
· Cos ts, direct costs
· Effectiveness
· Efficiency
· Efficacy
· Efficiency, productive efficiency
· Efficiency, allocative efficiency
· Elasticity, elasticity of demand
· Elasticity, price elasticity
· Equity
· Health care delivery system
· Health care financing
· Health care, hierarchy of health care
· Health facilities
· Health insurance
· Health maintenance organizations
· Health manpower
· Health planning
· Health policy
· Health promotion
· Health services administration
· Health status indicators
· HMO, IPA
· HMO, prepaid group practice
· HMO, prepaid practice Organization
· Marginal analysis
· Medical legislation
· Medically under-served
· Medicine, alternative medicine
· Medicine, biomedicine
· Medicine, complementary medicine
· Needs assessment
· Payment, per capita
· Payment, per diem
· Payment, third party
· Price, price discrimination
· Primary health care
· Public health administration
· Public health information system
· Public health law
· Quality assurance
· Quality improvement
· Quality, total quality management
· Reimbursement, fee for service
· Reimbursement, 3rd party
· Review, concurrent review
· Review, utilization review
· Risk pooling
· Scale, economies of scale
· Utility = satisfaction
· Utilization, bed-days
· Utilization, patient days
· Value of money, compounding
· Value of money, discounting
UNIT OUTLINE
6.4.1 HEALTH ECONOMICS
A. Terminology
B. Economic Concepts, Assumptions, and Tools
C. Controversies of Economic Analysis in Health
D. Measuring Costs
E. Measuring Health Outcome
F. Economic Analysis
G. Decision Analysis
E. Role of Epidemiology
B. Sources of Health Care Finance
E. Issues in Health Care Financing
A. Terminology and Concepts
B. Primary Health Care
C. Secondary and Tertiary Health Care (Hospital Care)
D. Program Evaluation
E. Quality Assurance in Health Care Delivery Services
6.4.1 HEALTH ECONOMICS
A. TERMINOLOGY
DISCIPLINE OF HEALTH ECONOMICS
Economics is a discipline that deals with scarcity of resources. It can be descriptive, explanatory or evaluative. Descriptive economics is used to describe medical care (such as number of physician visits or number of bed-days of hospitalization) or health status (such as morbidity, mortality, and functional capacity). Explanatory economics deals with demand and supply issues in health care, average and marginal costs of health interventions, and markets (competition and monopoly). Evaluative economics analyzes the allocation of healthcare resources in terms of efficiency, accessibility, equity, and fairness). Health economics is application of micro-economic tools to health. It studies supply and demand of health care services and their effects on the population. Measurement of quality and health care rationing are practical manifestations of the integration of medicine and economics. Economic appraisal is employment of economic tools to make allocation decisions. There are controversies whether it always works in health markets. Health care ethics are principles used to solve the conflict between healthcare and economics. Healthcare wants to maximize health benefit. Economics wants to minimize utilization of scarce resources of to use them in the most efficient way which translates in practice into restricted health care del ivery. Agency distinguishes the healthcare from other markets. Economic transactions in health have a special feature in that the physician acts as the agent of the patient.
ECONOMIC TERMINOLOGY
Efficiency is the ratio of input to output. It is also defined as cost per unit produced. High efficiency is achieved when a service is rendered with use of minimum resources. Inefficiency is a waste of resources. Effectiveness is ability to accomplish a defined task. It is the degree to which organizational goals and objectives are achieved. Efficacy, a measure of outcome, is the net impact of an intervention under ideal conditions. Efficacy assesses outcome under idealized conditions. An externality is said to occur when someone external to the market transaction (neither a buyer or a seller) is directly affected by the transaction. For example measles immunization has an externality in benefiting those not immunized by limiting spread of disease. Utility is economic jargon for satisfaction. A good is desired because it generates utility. If goods are of the same price the consumer will prefer those that provide higher utility. Utility like many economic measures is measured at the margin. Healthcare is subject to the law of diminishing marginal returns or law of diminishing marginal utility. Each additional unit of health care provides less marginal utility than the previous one. Utility can be positive or negative. Po sitive utility increases satisfaction whereas negative utility decreases satisfaction. Need is defined based on objective criteria. Usually professionals define the medical needs of a patient. In recent years the patient’s opinion has become increasingly involved in such discussions. Needs cannot always be met in full. Want is subjective. It is an ind ividual’s own assessment of health ‘wants’. Demand is the result of an individual seeking healthcare and converting a health want into a health demand. Supply is the response to the demand by suppling the goods or services demanded. Marginal values are measured at the margins of curves of demand, supply, benefit, and utility. Economists talk of marginal costs and marginal benefits. Marginal cost is the amount of money that an ind ividual is willing to pay for an additional unit of a commodity. Marginal benefit or marginal utility are benefits of satisfaction that an ind ividual gets from getting that extra unit of the commodity. In monopolistic markets, there is a gap between price (indicator of marginal benefit to buyers) and the marginal cost of production resulting in an inefficient level of production. Diagnosis related group (DRG) is a grouping of related medical procedures into one category for purposes of reimbursement under systems of managed care and cost control such as HMO and PPO. Service capacity is the production frontier of a medical facility. Utilization is amount of capacity actually used. Equity is a measure of justice or fairness but does not always translate into equality. Equity is easier to describe in competitive markets and is difficult to define easily in the special case of health care. It may be defined in different ways: equal health status, equal expenditure per capita, equal inputs per capita, equal access for equal need, equal utilization for equal need (horizontal equity), or geographical equity (distribution of resources such that no area is under-served). Achieving equity may be contradictory to achieving efficiency. Geographical equity is about geographical distribution of health care resources such that there are no underserved areas. One of the objectives of social insurance is to assure equity. Four approaches or theories of equity have been advanced: entitlement, utilitarianism, maximin, and equality. The entitlement theory argues that health care resources should be distributed according to the ability of individuals to pay for them. It argues individuals are entitled to what they have acquired justly; the rich are entitled to better health care because they can have their own personal resources. The utilitarian approach argues that health care resources should be distributed in such a way that the society as a whole benefits from the efficiency that results. The maximin approach argues that health care resources should be allocated in such a way as to maximize benefit to the least advantaged. Equality argues that everybody should be treated in the same way. Vertical equity is allocating the same health care resources to the same disease conditions or diagnoses without regard to who is the patient. Horizontal equity is allocating the same health care resources to different persons.
TIME VALUE OF MONEY
Present and future value: Economists compute present and future values of money when they want to make analyses that involve monetary comparisons such as cost-benefit analysis, cost-effectiveness analysis, and cost-utility analysis. This is because the purchasing value of money changes with time due to inflation. Another consideration is the return on investment. Money invested in a health project may have a lower return on investment than money invested in the normal money markets.
Compounding is used to compute the future value of money. The future value can be computed according to the formula FVn = PV (1 + i)n where FVn = future value at time n, PV = present value, and i = interest. If compounding is more frequent the above formula is modified as FVn = PV (1 + i/m)mn where m = number of periods in a year and n = number of years. The effective annual rate, EAR, is used to compare different compounding periods thus EAR = [1 + (1/m)]m – 1.0.
Discounting is used to compute the present value of money from the future value thus PV = FVn / (1 + i)n = FVn [ 1/(1+i)]n.
B. ECONOMIC CONCEPTS, ASSUMPTIONS, AND TOOLS
IMPORTANCE OF QUANTITATIVE ANALYSIS
Economics has moved from its humble beginnings as a social science to its position today as an empirical science using statistical techniques and models to analyse and build theorems. It therefore has a major contribution to make to the analysis of health care expenditures and advancing suggestions on more efficient utilization of the available resources. Services of health care organizations can be improved by application of quantitative methods that include epidemiological, statistical, and economic analytic techniques. Health economics represents an integration of medicine and economics in its concern for quality which is a medical objective and efficient allocation of health care resources which is an economic objective.
MOTIVATION FOR ECONOMIC ANALYSIS IN HEALTH
Rising expenditures in health care have forced economic analyses to understand the behavior of the medical care market and to find ways of controlling costs. The issue becomes more urgent when problems of quality and access persist despite increasing expenditure. Hospital expenditure is the biggest component of health care expenditure yet epidemiological data over the past century shows that preventive, socio-economic, environmental, and lifestyle factors have had a bigger impact on health than curative medicine. Three reasons have been advanced for the rising hospital expenditures: more utilization of health care services, use of higher quality and therefore more expensive services, and increase in the price of services. The three reasons may operate singly and with various degrees of synergy and antagonism.
MICROECONOMIC TOOLS IN HEALTH ECONOMICS
Health economics uses the following micro-economic tools and concepts: scarcity, production frontier, supply, demand, utility, elasticity, inputs, outputs, competition, and monopoly.
Scarcity, a major concept in economics, is the assumption is that resources are scarce and that any benefit accrued must involve a trade-off in other words of you want to get something you must be ready to give up something else.
Production: Economic activity involves production of goods or services. Scarcity imposes a production possibilities frontier (PPF) presented as a production possibility curve. There is a limit to production of goods because of limited of resources. A hospital or a factory has a production frontier beyond which it can not produce any more commodities. A producer aims at cost minimization in the production process so that he can be able to offer competitive prices on the market. He also aims at output maximization to maximize revenue. Economists analyze production using cost curves that either show variation of production with marginal cost or with average cost.. Units of healthcare are difficult to measure because they are intangible, are inseparable one from the other, and are of varying quality. Physician productivity is measured as patient visits, office visits, or billings. Quality measured in 3 dimensions: structure, process, and outcome. One of these dimensions may be of good quality whereas the other is not making it difficult to reach one judgment on the quality of healthcare overall. Production in health is measured as service utilization or as performance ind icators. The following are employed as indicators of health care utilization: percent of in-patient care, hospital admission rate, average length of stay, population per physician, number of physician consultations, dental visits, and pharmaceutical usage. Performance indicators include mean length of stay, number of procedures carried out, and staff per bed.
Demand and supply: Economists use supply and demand curves or functions in their analyses. Demand for and supply of goods are a function of price. Demand is willingness to pay for a given quantity of a product. The demand curve is a plot of price against quantity demanded and is an exponential curve because demand is inversely proportional to price. The supply curve is a plot of price against quantity supplied and is J curve because supply increases with increase in price. The intersection of demand and supply curves is an equilibrium point at which demand, supply, and price are optimum. This optimal point can be shifted by various economic factors such as health insurance, out of pocket expenses, and new technology. Health insurance coverage increases demand for and utilization of health care services. Out of pocket expenses decrease demand for health services. New medical technology may increase or decrease costs of medical care and therefore affect demand and supply. Other factors of demand are total income, time costs associated with seeking healthcare, tastes and preferences, and the initial health status.
Elasticity is a term that refers to response of one economic variable to another. Economists describe demand price elasticity and supply price elasticity. Price elasticity is changes in demand or supply as prices offered or paid change. Price elasticity is defined mathematically as percent change in demand / percent change in price. If demand is elastic it is responsive to prices. Demand for health services is largely inelastic with regard to prices.
Input-output: The health care scenario can be looked at as an input-output system. The inputs are manpower, materials, capital, and technology. The outputs are survival and well being. The input-output system has processes for conversion of inputs to outputs.
