0712P - THE TRANSITION FROM SAMPLE TO POPULATION EPIDEMIOLOGY

Professor Omar Hasan Kasule, Sr. University of Brunei Darussalam

ABSTRACT

This review is based on analysis of original research reports in one 2006 volume from each of three major epidemiology journals: The American Journal of Epidemiology, The International Journal of Epidemiology, and the European Journal of Epidemiology. A total of 149 research reports were included in the review. The pattern that emerged from the analysis was the tendency towards large epidemiological studies that utilize all available population-based data without resort to sampling. The tendency was to use data in existing data bases instead of field data collection. Developments in information technology enabled linkage between various data bases to extend the range of hypotheses that could be tested. The transition from sample epidemiology to population epidemiology had advantages and disadvantages. The main advantage was external validity (results of the study were applicable to the population). The main disadvantage was loss of internal validity that could be achieved in small studies with higher data quality and personal familiarity of the epidemiologist with the data. It is envisioned that in the future web-based data collection will be feasible. It will also be possible to use a wider range of data routinely collected online on citizens including credit card, shopping, and other financial transactions.

Key words: birth cohort, defined population, data linkage, large data set,  

Introduction

Epidemiological research is moving in several directions one of the most exciting being the transition from research based on population samples, using subjects counted in the tens or low hundreds, to the start of large population-based studies, using subjects counted in thousands and millions.

The preference for large studies was either motivated by editorial policy or was motivated by the fact that authors increasingly submitted large studies. The connection between the two motivations is undeniable. There were however small studies with subjects in the low 10s that got published because of their quality (1) but these were an endangered species.

A theoretical discussion can be made about what is the main driver of the new epidemiology?. Is it a desire for large studies (possible only with use of large data bases) or is it availability of large data bases (no need for sampling since the population data can be analyzed easily). My inclination is to the latter option because large studies are way above the minimum study size required for statistical validity.

Three epochs in the development of epidemiological research in relation to data collection can be identified. The pre-1950 epoch can be called sample epidemiology because studies were based on data collection from samples with the attempt being made to make the samples as small as was compatible with statistical validity. The number of subjects was in the tens or low hundreds which minimized the cost of epidemiological research. The second phase, 1950-1980, witnessed larger studies using cohorts and defined population groups that became increasingly easy to assemble because of developments in information technology. The third epoch starting in about 1980 witnessed the emergence of a brave new world of epidemiological research using large population and health data bases with the number of subjects counted in the hundreds of thousands. By the early 1980s information technology had developed to the level that epidemiologists could study the whole population without the need to sample or use specific cohorts. That was the birth of what I want to call population epidemiology. The transition from sample to population epidemiology, with serious practical and theoretical implications, has produced an arm chair epidemiologist who designs and analyzes large data studies using information from data bases many of them already online.

I am proud of having witnessed the birth of population epidemiology. I was in the generation of epidemiologists who in the early 1980s made the transition from using hand calculators to desk top personal computers for data analysis. The newly developed information technology led to far-reaching changes in the practice of epidemiology. Epidemiologists realized that basic socio-demographic and health-related data about the whole population was collected routinely and was stored unused in government and non-government electronic data bases. They also realized that the new information and communication technology could enable them identify and follow up research subjects as well as collect data from and/or about them without even meeting them physically. The ability to link various data-bases enabled assembling data on a single individual from several data bases and to carry out arm chair adhoc research. A new era for epidemiology had dawned.

Before the information age, we distinguished between the field epidemiologists (who collected and analyzed data) from the arm chair epidemiologists (who dabbled in theoretical epidemiology) and did not want to ‘dirty’ their hands with field data collection. Today arm chair epidemiologists collect and analyze data while sitting in their offices.

Methodology of the review

Original research reports that involved data collection and analysis were identified in volume 163 of the American Journal of Epidemiology, volume 35 of the International Journal of Epidemiology, and volume 21 of the European Journal of epidemiology. The following basic characteristics of each report were abstracted: type of study design (cross sectional, case control, and follow up), type of study population (defined group, general population, and ongoing study), type of data collection (new data collection, routinely collected data, previously collected data), and total study size. The mean number of subjects was computed for each grouping of research reports. The computations were carried out separately for studies below 100,000 and those above 100,000 subjects. Excluded from the computation of means were research reports based on large national populations like that of the US.

Statistical Results of the Review

The mean number of subjects in birth cohorts at enrolment was 13,614. Table #1 shows the mean number of research subjects according to study design and data collection methods for the rest of research reports. The data shows a tendency towards large studies above 1000 research subjects.

Sample Epidemiology

To understand the brave new world of population epidemiology we need to remind ourselves of the erstwhile sample epidemiology. In this review the word population is used in its true meaning of referring to a large number of humans and not in its statistical meaning that refers to a set of objects (humans, non-humans, or events) with a common observable characteristic or attribute.

Before the information age, epidemiologists made other researchers envious because they could get information easily from small samples and could make inferences about the general population at minimal expense. Sampling for survey research underwent a lot of change since it was first introduced in the closing years of the 19^th century. Sophisticated sampling methods and theories were developed to ensure that sample-based inferences reflected population reality. Statistical analytic techniques suitable for small samples (the student t test and Fisher’s exact tests) were developed for analysis of very small samples because large sample statistics did not give valid answers. The vision was to be able to reach valid inferences using hand calculators and from the smallest sample possible.

In the early period there was no alternative to small samples. Sophisticated data management and data analysis software capable of handling large data sets were not yet available. Epidemiologists preferred sample to population studies because data collection from a sample was logistically easier and financially more cost effective (the biggest bang from the least buck in terms of manpower, time, and money). Data from samples was considered more accurate and of higher quality because the epidemiologists had a smaller number of research subjects to work on and could have ‘personal’ knowledge of the research subjects and their data. This knowledge could enable epidemiologists spot inconsistencies and errors in the data. It could also enable them identify potential confounders more easily and realistically.

A sample was supposed to be a representative subset of the population but this might not be true in practice and disastrous conclusions could result as happened in the US presidential election of 1936 (2). Sampling started by defining a sampling frame which was enumeration of the population by sampling units (a technical term for individuals to be sampled). The arduous task was assembling the sampling frame; the actual sampling being thereafter relatively easier.

At the beginning simple random sampling was used when the population was approximately homogenous. It was realized that simple random sampling did not perform well in representing various sub-groups of a heterogeneous population. Stratified random sampling was developed to make sure that that the eventual sample correctly represented the population heterogeneity. In this type of sampling the population was divided into approximately homogenous groups and simple random sampling was carried out in each group separately with the samples derived being combined to make the study sample. Other techniques used to improve the practical logistics of random sampling were: sampling with unequal probabilities (if it was desired to over-represent one segment of the population), systematic sampling, cluster sampling, and multi-stage sampling. Development of computer technology and existence of data-bases on local and area wide networks make simple random sampling much easier because construction of sampling frames became easier and databases over long distances could be sampled and analyzed while sitting at one’s office desktop computer.

Probability theory enabled inferring sample data to target population. Probability theory also enabled assessment of precision and avoidance of bias in sample selection. If the sample was selected at random and if the assumptions of the central