Competition: economists assume that under theoretically perfect market conditions of free competition, cost of goods to consumers as well as production efficiency will be optimal (pareto efficiency). Under pareto efficiency the situation is the best that can be and there are no longer and further mutually beneficial changes. In reality this perfection is an abstract that has relevance because real markets offer various degrees of competitiveness. The health market is usually far from pareto optimality. Medical tend to be monopolistic and less competitive. Medical markets cannot be competitive because of information asymmetry.
ANALYTIC TOOLS IN HEALTH ECONOMICS.
Economic modeling: Econometric analysis of health economics applies a wide range of statistical tools to economic problems. It translates concepts into models that can be tested using empirical data. The models form the basis for developing economic theories. There are controversies about the appropriateness of economic models in general economics and health economics. Economic models are used as metaphors for reality. The basic economic model is E = f(Y) where E = expenditure and Y= income. This relationship can be modeled by a regression equation as E = a + bY. In the field of health the relationship between income and expenditure may be more complicated than in open competitive markets making its reduction to a model difficult if not impossible.
Hypothesis testing: Formulation and testing of hypotheses following the scientific method is the core work of econometrists. The techniques of continuous data analysis, discrete data analysis, and regression are used to test hypotheses. Regression is used to model economic data such as demand curves, supply curves, and estimation of elasticities. Regression models are valid if they do not violate 5 assumptions: zero expectation of error, constant variance, zero x-y covariance, zero autocorrelation of y. The errors around the regression line being random are expected to have an expectation of zero. If the expectation is not zero, inference is made that systematic errors do exist. The term homoscedacity is used to refer to the constant variance of y for all values of x. For the model to be valid the covariance of x and y must be zero. Different values of y must be independent of one another i.e. no autocorrelation.
FREE COMPETITION
Economic analysis can be characterized by assumption of free competition, rationality, use of abstractions, use of marginal analysis
The theorems of economics apply to a situation of perfect/free competition. This situation occurs when there is free entry and exit from the market, perfect information for both the buyer and the seller, a homogenous product, numerous buyers and sellers, no control over the price by any buyer or seller, no significant externalities, no public goods, no natural monopoly, and market operation under conditions of uncertainty. These assumptions apply only partially to the medical market. There are barriers to entry of new physicians, restrictions on price setting, and regulations on building new hospitals. The medical profession places many barrie rs to entry of new physicians into the profession by use of licensure laws. This keeps the field restricted and protects physician incomes. The government interferes in price setting directly by legislation or through regulations associated with social and other types of health insurance coverage. Building of new hospitals requires government approval. A certificate of need based on local needs must be issued before a new hospital is approved.
Availability of information is not perfect in the health market partly because of restrictions on advertisement. There is information asymmetry between the care providers and the care consumers. Physicians are in control of medical transactions because the patients know so little and cannot in any way influence of challenge physician decisions.
There is a greater tendency to monopoly in the health market than in other markets. Health care institutions like hospitals and insurance companies are few in number which makes it easier for them to exercise monopoly or semi-monopoly powers. It is generally assumed that monopolistic markets restrict production and charge higher prices that markets that allow free competition. Health care services are not a homogemous commodity; they vary in quality and number even within the same local area. This makes competitive pricing difficult since the consumer is not able to compare different providers in a rational, reliable, and consistent way. The providers of health care services and insurance companies are in a very powerful position as far as setting prices is concerned. This upsets the operation of a free and competitive market leading to economic inefficiency.
There are more externalities to consider in the health market than in other markets. Unlike other markets the motivation in health care is not profit alone. Many services are provided by not for profit organizations and by the government. Some services are provided as public goods by legislative mandate.
The medical market operates under more certainty than other markets. The expected demand can be estimated fairly accurately from epidemiological data. The physician ind uces demand. The caregiver or the insurance change prices virtually at will. Free competition cannot operate in the presence of supplier-induced demand. The physician who provides care determines demand for health care demands as he advises his patients. He then proceeds to be a supplier that satisfies that demand. In other markets demand is generated by the consumer who is considered sovereign under no obligation or dependency on the supplier. The phenomenon of the physician acting as an agent for the patient in generating demand gives rise to the phenomenon of supplier-induced demand (SID). Under SID the physician supplies unnecessary care in his self-interest leading to an anomalous economic process in which supply determines demand. There is agreement that SID exists but there is disagreement on whether it leads to unnecessary demand.
RATIONAL & SOVEREIGN CONSUMER BEHAVIOR
Economics relies a lot on analyzing and predicting consumer behavior. Consumer behavior is determined by consumer preferences and budgetary constraints. Consumer preference is more difficult to measure and describe empirically because human motivations and actions are very complex being influenced by human psyche, culture, and life experiences. Economists generally believe that a consumer relies on informal marginal utility analysis in making decisions about preferences. Economists assume that consumers make rational decisions in their best interests when making decisions about their demands. They assume that consumers are able to rank their needs and nev er purposely make choices that make them worse off. They assume that consumers choose what will bring them the biggest benefit for the least cost.
ABSTRACTION
Economic analysis is abstract. It simplifies a complicated reality to enable understanding. The simplification is not necessarily a distortion but is an attempt to leave out details so that the forest can be seen and not the individual trees. It is however not always easy to apply the abstract conclusions to the practical reality.
MARGINAL ANALYSIS
Economic analysis uses models that analyze at the margin. They generally try to predict what changes would occur in one variable on a unit change in another variable. Marginal analysis enables understanding of trade-offs and opportunity costs. Conclusions are then built on these marginal changes. Economics talk of marginal cost, marginal revenue, marginal benefit, and marginal revenue.
C. CONTROVERSIES OF ECONOMIC ANALYSIS IN HEALTH
INTRODUCTION
Application of economic analytic tools to health generates the following controversies: scarcity v.s. mal-distribution, health as a commodity, supplier-induced demand, role of price in health, restricted competition, and government subsidies.
SCARCITY
The scarcity of resources that economists talk about is true in given time and space circumstances. The creator of the universe provided sufficient resources for humans. Scarcity is due to mal-distribution and not to absolute scarcity. Human history has also given the lie to the scarcity assumption. Humans have always been able to discover new food and energy resources when old ones seemed to be nearing exhaustion. Wood was the first source of energy used. Later fossil fuels were used to provide heat. Electric and nuclear energy were later discovered. It is conceivable that new sources of energy will be discovered to replace fossil fuels that are polluting the atmosphere. Malthusian economists had predicted a food crisis in the world because population was increasing geometrically and food production was increasing arithmetically. The prediction was proved wrong by new agricultural technology of the Green Revolution. Human population has increased tremendously with no signs of starvation at the global level. Cases of starvation or food shortage that have been observed are either due to civil disturbances of mal-distribution. Some societies have overabundance of food and suffer from diseases of over-nutrition whereas others have les food and suffer from diseases of under-nutrition.
HEALTH AS A COMMODITY
Conventional economic analysis considers health as a commodity. Health is a complicated entity, little understood, and impossible to measure accurately. Healthcare on the other hand can be considered a commodity under the ordinary economic principles of demand and supply. The ultimate objective is health but it cannot be measured easily. What can be measured is health care. The relation between health and healthcare is complicated and is non-linear. Health status and health outcome are not determined by health care alone. The historical record shows that decline in mortality starting in the 18th and 19th century Europe was not due to medical interventions but due to public health measures, improved environmental conditions, and improved sanitation. Change in life style is a major determinant of health status and health outcome. Socio-economic status is a determinant of health status independent of health care. Health is a capital good and an investment. It is not bought but is produced by the individual. This explains the big role of SES and education on health. The number of years of schooling is a determinant of health status through an unknown mechanism. Most probably schooling develops discipline in the child and youth to wake up every morning, go to school, sit in the class room and obey all the rules. Such a person has more self-discipline and is likely to take correct measures to protect and promote health.
It seems that medical care has had its impact on special segments of the population. Improved access has improved the health status of minority groups that were previously excluded. Infants and children mortality and morbidity have greatly reduced because of specific medical interventions.
Health is measured by morbidity, mortality, and disability.
CONSUMER SOVEREIGNTY
The assumption of a sovereign, rational consumer who makes decisions on demand in his/her best interests is very difficult to accept by someone living in society and seeing the folly of ordinary humans squander their resources on acquiring goods and services that do not benefit them but do cause them harm. The economists counter by saying that economic assumptions are amoral. Consumers may make decisions that satisfy their own inner needs unknown to those outside. These needs are rational and beneficial to the consumers at the time of making the decisions. Outsiders, in ignorance of the inner secrets in the consumer’s mind, may look at the decisions as irrational, of no benefit, and perhaps harmful. Economists talk about consumer sovereignty which is difficult to defend when we look at real life situations. Consumers are pressured into purchasing goods and services that are not in their best interests by commercial advertisements. Consumers cannot be sovereign in their decision making when they have imperfect information about the quality of goods or services bought. They also have imperfect information about the right prices. The counter argument of economists is that in an open and competitive market situation, the suppliers cannot collude to mislead the consumer. Rival advertisements and claims provide a lot of data about goods and services and help the consumer become better educated. The assumption of consumer sovereignty is challenged more in the health market because of the need for health. Need for health is a very strong drive different from demand for goods and services in the usual markets. Ordinary consumers can choose what to buy and what not buy according to their taste and resources. The sick on the other hand do not have such a choice in cases of serious disease. They are forced to buy health care at any price. Thus patients do not have consumer sovereignty in reality.
SUPPLIER-INDUCED DEMAND
The conventional analysis of demand in free competitive markets can not be transferred directly to health because in health the health care provider makes decisions or recommendations about demand and then wears another that to become a supplier. The term supplier-induced demand is used to refer to the situation in which the physician controls both demand and supply. The physician behaves very differently from a normal businessman because of the information and power gap between the physician and the patient that prevents normal business negotiations and transactions. Healthcare consumers have relatively less information about the product demanded than consumers of other products. Patients do not normally actively seek to acquire information because of the implicit trust in their physicians. Consumers also look towards physicians as powerful and authority figures whose opinions cannot be challenged. It is assumed that physicians as consumer agents will make rational decisions on behalf of the patient.
It is s better to analyze utilization of health care services instead of analyzing health care demands to avoid the trap of the supplier-induced demand (SID).
ROLE OF PRICE IN HEALTH
The demand for health is such a strong human drive related to survival that it cannot be modulated fully by price as are commodities in open competitive markets.
RESTRICTED COMPETION
The healthcare market has several restrictions that do not allow the forces of free market competition to operate and regulate prices.
GOVERNMENT INTERVENTION
Government intervention in the health market is posited on two theories: public interest and special group interest. Government intervenes when market mechanisms fail to distribute health care resources fairly. Government intervenes in the health care market to control monopolies, regulate care delivery, and insure public good. Government intervention takes several shapes: subsidy for hospitals, building government hospitals, regulation of drugs, mandated insurance coverage, taxation policies for hospitals and providers of equipment and supplies, public health regulations, and regulatory actions (fee and rate controls, quantity and capacity control through issue of certificates of need, and quality controls). Government subsidies as well as direct government involvement in healthcare delivery distort the healthcare market from the perfect competitive model. Government is in many countries the major health insurer.
D. MEASURING COSTS
PRODUCTION COSTS
Specific cost terminology is used in economic analysis. The components of cost may be described as direct costs or indirect costs. They may also be described as fixed costs, variable costs, and marginal costs. The direct health care cost of an intervention is the actual monetary expenditure for that intervention and includes the costs of test, drugs, supplies, rent, and equipment maintenance. The direct non-health care costs include patient transportation costs and patient time costs. The indirect costs are overhead expenditures. Total cost (TC) is the sum of fixed costs (FC) and variable costs per unit (VC) x number of units. Cos ts can be related to revenue and volume. At the brea k-even point, total cost is equal to total revenue. Fixed costs can be managed in three ways: changing them into variable costs such as renting instead of buying, decrease of staff, or increase in volume. Discounting of future health costs is used in economic analysis. It is often difficult to determine the appropriate discount rate.
INTERVENTION COSTS
Units of medical care are difficult to measure because they are intangible, are inseparable, and are of varying quality.
The concept of cost is different from that of opportunity cost. The concept of opportunity cost is used by economists in their analyses. It is based on the concept of scarcity and is akin to the saying that you can not eat your cake and have it at the same time. Scarcity of resources requires that getting something means foregoing something else. Opportunity cost of a resource is its total value in another use. For example when funds are used for vaccination they cannot be used for education. Thus the opportunity cost (or cost of lost opportunity) is a true refection of health costs. Also included among costs are family or home caregiver costs or cost of lost wages. The concept of opportunity cost is used in priority setting. It is a better measure of the true costs of an intervention that the dollar amounts paid for equipment, supplies, and labor.
D. MEASURING HEALTH OUTCOME
Health has two dimensions: duration of life and quality of life. Both must be considered in the assessment of health outcome. A prolonged life with low quality is not a highly desired goal. It may be better to have a shorter life of higher quality. Duration of life is easy to measure using life expectancy. Quality of life is a less tangible entity difficult to measure accurately and includes disability, functional status, and generic measures of health status. The following instruments are used to assess generic measures: the Sickness Impact Profile (SIP) has 136 items in 12 categories and takes 20-30 minutes to complete; the Short-form Questionnaire SF-36 has 36 items; the Quality of well being scale (QWB), the EuroQuol Quality of Life Scale, Nottingham Health Profile (NHP).
E. ECONOMIC ANALYSIS
INTRODUCTION
The purpose of economic analysis is to evaluate projects. There are 4 basic types of economic evaluation in public health: cost minimization, cost benefit analysis, cost effectiveness analysis, and cost utility. Cost minimization is the easiest of the economic evaluations because it uses monetary units directly to make a decision. It is choice of the least costly of 2 or more interventions that have the same effectiveness or outcome. Cos t benefit analysis is economic appraisal that addresses allocative efficiency. It compares marginal benefit to marginal cost. Cost effectiveness analysis addresses meeting given objectives at least cost. CEA minimizes costs and is the first stage of CBA. Interpretation must take into account general background data such as population, total employment, and gross domestic product.
COST BENEFIT ANALYSIS (CBA)
CBA compares monetary costs with monetary gains. It measures the costs of an intervention and the benefits of the intervention in the same monetary units. Nett benefit = (Total benefit + averted costs) – total costs = total benefit – (total cost + averted cost). If we take present value of money into consideration, Nett benefit = t=1∑t=T (Bj – Cj) / (1 + r)t where r = discount rate. The Cost-benefit criterion can be computed as {t=1∑t=T [Bj / (1 + r)t] }/ {t=1∑t=T [Cj / (1 + r)t] } CBA is a type of marginal analysis that compares increase in cost with increase in benefit. CBA is undertaken because there are no market mechanisms to determine when marginal cost = marginal benefit. It rests on the principle that society’s welfare will be improved if the benefits of a health intervention exceed its costs. The intervention is allowed to go ahead if benefits exceed costs. It is stopped if costs exceed benefits. CBA is necessary for making efficient allocation decisions because resources are limited.
CBA has three main problems: complete identification of all relevant costs and benefits, assigning monetary values to benefits and some of the costs, and determining the appropriate discount rate for projecting monetary values into the future. Health expenditures are mainly of three types: hospital expenditures (wages, equipment, and supplies), community expenditures, and others like ambulances. Benefits include the cost of life saved but no consensus has ever been reached on valuing human life. The future income stream is empirically used to measure value of life but it has many disadvantages. The reluctance of health professionals to put a monetary value to life has made them avoid CBA and resort to CEA and CUA.
CBA is used for evaluation (screening programs, alternative treatment procedures, and technology), selecting policy alternatives, medical research, and setting regulatory measures. CBA enables making a choice between two or more interventions with different outcomes and effectiveness. For example in the case of polio vaccination the cost of the vaccination program is compared to the amount of money saved by not hospitalizing and taking care of polio victims. It is however difficult to put a dollar value to intangibles like value of life, suffering, and quality of life. Cost benefit analysis addresses the issue of allocative efficiency that answers the questions: ‘is it worth achieving this goal?’ and ‘ are the costs higher than the opportunity costs?’.
It is not easy to assess the benefits of health intervention because they are not directly measurable and measures used may not be complete or objective. Three indirect measures are used for health benefits: the human capital approach, willingness to pay, and cost savings. The human capital approach assumes that a healthy individual is an economic asset and his economic productivity (earnings) can be measured and can be attributed to the health intervention. The willingness to pay uses the health consumer as the judge of the worth of a specific intervention. The amount of money that a consumer in a free market situation is willing to pay for a specific health intervention can be used as a measure of the benefit of that intervention. The cost savings approach computes the benefit of the intervention as the difference in health and other costs with intervention and without intervention.
In some cases the benefits may be delayed for years after program implementation. It is also difficult to allocate savings benefits between the consumer and the provider. Benefits today may have to be discounted in order to compare favorably with future benefits; the formulas used are not always exact. Benefits have also to be considered in view of the priority of the program. Some issues have higher priority than others.
COST EFFECTIVENESS ANALYSIS
Cost effectiveness analysis measures technical efficiency of an intervention. Costs are computed monetary terms and benefits are expressed in their natural units. CEA is used instead of CBA because it does not involve monetary evaluation of benefits which is very difficult. Cost effectiveness is computed as nett costs / benefits. The nett costs are the intervention costs + costs of side effects – direct medical costs saved. A ratio of cost to health effects is computed for each intervention. Cost effectiveness analysis enables a comparison of costs alternative disease control strategies with benefits of each alternative measured in the same units. For example for HBV we may compare the costs of three alternative approaches: no vaccination, universal vaccination, and vaccination preceded by screening. The cost of each alternative is computed and the cheapest is adopted. Cost effectiveness analysis evaluates benefits against an acceptable cost. The cost of saving a life-year is estimated. Most sophisticated analysis may involve adjusting for quality of life. Cost effectiveness analysis has several limitations that must be taken into account while making public health decisions. Cost data is difficult to obtain. Often charge data is used but it is not a good substitute because what is charged is either below or above the actual cost depending on market factors. The health effects of an intervention are not easy to measure accurately and because unquantifiable value judgments are involved. Life expectancy and QUALY are sometimes used but they are not considered perfect.
COST UTILITY ANALYSIS (CUA)
Cost utility analysis values benefits of health services in terms of utility. The most commonly used measure of utility is the number of years of life gained due to the health intervention. Cost utility analysis enables making a decision which of 2 or more interventions is better per cost unit when the outcome measure reflects the values and preferences of society. Since CUA is based on consumer preferences, it is of limited application due to its many value-laden assumptions. Cost utility analysis uses years of health as a measure of outcome. The most popular are: Years of Potential Life Lost (YPLL), and Healthy Years Equivalent (HYE). Quality of life is measured as years of healthy life (YHL), quality adjusted life years (QALY), and Disability adjusted life years (DALY).
F. DECISION ANALYSIS
Two criteria are used in making decisions: expected payoff and break-even analysis. The expected payoff is computed as the product of the probability of payoff and the payoff. Payoff is equivalent to profit. The best decision alternative is one with the best payoff. Break-even analysis seeks to identify a point at which the payoff of a certain intervention is equal to the payoff of non-intervention.
The US Public Health system has federal, state, local, and quasi government components. The Department of Health and Human Services (DHHS) is the largest federal department in terms of budget. It has 10 regional offices and 4 main divisions: the Public Health Service (PHS), The Health Care and Financing Agency (HCFA), the Social Security Administration (SSA), and the Administration for Children and families (ACF). PHS is involved in community health. It was started in 1798 as the Marine Hospital Service. It was renamed the PHS in 1912. PHS has 7 divisions in addition to the uniformed Public Health Service. The 7 divisions are the National Institutes of Health (NIH), the Food and Drug Administration (FDA), the Centers for Disease Control (CDC), The Health Research and Services Administration (HRSA), the Indian Health Service (IHS), The Substance Abuse and Mental Health Services Administration (SAMHSA), and the Agency for Toxic Substances and Disease Registry (ATSDR). HCFA oversees spending of Federal funds on health care including Medicaid and medicare. SSA administers three programs: social security, supplementary income, and Aid to Families with Dependent Children (AFDC) programs. Social security is funded by worker contributions and pays benefits to retired workers and their dependents. Supplementary Income is paid to the elderly and the disabled. The Administration for Children and Families (ACF) provides family assistance, refugee resettlement, and child support enforcement.
The state health department provides a link between a link between federal and local health authorities. State health agencies are concerned with health regulation, supervision of health professions, and have the following services: communicable disease control, chronic disease control, vital and health statistics, environmental health, health education & health promotion, health services, maternal and child health, occupational and industrial health, dental health, laboratory services, public health nursing, and veterinary public health.
Local health authorities act at county and city levels. They provide health services to their communities as mandated by state law.
Local Health Departments undertake activities in communicable disease control, environmental health, food hygiene, school health, preventive screening, and others. Communicable disease control includes immunizations, special clinics such as TB and STD, environmental sanitation, pest control, contact tracing, and disease surveillance. Environmental protection activities include waste disposal, noise abatement, nuisance control, monitoring pollution (air, water, and soil), and hazardous waste disposal. Food hygiene activities include inspection of food products, inspection of food processing, and restaurant inspection. Special clinic services include family planning, dental clinics, maternal and child health clinics, geriatric clinics, and sexually transmitted disease clinics. Preventive screening programs are undertaken for hypertension, breast cancer, cervical cancer, and chest x-rays for tuberculosis. The departments also supervise home care plans, disaster planning, and voluntary services. (Page 325 John M Last: Public Health and Human Ecology 2nd edition Prentice Hall International Inc ? year)
Quasi-government organizations include the Red Cross, the National Science Foundation, and the National Academy of Sciences. The American Red Cross founded in 1851 by Clara Burton is the official liaison of the federal government in disasters. It also makes contacts between members of the armed forces and their families. The National Science Foundation funds research. The National Academy of Sciences advises the government.
Non governmental agencies are Mothers Against Drunk Drivers, the American Cancer Society, the American Health Association, the American Lung Association,
United Way,
Local Way , Rockfeller Foundation, Ford Foundation, the Commonwealth Fund, the Kaiser Foundation, the Kellog Foundation, the Millbank Memorial Foundation, and other corporate foundations.
United Way,
Local Way
· Relation between economic conditions and health status indicators
· Emerging communicable diseases
· International travel and health
· Pan american health organization
· Public health in africa
GOALS AND OBJECTIVES OF INTERVENTION
A goal is a future event towards which the effort is directed. Objectives are steps taken in pursuit of the goal. Objectives are a more concrete way of stating the goal. The goals of public health intervention include: raising awareness of the health problem, increasing knowledge and health skills, change of attitudes and behavior, access to health care, reduction of risk and finally improved health status.
TYPES OF INTERVENTION
Public health intervention takes the following shapes: (a) behavioral modification (b) environmental control (c) legislation (d) social engineering (e) biological measures (f) screening for early detection and treatment of disease. Intervention can be in the form of micro intervention activities or macro intervention activities.
Behavioral modification: The objective of behavioral modification is largely achieved by empowering the individual to make positive decisions about health. The individual can be influenced by education (information and skills), persuasion (communication & social reinforcement), motivation (reward and punishment), and facilitation (access and availability). These four influences are affected to a large extent by public policies and the economic system. Behavior modification is used for smoking cessation, exercise, and dietary change. Traditional educational activities are printed material, class room instruction, and use of the mass media. Other communication activities are billboards, bulletin boards, films, flyers, direct mail, newsletters, pamphlets, posters, video, and audio tapes. Behavior can also be modified by changing the environment eg no smoking zones, healthy foods in restaurants and vending machines.
Environmental control: Intervention is possible against air pollution, water pollution, soil contamination, and food contamination.
Legislation: Legislation can be passed to improve health. Laws on health insurance can regulate private insurance companies and assure employer-provided insurance. Legislation can require health coverage for the poor and the uninsured. Special programs may be set up for example for migrant workers, nomads, and aborigines.
Social intervention: There is a strong link between socio-economic status and health. The experience of the some countries and some specific projects proved that social mobilisation can effect major improvements in health even with few resources. Social intervention could be by improving income or education. Overall improvement in health was experienced in : China , Cuba , Kenya (Machakos project), and India (Kerala state). The social intervention could also be for a specific disease as in the following projects: (a) The cardiovascular disease project in Finland (Keralia) (b) the Multiple Risk Factor Intervention Trial in the US , and the Onchocerciasis Control Project in West Africa . Social intervention requires community participation for success. A question may be raised how far can social controls go? What is the balance between individual responsibility and external control?.
Biological measures: The following are examples of biological interventions against disease: immunization, chemoprophyllaxis, immunoprophyllaxis, use of pesticides, taking supplementary vitamins.
Screening for early detection and treatment of disease: For example screening for breast and cervical cancer.
INTERVENTION IMPLEMENTATION
A pilot project is necessary to test and fine tune the intervention procedures. Implementation should be phased for best results.
EVALUATION OF INTERVENTION
The evaluation criteria must be set during the planning stage. Evaluation can be carried out during implementation. Summative evaluation is carried out at the end ot the project.
E. ROLE OF EPIDEMIOLOGY
Case control studies:
Expenditure on health is rising in all countries. Health expenditure constitutes a high proportion of GDP. Per capita health expenditure is also very high. There are in addition unrecorded health expenditures such as out-of-pocket payments and payments in-kind. The rising costs of health have forced a review of the issue of access. Managed care is managed costs. Rationing health care and prioritizing coverage have also been considered. The reasons for rising health care costs are higher demand for care, higher wages of health care workers, more sophisticated and expensive medical technology. The traditional cost containment strategies included insurance deductibles, co-payments, and exclusion of certain services from coverage. The new cost containment procedures are prospective payments based on DRGs and managed care. Managed care interferes with physician autonomy and decision making. It uses the following method to control costs: pre admission review and certification, emergency room review, concurrent reviews to justify keeping the patient in the hospital, discharge planning, second opinions, and gatekeepers to check referrals to specialists.
B. SOURCES OF HEALTH CARE FINANCE
INTRODUCTION
Several methods are used to pay for health services: direct physician payments, national insurance, social insurance, commercial insurance and third party payors.
Physician payments may be fee for service or may be on a capitation basis. The US is the only developed country without a national health insurance plan. Canada has provided a national health insurance since 1972 financed by federal and provincial governments. In this model, the government is the Third Party payor. All conditions are covered and the patient chooses the physician he/she wants. Social insurance covers poverty, old age, disability, health and unemployment. Medicaid and medicare in the US are forms of social insurance.
Insurance and third party payors rely on actuarial principles in pooling risk in order to set the premiums. They protect the consumer from a high risk of catastrophic medical bills. Payments by 3rd party payors may be fixed payments per capita irrespective of services rendered or may be variable payments according to services rendered. Insurers prefer to insure groups rather than individuals in order to benefit from the healthy worker effects. The epidemiological profile of the insured is used in the determination of the premium. National insurance is popular in some countries.
Some terms relating to insurance must be known. The premium is the monthly or annual payment in lieu of insurance coverage. Some insurance plans demand co-payments and deductibles in addition to the premium. Insurance coverage may increase demand.
METHODS OF REIMBURSEMENT IN THE US
Fee for service, direct payment by the patient, is now limited. The common methods are third party payers (Medicaid, Medicare, and private insurance), Health Maintenance Organization (HMO), preferred provider organization, PPO, and Neighborhood Health Centers. Medicaid and Medicare are forms of social insurance provided by by the US government. Private insurance may be commercial or may be group insurance for employees of large corporations. HMO pre-payment plans have several models: employment of salaried physicians working at HMO facilities, HMO contracting with a physician group, and HMO contracting with individual physicians. Under the Preferred Provider Organization (PPO) arrangement the patient chooses a physician who is paid a fee by a third party payer. Neighborhood health centers operate in poor areas.
Commercial private health insurance is an American invention but its origins cab be traced to ancient Chinese who paid their physician only when they were healthy. Payments were stopped as soon as the patron fell sick and would not be resumed until after recovery. Severe penalties were imposed when the patron died (Quoted in Epidemiology, Biostatistics and Preventive Medicine by James F Jekel, Joann G Elmore, and David L Katz (eds.) from Prussin JA and JC Woods Development of Health Insurance. Topics in Health Care Financing 2:1-12, 1975.). Insurance is a mechanism for spreading risk and cost among a large number of persons. Those with known higher risk pay a higher premium for example smokers, bad drivers, and those with pre-existing medical conditions. The health insurance policy has the following elements: premium to be paid, the deductible that is the minimum paid by the patient and above which coverage starts, co-insurance provisions, co-payment provisions, fixed indemnity that is the maximum amount payable by the insurer, and exclusion of specific health conditions from coverage. Very large organizations can be self insured by collecting premiums from their employees and paying for medical costs. Several types of insurance coverage are available: disability insurance, hospitalization insurance, major medical insurance, regular medical insurance, optical insurance, surgical insurance etc. There are in addition special types of insurance: medigap to cover out of pocket expenses, specific disease insurance, hospital indemnity insurance, and long term care insurance.
METHODS OF REIMBURSEMENT IN CANADA
The National Health Insurance in Canada reimburses the provider the full cost of care and is funded by the state.
METHODS OF REIMBURSEMENT IN EUROPE
Health care in the Netherlands is financed by a mixture of social and private insurance. The providers are mainly private physicians or hospitals. Health care is financed by social insurance with a mixture of private and public providers in Belgium , France , Germany , and Luxembourg . Health care is financed by taxation with mainly public providers in Denmark , Spain , Greece , Ireland , Italy , Portugal , and the UK .
The UK National Health Service was introduced in 1948. It is publicly funded from taxes. General practitioners work under contract with NHS. Denmark has a tax-sponsored national health care system. Sweden has a national health insurance. Norway provides free care for its citizens. Finland runs an obligatory health insurance. Netherlands and France have national health insurance.
Social insurance in Germany : The term insurance here is used in terms of protection and not in the actuarial sense. Health care is paid for by payroll taxes to which both the employer and employee contribute. Federal and local governments cover the unemployed or the unemployable. Those earning above a certain income level can opt for private insurance.
METHODS OF REIMBURSEMENT IN OTHER COUNTRIES
C. ISSUES IN HEALTH CARE FINANCING
Distributive justice:
· Access to health care: financial, cultural, distance barriers
· Affordability of health care
· Quality of health care
Allocation priorities
· Rural vs urban
· Curative vs preventive medicine
· Administrative costs vs actual care
A. TERMINOLOGY and CONCEPTS
DEFINITION
Definition a health care system: The health care system can be described as resources, organization, and management
GENERAL DESCRIPTION OF A HEALTH CARE SYSTEM: The health care system is described using the following attributes: availability, adequacy, accessibility, acceptability, appropriateness, assessibility, accountability, completeness, comprehensiveness, and continuity.
COMPONENTS OF A HEALTH CARE SYSTEM: A health care system consists of institutions, human resources, information systems, finance, management, and organization, environmental support, and service delivery.
FACTORS OF A HEALTH CARE SYSTEM: The factors that determine the nature of a health care system are demographic, cultural, political, social, and economic. The political ideology may be favor private, public, or mixed health care. Political ideology explains variations in health care systems. In capitalist countries like the US and the Philippines , the health system is entrepreneurial. The system is welfare oriented in Canada , Japan , Australia , and Peru . Some countries provide a comprehensive health system like UK , Scandinavian countries, and Sri Lanka . Cuba and China have socialist health systems. Available economic resources determine the type of system that the country or the community can afford.
MODES OF HEALTH CARE DELIVERY
Physician office: The solo physician practice is the traditional mode of health care delivery and has survived many changes in the organization and technology of medicine.
Health Maintenance Organization, HMO: HMO is an alternative to the fee-for-service model of health care delivery. It is a combination of health insurance and health care delivery. Participants pay a fixed amount for health care coverage (with some restrictions). The HMO can be not for profit or can be linked to a mother hospital. It may be owned by insurance companies, by physicians, by consumers, or by companies to take care of employee. An HMO enrollee has a limited choice of providers. The HMO makes money by emphasizing health promotion and primary care to decrease service demands. HMOs control costs because they are also the insurers. There are 4 organizational models of HMOs. In the staff model the physician is salaried and works on the premises of the HMO. In a group model a group of independent physicians contract to provide professional services for the HMO. In a network model 2 or more group practices contract to provide services to the HMO. In the independent practice association (IPA) model, individual physicians contract with the HMO to treat patients on the basis of a capitation fee per enrollee.
Preferred Provider Organizations: PPO is a compromise between the traditional fee-for-service and the HMO. It negotiates fixed prices for private physicians and charge participants a premium. Under the PPO arrangement, subscribers get care from a limited number of physicians and hospitals with whom the payor has contracted. The providers benefit from predictable income and payors benefit from lower payments.
Ambulatory Care Centers: These have lower costs and employ fewer staff. They are either emergency centers or surgical centers
HEALTH CARE PERSONNEL
Health care personnel are classified as independent providers, limited care providers, nurses, allied health professionals, and public health professionals. Independent providers are physicians practicing allopathic medicine in addition to practitioners of osteopathy, chiropractic, acupuncture, naturopathy, homeopathy, and naparopathy. Limited care providers provided a limited and very specialized range of services: dentists, optometrists, podiatrists, and psychologists. Nurses are LPN, LVN, and RN. Allied health professionals include dietitians, occupational therapists, radiographers, medical technologists, medical record keepers, medical laboratory technicians, and audiologists. Public health personnel are involved in preventive medicine activities such as health education, environmental sanitation, and occupational safety. Physician supply and distribution is affected by economic factors. Solution of problem of physician mal-distribution is not easy.
HEALTH CARE FACILITIES
Physician offices …. Hospitals are classified as community, specialist, or teaching hospitals. Some are full-service hospitals while others are limited service hospitals providing a narrower range of services. Some hospitals are classified as short stay whereas others are classified as long stay hospitals. As regards ownership, hospitals can be public, private, or voluntary non profit hospitals. For profit hospitals may be independent or part of a chain. Not for Profit hospitals may be community hospitals, church hospitals, charity hospitals, city, county or state hospitals, or federal government hospitals. Not for profit hospitals either provide for unmet need or take care of special interests. Nursing homes. Out-patient services (ambulatory services): These may be a solo physician, a physician partnership, a group practice, at HMO, at hospital. Emergency room services. Comprehensive health service programs. Health care maintenance organizations (HMO). Rehabilitation centers. Continuing care facilities
FINANCIAL MANAGEMENT
Cost: Costs may be fixed or variable. Regression analysis is used to separate fixed from variable costs. Costs may be direct or indirect. Cost and revenue centers must be identified. Cost allocation to various cost centers may not always be easy or straightforward.
Cash: Cash flow management is very important because cash is the king since it pays bills. Financial managers are risk averse. An organization may have to shut down if cash flow is negative. The basic accounting identity shown on the balance sheet is assets = liabilities + owner’s equity. Assets may be current or fixed. They may be tangible or non-tangible. The income statement shows profits, losses, revenues, and expenses. Net income = cash inflow – cash outflow. In accounting documentation debits are on the left-hand side and credits are on the right-hand side. Debits must equal credits. In accrual accounting revenue is recorded at the time that the service is performed and expenses are recorded at the time that they are incurred. Cash accounting records revenues and expenses at the time that cash is received or is paid out. Cash flow management is easier when accrual accounting is used.
Time value of money: Compounding is used to compute the future value of money. Discounting is used to compute the present value of money.
Budgeting: Budgeting can be bottom-up or top-down. Budgeting may be incremental or may be zero-based. The budget may be fixed or variable. An organization has a cash budget and a capital budget. It also has a revenue budget and an expense budget.
Valuing assets: The accounting book value of assets does not reflect market values.
Analysis of financial statements: Various ratios are used in the analysis of financial statements comparisons being made with industry standards. The following are the ratios commonly used: liquidity, activity, leverage, and profitability ratios.
B. PRIMARY HEALTH CARE
DEFINITION AND SCOPE
History: a conference held in Alma Ata in 1978 formulated the concept of primary health care under the slogan of health for all by the year 2000. From available data the slogan has not been fulfilled but many strides have been made in the past 20 years towards the target
Definition: Primary health care (PHC) was defined by the World Health Organization in 1978 as essential health services universally accessible to individuals and families in the community by means acceptable to them through their full participation and at a cost that the community and the country can afford. It forms an integral part both of the country's health system of which it is the nucleus and the overall social and economic development. WHO’s slogan ‘Health for All by 2000’.
Entry: Primary health care is the frontline or point of entry of an individual into the health care system. It is centered on the individual and not the organ system or disease. It is provided at physician offices, clinics, and other patient facilities.
Content of PHC: PHC is a comprehensive care for common diseases including prevention, screening, diagnosis, and treatment. The content of PHC varies from place to place. In rich and developed countries, PHC may be very sophisticated medical procedures. In poorer countries, it may be simple and rudimentary. What is important is to make sure that PHC is relevant to local needs and is affordable by the community concerned.
Community participation: The local community must participate in the formulation of PHC. PHC must conform to their culture and local circumstances.
PHC and equity
WHO declared that PHC rests on 8 elements: health education, food supply and proper nutrition, safe water and basic sanitation, maternal and child health services including family planning, immunization against major infectious diseases, prevention and control of locally endemic diseases, appropriate treatment of common diseases and injuries, and provision of essential drugs.
Health education are learning activities that enable individuals to voluntarily make decisions, change behavior in order to enhance health.
Health protection: Health protection includes accident prevention, occupational safety and health, environmental health, food and drug safety, and oral health
Preventive services: Preventive services include Maternal and Child Health, Screening, Clinical Preventive Services.
NEEDS ASSESSMENT
A distinction must be made among three closely related terms: need, want, and demand. Need is an elusive concept. It is conceived differently by the population and the planners. Need assessment is determining the gap between the ‘is’ and the ‘ought’. Service needs are real needs as assessed by health professionals. Service wants or service demands are perceived by the community. Data for needs assessment is of two kinds: service needs and service demands or wants. Data on service needs is provided from epidemiologic data, health data, health risk appraisals, information from stakeholders, and literature review. Data on service demands is obtained by consulting the target community, consulting opinion leaders, a survey of the target population, and observation of the target population. Needs assessment proceeds in 5 steps: determining present health status, assessing the environment, identifying and prioritizing existing programs, assessing service deficits in light of existing programs, dealing with the problems, and validating the needs.
COMMUNITY HEALTH PROFILE
Overall health status is assessed using mortality statistics, life expectancy, years of potential life lost (YPLL), Disability Adjusted Life years (DALY. P 187 Mackenzie Intr to Comm Health), and results of Nutritional and Health Surveys. The health profile should also be studied for specific segments. The health status of infants and children is assessed using the infant mortality rate (IMR) and 1-4 year mortality rate. The health status of adolescents and young adults (ages 15-24) is more difficult to assess using a few indices. Mortality in this age group is more among males and is due to homicide, suicide, and motor vehicle accidents. STDs are a major cause of morbidity. This age group has behaviors and life style choices that endanger good health: alcohol and drug abuse, tobacco use, fighting, and promiscuous sexual activity. The health status of adults is assessed using mortality. The leading causes of adult mortality are cancer, heart disease, stroke, injuries, liver disease, chronic lung disease, homicide, HIV, and diabetes mellitus. Health behaviors and lifestyle choices that impact good health negatively are smoking, poor nutrition, lack of exercise, alcohol, and neglect of screening for disease. The health status of seniors >65 years is assessed using morality and morbidity. Mortality is falling, life expectancy is increasing but issues remain about quality of life. The main cause of mortality are cancer, stroke, COPD, pneumonia, and influenza. Morbidity is due to chronic conditions (HT, OA, CHD, and DM) or impairments (hearing, cataracts, orthopedic). Alcohol, tobacco, and obesity are important behavioral problems.
THE DISADVANTAGED
Community diagnosis must be sensitive to the special needs of disadvantaged minorities that live in poverty. Poverty can be defined in absolute terms and in relative terms. Economic inequalities translate into health inequalities. The disadvantaged are one parent families, the unemployed, the sick or injured, the disabled or handicapped, the elderly, the immigrants, and racial or ethnic minorities.
ROLE OF EPIDEMIOLOGY IN PRIMARY HEALTH CARE
Epidemiologic methods are used in the assessment of subjective and objective aspects of health. Epidemiologic data on morbidity and mortality is used in planning and evaluatoing health programs. Epidemiologic studies, observational and experimental, are used to define specific etiological relations. Epidemiological tools are used in screening, investigation, and control of disease. Clinical epidemiology guides clinical practice. Epidemiological methods are used in health services to plan, implement, and evaluate health interventions. Evaluation consists of cost-benefit, cost-efficiency, and cost-effectiveness analyses.
C. SECONDARY and TERTIARY HEALTH CARE (HOSPITAL CARE)
Secondary care, inpatient or outpatient, is provided in institutions to treat those already ill in order to alleviate disease or injury. Tertiary care is very expensive and is technologically intensive. Considerations of economies of scale are involved in secondary and tertiary care to make sure that one institution serves as wide a region as is possible.
ROLE OF EPIDEMIOLOGY
D. PROGRAM EVALUATION
DEFINITION
Program evaluation is study of effectiveness, outcomes, efficiency, goals, and impact. Evaluation may be uses clinical assessment, health services research, health policy analysis, and epidemiology. Health plans are assessed based on the basis of accessibility, satisfaction, technical quality of care, efficiency and cost-effectiveness, and financial stability.
TERMINOLOGY
A project has a start and an end and is carried out only once. A project goes through the the phases of concept, A program is carried out repetitively with no start or end. Process evaluation: is evaluation of the processes involved in health care without reference to the output. Efficient processes normally lead to better output. There are cases in which the processes are efficient but the output is not as expected due to other factors that have to be taken into consideration. Outcome evaluation focuses on results. The following are used as outcome measures: mortality, morbidity, patient satisfaction, quality of life, degree of disability or dependency, and any other specific end-points. Evaluation may be described as qualitative evaluation or quantitative evaluation. Participatory evaluation has many advantages over non-participatory evaluation.. Performance criteria are set and records are analyzed to determine the proportion that meet the given the criteria. Data on performance is usually obtained from a sample. Indicators used in evaluation may be sentinel indicators or rate-based indicators. Both process indicators and outcome indicators may be financial, clinical, or organizational. Efficacy of drugs and treatment techniques can be tested in controlled experiments. These can be randomized trials, before and after comparisons, comparison of utilizers and non-utilizers, and case control studies. Effectiveness involves testing effects in real life. Efficiency is computation of the cost-benefit ratio. Access to care is a measure of how easy is it for patients to obtain care at a given facility. Use of services measures how many units of diagnostic or treatment services are used. Other terminology used in program evaluation is program intervention, program merits, program objectives, program activities, program outcomes, program impact, program expenses and costs.
COMPONENTS OF PROGRAM EVALUATION
Program evaluation starts with formulation of questions to be answered on demographics, activities, effectiveness, costs, and the general environment. The answers to questions provide the data on which the evaluation is based. Baseline and interim data may also be used in the evaluation. Question formulation is followed by setting standards, determining an evaluation methodology, collecting data, analyzing data, and preparing an evaluation report.
WHAT TO EVALUATE
Evaluation of health services in in three areas: structure, process, and outcome. Items to look at in structure include the physical plant, technical resources, human resources, lines of communication, and budget allocations. Process indicators include appointment systems, services provided, diagnostic tests done, preventive procedures, education classes, and referrals to other agencies. Items to look for in outcome evaluation are death and morbidity rates of various diseases and conditions, disabilities, discomfort and dissatisfaction of patients. (Page 332 John M Last: Public Health and Human Ecology 2nd edition Prentice Hall International Inc ? year).
EVALUATION STUDY DESIGN
Three study designs are used: pre and post assessment in one group, randomization into 2 or more groups, and use of a control group. Inclusion and exclusion criteria have to be defined.
DATA COL LECTION
Evaluation may be based on existing data or on freshly-collected data. Data for evaluation may be obtained from various sources: medical records, vital statistics, review of published and unpublished literature, surveys, tests of achievements, observations, interviews of patients and providers, physical examinations, and clinical scenarios. Medical records may be computer or paper medical records.
DATA ANALYSIS
Data scales used are qualitative (ordinal and nominal) and numerical (continuous and discrete). The student t test is used to analyze continuous data. The Chi-square is used to analyze discrete data. Meta -analysis can be used as super analysis of evaluation data.
Three methods are used for ind icator analysis: data trends, threshold, and guidelines. Data trends, increasing or decreasing, point to consistent changes. A threshold may be set for an indicator beyind which furth er investigation and action are called for. Guidelines may be set in such a way that specific actions are taken if ind icators reach certain levels.
ANALYSIS OF WAITING TIMES
Defining the problem: Waiting time is a waste of time to the patient. To the care provider it is an ind icator of demand for services. Too long lines and few care personnel is a problem. It cannot be solved simply by increasing the number of providers because demand fluctuates and there may be times when the care providers are idle. Each service system has a limited capacity and its performance cannot be improved unless more resources are added to the system. A balance must therefore be established between reducing waiting times and adding resources to the system. The traditional way of establishing the balance is to use an appointments system, and first in first out (FIFO) except in emergencies.
Description of waiting times: Waiting times are random events. ‘Channels’ refers to the number of lines and servers refers to number of care providers. Balking is when a patient refuses to enter a line that is too long. Reneging is when a patient stays in the line for some time and finding it too long decides to leave. Batching is when several people usually members of the same family enter at the same time. Jockeying is changing from one line to another that is thought to be shorter.
Queuing Theory uses the Poisson distribution to describe arrivals and the exponential distribution to describe service times. The exponential is the inverse of the Po isson. The Poisson distribution is represented as P(x) = λx e-λ / x! where P(x) = probability of exactly ‘x’ arrivals, x = actual number of arrivals in a specific time period, λ = mean arrival rate, and x! = x factorial. The exponential distribution is represented as P(t) = 1 – e-µt where P(t) = probability of serving a given number of patients in time t, t = service time, e = 2.71828 and µ = mean service rate. For a single server single line system (same as the appointments system), P0 = 1 – λ / µ, L = λ / (µ - λ), W = L/ λ, Lq = λ2, Wq = Lq / λ, Pw = λ/µ where λ =expected number of arrivals per time period (mean arrival rate) and µ = expected number of services possible per time period (mean service rate), and P0 = probability that server facility is idle, L = average number of patients in the system, W =average waiting time, Lq = average number of patients in the queue waiting for service, Wq = average time a patient spends in the queue waiting, Pw = probability that a patient must wait. For a multiple server system, P0 = 1 / [{n=0Σn=s-1 1/n! (λ/µ)n }+{(1/s!) (λ/µ)s (sµ/sµ-λ), Lq = P0 . [(λ/µ)s+1] / [s.s!{(sµ) – λ)/sµ}2, and Pw = P0 . {1/s!} {λ/µ}s {sµ/(sµ-λ)} where P0 = probability that all servers are idle, s = number of servers, Lq = average number of patients in the queue, and Pw = probability that an arriving unit must wait. Formulations are also available for multiple server single line systems and multiple server multiple line systems. A general rule in queuing theory is that on average the arrival rate must be less than the system service rate.
ANALYSIS OF CAPACITY
Excess capacity should be built into the system to respond to unexpected demands. Output is measured as service units, patient-days, clinic visits, and number of procedures. In some cases capacity cannot be measured accurately and the peak volume is used instead. The production frontier is a straight line if resources are plotted against volume of patients. The region below the line is production possibility and above the line is production impossibility. The capacity analysis model compares capacity with output. A change in resource mix may enhance the system’s efficiency without necessarily increasing the total resource outlay.
EVALUATION OF PROJECT IMPLEMENTATION
A project goes through the phases of concept, definition, implementation, and evaluation. A project may be simple or complex. Various methods of project management and control are also tools for evaluating the implementation. These include PERT, CPM, and GANTT charts. Program Evaluation Review Techniques (PERT) is a management support system or a management tool that enables a manager to evaluate and control a project. Before applying PERT the project must be broken up into its component parts, a process called work breakdown structure (WBS). PERT shows the start and end of each activity, the critical and non-essential activities, the immediate predecessor of each activity. Critical Path Method (CPM) is similar to PERT. The GANTT chart shows the status of various project activities. It however does not show the sequence of the activities. It is therefore difficult to tell which activities will precede which in the critical path.
Financial evaluation of projects: Financial evaluation of a project involves computing the present value, the internal rate of return, and the adjusted rate of return. Discounting is used to determine the present value of money. The discount rate (alternatively called the cost of capital) is not easy. The best approach is to use the rate from the return (also called opportunity cost) from alternative low-risk investments (such as bank accounts, treasury bills, treasury bonds, and treasury notes). The internal rate of return is the actual return of the project. The adjusted rate of return is obtained after adjusting for market financial rates. Annuity is even cash flow every year like mortgage income. Annuities are of two types, ordinary and annuity due. Ordinary annuity is paid at the end of the period. Annuity due is paid at the start of the period. . Cash flow is critical to organizational survival. Nett cash flow = cash inflows (revenues) – cash outflows (expenses). Cash flow may be negative or positive. Future cash flows must be discounted in comparison with the present value of money. Depreciation is an expense that does not involve cash. The payback period is the time during which the investment can be recouped.
ISSUES OF VALIDITY
Validity ind icates that the instrument measures what it is supposed to measure. Reliability assesses measurement error. A reliable instrument will have a low measurement error and repeated measurements will show stability. Validity is described as content validity, face validity, criterion validity, and construct validity. Face validity is based on appearances. Criterion validity is based on use of criteria either for predictive validity or for concurrent validity. Predictive validity is a measure of how well a certain criterion predicts the future. Concurrent validity is an assessment of how well a measure agrees or concurs with other measures. Construct validity is experimental demonstration of how well a measure distinguishes groups of people.
UTILIZATION REVIEW
Utilization review has three objectives: ascertainment that procedures are medically necessary, ascertainment that level of service intensity is appropriate, and ascertainment that the cost is appropriate. Review may be carried out pre-admission, during hospital stay (concurrent review), and retrospective review. The conflict between demand for higher quality and minimum costs arises in utilization review.
EVALUATION REPORT
The report must have an abstract or an executive summary. The contents of the report are the introduction, methods, results, discussion, and recommendations.
ROLE OF EPIDEMIOLOGY IN THE EVALUATION OF HEALTH CARE
Epidemiology provides basic data on which evaluations are based.
E. QUALITY ASSURANCE in HEALTH CARE DELIVERY SERVICES
OVER VIEW
Definition: Quality assurance and peer review are control tools in the health care industry. Assessing quality of medical care by structure and outcome. Quality assurance (QA) is formal and systematic identification, monitoring, and overcoming problems in health care delivery. Quality improvement (QI) is a management philosophy to improve organizational performance. Total Quality Management (TQM) is a participatory and systematic approach to planning and implementing continuous improvement in quality. The term audit is sometimes used to refer to quality review. Benchmarking is establishing targets based on leading performance indicators of the industry concerned. Quality is different from the perspective of the patient and that of the caregiver. The definition and measurement of quality are still a dilemma. Quality in consumer economics is easy to measure since it is based on consumer satisfaction. Quality in industry can be quantified easily and its determinants can be identified and can be incorporated into the worker training and service delivery systems. Quality in medicine is seen more as what is wrong and not what is right. Outcome measures of quality have the disadvantage that they are probabilistic with no consistent relation between health intervention and health outcome. The price of a health intervention can be measured but not its value. There is no consensus on what is appropriate intervention. Generally assessment of quality covers personnel, facilities, processes, and outcomes. An exact definition of quality of health services is elusive. It may be defined as maximizing patient well-being, improvement of life, or desired health outcomes. To avoid confusions, quality in any specific discussion must be defined empirically and contexually.
History: The concept has evolved through 3 stages: Quality control (QC) through Quality assurance (QA) to Quality improvement (QI). Wide-spread use of computers in hospitals has increased the need for and ability to carry out quality reviews since data is readily available. TQM started in ind ustry and was then applied to the medical field.
Purposes: QA is normally part of good clinical practice being a continuous monitoring tool to make sure that care given is up to expectation. It is also required in some situations of accreditation and even licensure
Principles: The 4 major principles of TQM are intrinsic motivation, review of systems (problems are in systems and not individuals, use of the scientific method (hence the involvement of epidemiology), and adult learning to change behavior in view of the findings of the quality process.
Common quality problems in health care: The common problems in medical quality are: insufficient knowledge, defect in the system, and different behavior and performance. Recording of the clinical data is the corner-stone of QA reviews. The QA reviewer can not attend all medical procedures and will have to rely on the records for evaluation. The records must be a faithful representation of what actually happened. Data problems are usually: incomplete data, inconsistent data, and data without record of time. Uniform reporting of data facilitates quality reviews.
Process of QA: Quality assurance involves planning, action, checking, action, and returning to planning. The processes of QA can be summarized by the mnemonic: FOCUS-PDCAE. Finding a process to improve. Organizing a team. Clarifying current knowledge. Understanding the process and causes of the problem. Selecting procedures to improve. Planning data collection & determining what data to collect. Data collection and analysis. Checking data to see opportunities for improvement. Acting to improve the process. Evaluation
QUALITY INDICATORS, CRITERIA and GUIDELINES
Quality ind icators are mortality, morbidity, patient satisfaction, and various rates. The indicators must be assessed for validity, reliability (precision), and acceptability because of random and systematic errors.
Consensus guidelines must be developed for each clinical situation to be a bench-mark against which clinical performance can be evaluated. Good Clinical Practice (GCP) is a set of guidelines that have been developed and they undergo continuous revision. They are not a universal prescription since each situation will have to be treated differently. Clinical protocols are developed for dealing with specific diagnostic categories of specific procedures. Nursing guidelines or standards
METHODS and PROCEDURES of QA REVIEW
Types of QA reviews: QA review may be concurrent or retrospective. Concurrent review occurs when the reviewer attends and directly observes health care delivery such as attending a ward round, an operation, or an out-patient clinic. Retrospective review normally depends on review of records or interview of patients and health care providers. Quality review may be discipline specific (e.g. surgery or obstetrics) or site specific (heart, and lung).
The QA reviewers: The QA reviewers may be independent clinical auditors from outside or may be part of the health care team assigned the special function of QA. In most cases QA review by a committee gives best results. Many institutions train nurses to be QA reviewers and they report to the institutional or departmental QA committee. Peer review is when a person of persons of equivalent professional status carry out the review.
What is reviewed: QA in hospitals centers around review of the patient charts. The following records are reviewed: physician notes, nursing notes, pharmacy records, dental records, etc. Additional documents may be reviewed as necessary. Other aspects reviewed are morbidity and mortality figures, waiting times, the ratio between primary and secondary care. Physician performance is assessed based on knowledge and skills, observation, an clinical audit.
Method of review: The aim of QA review is to ascertain compliance with the given guidelines. If a deviation is found, it is documented as well as its surrounding circumstances. It is discussed at the departmental QA committee. The committee will suggest actions to be taken to alleviate the deficiency and map out an implementation plan.
Follow-up: The QA review process is cyclical. The QA reviewers must follow up on the recommendations of the QA committee and ascertain that they have been followed.
QUALITY REVIEWS FOR CLINICAL TRIALS
QA in clinical trials is very important because mistakes have devastating effects ie can approve an ineffective or even dangerous drug for public use. QA in clinical trials involves data monitoring, data management, and data analysis. Problems of quality in clinical trials arise because of multi-center trials with each institution using different approaches
EPIDEMIOLOGICAL METHODS USED IN QA PROGRAMS
Epidemiology provides data and to provide comparison tools used in quality studies. It studies the impact of quality on health outcome by comparing rates (incidence, prevalence, and risk). It also deals with issues of validity and reliability in quality measurements. Data can be obtained from routinely collected data or from special studies (cohort, case control, and cross-sectional). For certain quality problems the usual methods of documenting a deficiency, discussing it, and suggesting solutions may not be suitable. Specific epidemiological studies are used to investigate the causative factors of the problem and to evaluate the impact of interventions. Case control, cross-sectional, cohort, randomized, and quasi experimental studies are used. A sampling plan is made. Variables to be investigated are selected. The reliability and validity of the instrument are determined. Data collection may be in person, by mail, or by telephone. Incidence, prevalence, odds ratio, and risk ration are epidemiological measures that can be used to describe QA phenomena
UNIT 6.5
Learning Objectives:
· Literature search
· Fallacies of numerical reasoning: Lack of/use of wrong denominator
· Misleading graphs: scale and omitting the origin
· Check-list for reading a scientific article
· Writing and publishing a biomedical article
Key Words and Terms:
· Meta Analysis
· Scientific Misconduct
· Old Boy Network
· Publication Bias
· Data retrieval
· Literature search
· Data retrieval
· Database searching
· Data Retrieval
UNIT OUTLINE
6.5.1 LITERATURE SEARCH
A. Documents
B. Sources of documents
C. Document retrieval
6.5.2 CRITICAL READING OF A JOURNAL ARTICLE
A. Title and abstract, and introduction
B. Introduction
C. Materials and methods
D. Results
E. Conclusions
6.5.3 ABUSE or MISUSE OF STATISTICS
A. Introduction
B. Principle of parsimony
C. D. Fallacies of numerical reasoning
6.5.4 SCIENTIFIC WRITING
A. Goals of Scientific Writing
B. Title, Abstract, and Introduction
C. Materials and Methods
A. Results.
E. Discussion and Conclusion
6.5.5 SCIENTIFIC PUBLISHING
6.5.1 LITERATURE SEARCH
A. DOCUMENTS
DEFINITION AND TYPES OF DOCUMENTS
A document is stored data in any form: paper, book, letter, message, image, e-mail, voice, and sound. Some documents are ephemeral but can still be retrieved for the brief time that they exist and are recoverable. Documents of medical importance are usually journal articles, books, technical reports, or theses.
DOCUMENT SURROGATES
A document surrogate is a brief extract of the original data that help in the retrieval of the whole document. Examples of document surrogates are: document identifiers, abstracts, extracts, reviews, indexes, matrix representations, term extraction, term association, lexical measures, and trigger phrases. An index must be exhaustive and user-specific. A matrix representation has columns representing terms and rows representing documents. Term extraction is identifying terms that are important in a document by their low frequency according to Ziff’s law that states that the rank of importance x frequency = constant. Term association is looking for a pair of terms that occur near one another like ‘information’ and ‘retrieval’. Lexical measures of term significance use specialized formulas. Trigger phrases are terms like table, figure, and conclusion. There are about 250-300 common grammatical words that account for 50% of any text such as the, of, and, to a, in etc. These have to be excluded from queries with little loss of efficiency. Stemming algorithms remove the ends of words and leave only the roots. The thesaurus can help in the retrieval because it gives synonyms and antonyms of words. Homographs are words that have the same spelling but different meanings. Homonyms are words with the same sound but different spellings.
DATA FORMATTING
Data may be formatted in tables of several types of databases (relational, hierarchical, and network). It may be unformatted such as images, sound, or electronic monitoring in the hospital. Formatted documents are easier to retrieve.
FILE STRUCTURES
Files may be sequential files, indexed files, tree structured files, and clustered files.
B. SOURCES OF DOCUMENTS
ON-LINE DATA-BASES
MEDLINE was established in 1971. Every year 400,000 articles from 3,700 journals are added and are indexed using medical subject headings (MESH). GRATEFUL MED is a query language used to search MEDLINE. PDQ is a data base about cancer
ON-LINE JOURNALS
BOOKS
TECHNICAL REPORTS
THESES AND DISSERTATIONS
MASS MEDIA
C. DOCUMENT RETRIEVAL
QUERIES FOR RETRIEVAL
Retrieval technology for formatted character documents is now quite sophisticated. Retrieval technology for images is still in its infancy. Queries are short documents used to retrieve larger documents by matching, mapping, or use of Boolean logic. Examples of Boolean logical connectors are AND, OR, Not etc. For example a query may be written to retrieve animals AND plants BUT not machines OR minerals. Queries may be in the form of natural language. They may also be written in probabilistic formulations. Fuzzy queries are becoming popular because they can retrieve documents where the more rigid queries fail. A data query can be in the form of a computer program that has terms or keywords used in the data retrieval process. Document retrieval is easier if authors use a controlled vocabulary.
RETRIEVAL BY MATCHING
In the most common form of retrieval, the query is matched to the document being sought. The matching must be significant enough to retrieve the right document. Determinations of what is significant must be made. Not all terms in a query are equally important. It may be necessary to give different terms different weightings. Filtering is used to limit the range of search for example limiting the search to certain years of publication or by language. Sometimes a false drop is made by picking a false document that matches the query. Retrieval can be carried out bu submitting a user profile which then acts as a query. The profile includes language, educational level, job, interests, and types of journals searched.
OTHER METHODS OF RETRIEVAL
Natural language processing uses syntactic or semantic analyses. In citation analysis the suer goes for documents cited in the footnotes. Use can be made of hypertext links made by the author to other documents. Image and sound processing can also be used.
EFFECTIVENESS OF RETRIEVAL
The following 2 x 2 table can be used to assess efficiency of retrieval
RETRIEVED? | ||
RELEVANT? | YES | NO |
YES | a | B |
NO | c | D |
The precision of the retrieval is evaluated as a/(a+c). Recall is evaluated as a/(a+b)
Fallout is evaluated as c/{N – (a+b)}. Generality is c+b/N. Other methods of assessing efficiency are: (a) use of the coverage ratio which is the proportion of documents known to the user that are actually retrieved. (b) The novelty ratio are relevant retrieved documents previously unknown to the user.
6.5.2 CRITICAL READING OF A JOURNAL ARTICLE
A. INTRODUCTION
The concepts and procedures of study analysis discussed in this section are needed for critical reading of scientific literature. There are many scientific journals many of which are peer-reviewed and they try to maintain the highest standards. There are however lapses from time to time that allow poorly-designed, poorly-analyzed, or poorly-reported studies to be published. The reader must be equipped with tools to be able to analyze their methodology and data analysis critically before accepting their conclusions. These tools are provided in this section. The common problems in published studies are found in the following aspects: incomplete documentation, design deficiencies, improper significance testing and interpretation.
B. TITLE, ABSTRACT, and INTRODUCTION
The title must be relevant to the body of the article. The abstract is very important since it shows the focus of the study. It gives summary information that can be used for preliminary assessment of (a) study design and analysis (b) significance of the conclusions. The decision whether to read further can be made after studying the abstract. The introduction states the reason for the study. It reviews previous studies in order to establish the need for the study and indicate its potential contribution. It gives a general background and a historical perspective to the study. It may also refer to the population to be studied. The hypothesis of the study may also be stated in the introduction. Having a prior hypothesis prevents the pitfalls of going on a fishing expedition.
C. MATERIALS AND METHODS
The materials and methods section covers the following: study subjects, study design, data collection, data analysis, and assessment of errors. Study subjects involves identification of the study population, identification of the sample, methodology of sampling, methods of subject selection, eligibility and exclusion criteria, exposure and outcome criteria, sample size, study power, and representativeness of the sample. The study design may be experimental or observational (follow-up, cross-sectional, case control). An assessment is made of whether the case definition is clear and rigorous and whether the groups are comparable. Data collection involves records, measurements, or interview. An assessment must be made of the accuracy of measurements. Data analysis may be descriptive or analytic. An assessment is made whether the appropriate scale and test were used. Data normality is also assessed since the statistical test is determined by normality. The following are common design deficiencies: study design not appropriate for the hypothesis tested, lack of a comparison group, selecting a control sample from a population different from that of the study sample, and the sample size not being big enough to answer the research questions raised.
D. RESULTS
The reporting of results is sometimes selective showing only favorable outcomes. Missing denominators and numbers that do not add up are common deficiencies. The following are checked about descriptive statistics. Tables must be labeled well and completely. Marginal totals (rows or columns) must equal the sum of the respective cells. The column totals must equal row totals. Row or column percentages must add up to 100%s. Numbers in the table must reconcile with the text. A check is made on numerical consistency: rounding, decimals, units. The following must be checked about statistical tests: is hypothesis testing using confidence intervals or p-value?, are both negative and positive findings presented?. The following pit-falls are common with the t-test: not stating the degrees of freedom, not stating the CI, use of the t-test for non-Gaussian data, and Multiple comparisons.
E. CONCLUSIONS/DISCUSSIONS
The conclusion highlights the major findings without repeating the results section. Consistency of conclusions with the data and hypothesis, was there extrapolation beyond the data?, what are the short-comings and limitations? The statistical conclusions are evaluated after assessment of errors (random & non-random) and assessment of bias (misclassification, selection, and confounding). The following are common pitfalls in significance testing & interpretation: use of wrong statistical test, drawing inappropriate conclusions for example significant findings may not be important or important findings may not be significant, wrong interpretation of the age effect by ignoring the cohort effect. A distinction must be made between precision (lack of random error : size and efficiency) and validity (lack of systematic error). Validity is internal & external. Internal validity is ensuring that the study carried out was accurate in its findings. Internal validity is achieved when the study is internally consistent and the results and conclusions reflect the data. External validity is generalizability ie how far can the findings of the present study be applicable to other situations. External validity is achieved by several independent studies showing the same result. The following are common mistakes: numerator without denominator, inappropriate denominator, a missing comparison group, inappropriate comparison, missing standardization for age, loss of information due to censoring (loss to follow-up), inappropriate tests for rates based on person-year, using mean +/- 2SD on non-normal data, the Berkson's fallacy, and multiple comparison or multiple significance
6.5.3 ABUSE or MISUSE OF STATISTICS
A. LYING WITH STATISTICS
Statistics are often abused. Benjamin Disraeli, the 19th century British Prime Minister, is credited by Mark Twain for making the statement: ‘There are three kinds of lies: lies, damned lies, and statistics’ to which Frederick Mosteller, Harvard Professor of Statistics replied: ‘It is easy to lie with statistics but it is easier to lie without them’ (Demetri Kantarelis: Essentials of Inferential Statistics’ McGraw-Hill NY 1996.). Figures never lie but liars figure. Statistics can be abused by selection of a favorable rate and ignoring unfavorable ones. This is done by 'playing' either with the numerator or the denominator. The numerator and the denominator can be made wider or narrower.
B. PRINCIPLE OF PARSIMONY
Occam’s razor is a very common expression. William of Occum (1285-1349) was an English philosopher who formulated the maxim ‘entia nonsunt multiplicanda praeter necessitatem’ which translates that the assumptions to explain a phenomenon must not be multiplied beyond necessity’ This law was called the law of parsimony by Sir William Hamilton in 1853. Karl Pearson called it the canon of economy in 1892.
C. POOR DESIGN AND ANALYSIS
Before any critical review of any scientific literature we must make sure that the underlying research followed proper methodology in study design, data collection, and data analysis. This makes it necessary that anyone involved in reading literature must have a minimum of epidemiologic and statistical literacy. The study must have a clearly articulated objective that if reflected directly in the study hypothesis or hypotheses. The definition of cases, non-cases, the exposed, the non-exposed, and the comparison groups must be clear, consistent, and objective. The exposure must be defined and the method of its measurement must be adequate. The period of follow-up in a cohort study must be sufficient for detecting the outcome of interest. The sample size must be adequate and the study must have enough power to detect the outcome of interest. The methodology of data collection must be well documented and must be appropriate for the study design. Response rates must be adequate and missing data must be minimized. Causes of bias (information bias, selection bias, and confounding bias) must be considered and must be accounted for. Results must be documented fully and accurately.
D. FALLACIES OF NUMERICAL REASONING
Information presented as numerical and scientific may be based on false or wrong numerical reasoning. Statistics can be abused by selection of a favorable rate and ignoring unfavorable ones. This is done by 'playing' either with the numerator or the denominator. The numerator and the denominator can be made wider or narrower. Other mistakes are: giving a numerator without a denominator, and using an inappropriate denominator. It is misleading to make comparative statements without specifying a comparison group or using an inappropriate comparison. Berkson's fallacy and multiple comparison or multiple significance are other causes of false reasoning.
E. MISTAKES IN PUBLISHED PAPERS
The following pit-falls are common with the t-test: not stating the degrees of freedom, not stating the CI, use of the t-test for non-Gaussian data. Going on a fishing expedition. This is prevented by having a prior hypothesis. Multiple comparisons, failure to standardize for age, failure to take into account information loss due to censoring, using the wrong statisti cal formula, confusing continuous and discrete scales, using mean +/- 2SD on non-normal data.
6.5.4 SCIENTIFIC WRITING
A. GOALS OF SCIENTIFIC WRITING
CLARITY
The goal of scientific writing is clarity. The essen tial information must be communicated effectively. This requires that the author does not lose the forest for the trees by providing a lot of details that do not communicate a clear and definitive message. Clear writing is not only a help for the reader but also helps the writer clarify his or her thinking. Scientific writing is in the first place communication with the self. It shows the writer whether what is communicated is clear to himself. A writer who has no clarity about any subject will definitely fail in communicating it to others.
SENTENCES
Short concise sentences: Writing a good sentence involves both good choice and good arrangements of words. Short sentences are preferred. A sentence is considered overloaded if it has more than 20 words. Each sentence should have only one idea or piece of information. Stringing several ideas together in a sentence is bad writing.
Use of personal pronouns: Personal pronouns like ‘I’ and ‘we’ can be used in the paper. They help avoid the use of passive sentences that make the writing weak.
Use of verbs: Subject-verb agreement is a common mistake. The verb and the subject it refers to must agree. It is wrong to write ‘the student prefer’. The correct form is ‘the student prefers’. Helping verbs should not be omitted. It is wrong to write ‘the skin was opened and the peritoneum exposed’. The correct way is to write ‘the skin was opened and the peritoneum was exposed’.
Active and passive sentences: Each sentence should be active and should have an action verb. Passive sentences should be avoided. The action verb should be a principal part of the sentence and not hidden in a phrase or a clause within the sentence. It is better to write ‘ the drug lowered blood pressure’ than to write ‘ the drug caused lowering of blood pressure’. Noun clusters of more than 2 nouns should be avoided. It is better to write ‘indicators of blood pressure’ than to write ‘blood pressure indicators’. An adjective should not be added to a noun cluster. It is better to write ‘chronic disease of the heart’ than to write ‘chronic heart disease’. Care must be taken to make sure that the pronouns refer to the right nouns. It the pronoun-noun relation may not be clear in a sentence with 2 or more nouns or pronouns leading to confusion in meanings. In a worst case scenario a pronoun in a sentence may refer to a missing or unknown noun.
Parallel ideas: Parallelism helps avoid repetition thus improving the flow of the writing. It is for example better to write ‘cases had high blood pressure but controls did not’ than to write ‘ cases had high blood pressure but controls did not have high blood pressure. Hower writers often make mistakes with parallel ideas. Confusions arise when the parallel ideas equal logic or importance are joined by ‘and’, ‘or’, or ‘but’. Confusions also arise with the use of paired conjunctions such as ‘both… and …’, either… or …, ‘neither … nor …’ and ‘not only … but ….’. Care must be taken to make sure that the 2 parallel ideas are similar in form. It is good write ‘blood pressure was assessed and pulse pressure was calculated’ and not ‘blood pressure was assessed and pulse pressure assessed’. The term ‘compared to’ is overused and if often misused. It is better to use terms like ‘higher’, ‘greater’, or ‘lower’ to compare parallel things. Para llel ideas being compared must be equal or similar. It is better to write ‘results of this study are similar to results of previous studies’ and not ‘results of this study are similar to previous studies. It is possible to write more than 2 parallel ideas in a sentence but this must be discouraged.
Use of parentheses: Sentences, clauses, or phrases can be inserted in the middle of another sentence as a parenthesis enclosed within commas. Care must be taken to make sure that the general flow of the writing is smooth and intelligible.
Expressions often misused: The term ‘compared with’ is often confusing and should be avoided; use .more than’ or ‘less than’. ‘Could not’ meaning ‘unable to’ is sometimes confused with ‘did not’ meaning the action was not carried out irrespective of ability. ‘Did not’ is not the same as ‘failed to’. The term ‘marked’ or ‘markedly’ are misused widely. They have no meaning unless some form of quantification is attached to them. The term ‘significantly’ is wrongly used instead of ‘significant’ to indicate statistical significance.
PARAGRAPHS
Topic sentence and message: A paragraph starts with a topic sentence that is an overview of the message contained in that paragraph. The topic sentence must be short and simple. Each paragraph should convey only one message. Key terms are repeated in the paragraph for clarity and accuracy. The sentences following the topic sentence provide details and support for the topic sentence.
Paragraph organization: All steps of the logic relating to the paragraph message should be presented in the right order with no missing steps. The following are alternative logical orders of paragraph organization: least to most important, most to least important, concise to the detailed, time chronological order, problem followed by solution, or solution followed by the problem.
Links and transitions: Link terms such as ‘which is’ should be used when moving from one group of ideas to another. Transitions ensure continuity in the paragraph. These may be words, phrases, clauses, or sentences. Examples of transitions are: ‘therefore’, ‘thus’, ‘in addition’, ‘in contrast’, ‘however’, ‘for example’, and ‘on the other hand’. Transitions can be placed within sentences or can be placed between sentences.
Consistency: There must be consistency in the order in which information is mentioned. If certain objects were mentioned in a certain order in the introduction, they must be mentioned in the same order all through the writing. The writer should maintain a consistent viewpoint all through the paper and not appear to be jumping from point to point.
Emphasis: A piece of scientific writing must have a specific message to convey. This requires that this message be emphasized and be given prominence. Emphasis can be achieved by putting important information in power positions (the start and the end), identifying important information with appropriate labeling and clearly stating that it is important and not leaving the implication to the reader, repetition of important facts, and labeling important information. De-emphasis is achieved by condensing, omitting or labeling less important information.
B. TITLE, ABSTRACT, and INTRODUCTION
TITLE
The purpose of the title is to identify the main topic or message of the paper so as to attract readers. The title of the paper should contain the following: independent variable(s), dependent variable(s), the study subjects or materials, and statement of the main message like ‘to study the effect of’ , ‘to determine’ etc. A good title is unambiguous, concise, and contains important words.
ABSTRACT
The abstract is an overview of the report with a few significant details. It should be written to be read by both those who read the full paper and those who do not read the full paper. Normally the abstract should not exceed 250 words. The abstract should mirror the sections of the paper: introduction, materials & methods, results, and discussion. The subsections of the abstract may be ind icated as sub-titles. They may alternatively be signaled as follows. The research question is signaled by ‘to determine’, ‘to test the hypothesis..’, . The results are signaled as ‘we found’. The answer is signaled as ‘we found’ or ‘we concluded’. The implication or conclusion is signaled as ‘the results suggest..’ or ‘we conclude..’. The present tense is used to state the research hypothesis and the answer. The past tense is used for the experiment. An abstract is accompanied by keywords that are used for ind exing.
INTRODUCTION
The introduction should be short. It should start with stating the research question or research hypothesis and then go to elaborate. The transition should be from the known to the unknown and from the big bird’s eyeview picture to the detail. The introduction should mention the type of study, the study subjects or materials (substances, animals, persons). In some cases the introduction may briefly mention the proposed experimental approach to answering the research question. Results should not be mentioned in the introduction. The introduction should state whether the work is new or original.
C. MATERIALS AND METHODS
The aim of the materials and methods section is to describe the experimental techniques in detail sufficient for another trained scientist to replicate the procedures. The order of presentation is different for animal and clinical studies. For animal studies the order is: materials and animals, preparation, study design, interventions, methods of measurement, calculations, and data analysis. For clinical studies the order is: study subjects, inclusion criteria, exclusion criteria, study design, interventions, methods of measurement, calculations, and data analysis (Mimi Zeiger. Essentials of Writing Biomedical Research Papers. McGraw Hill New York ? Year). Independent and dependent variables should be identified. Intermediate results can be put in the materials and methods section. Final results should be put only in the results section. Details of sample size determination should be provided. The following terms are often confused with one another: ‘measured’, ‘calculated’, ‘estimated’, and ‘determined’. Measurement is using instruments. Calculation deals with numbers and formulas. Estimation is used in two senses as an approximation in measurements or as computation of statistical parameters. Determination is a general term for getting to a conclusion by use of the 4 methods above. The term ‘study’ is generic and can be confused with experiment that refers to only some types of studies.
B. RESULTS.
The results section presents the findings of the procedures carried out in the methods section. It should be brief and to the point. A distinction must be made between results and data. Result refers to summary information obtained from data analysis. Results of hypothesis-based studies should be in the past tense. Data of descriptive studies should be in the present tense. Data is the actual numerical information often presented in a summarized form. The result is presented followed by presentation of supporting data. Data are presented in the form of tables and diagrams (figures, bar diagrams, graphs, pie-charts, maps etc). Presentation of numerical data in text should be kept to a minimum. Only results relevant to the research hypothesis should be presented. Both negative and positive results are presented. It is considered scientific fraud to present only those results that the author thinks favor a particular hypothesis. The results section is written in chronological order. The most important results are presented before the least important. Magnitude of change should be presented as a summary statistic such as percentage change instead of presenting the raw data. Summary statistics normally used as the mean, the median, and the proportion. The mean should be presented properly as mean +/- standard deviation or standard error of the mean (SD or SE) with units of measure indicated. Measures of effect are normally the chisquare and the t statistics. Actual p values should be given instead of indicating <0.05 or >0.05. When specifying the sample size the type of sample should be explained for example ‘the sample was 20 rats’ instead of the sample size was 20’. Emphasis can be put on some results and not others. Not all the data from the study need be reported. Citing data in the tex t takes less space but is more difficult to read. A topic sentence is used to give an overview. Important results are put first. Figures used to present results must have a strong visual impact and must be simple. The following types of figures are used: line graph, scattergram, bar graph, histogram, and the frequency polygon. The title of the figure should reflect its contents. It must be labeled correctly. Symbols must be defined. The names of variables and units of measurement must be labeled appropriately. Tables must be properly titled and column headings clearly indicated. Footnotes, subscripts, and superscripts can be used.
E. DISCUSSION AND CONCLUSION
The discussion section states the research hypothesis, answers it, and supports the answers using data from the current study and other studies. It provides reasons to show that the answer to the question is reasonable. It explores and explains possible sources of error and bias. It also identifies and explains differences between the study results and published results. As part of intellectual honesty it discusses the strengths as well as the weaknesses of the study and how they impact on the interpretation of the results. Issues of validity and precision are also addressed. Also discussed is whether the result is new and how important it is..
References are used to acknowledge information obtained from others. The references must be the most recent and most easily available on the subject. Review articles are better than original articles. They may be journal articles, books, Phd theses, abstracts of meetings, or conference proceedings. The reference should be put immediately after the relevant tex t. If there are several references in a sentence, cite each reference at the relevant point and do not wait to put all of them at the end of the sentence. References should be written using the Vancouver style which is: Author. Title. Journal Year; Volume (number): starting page – ending page.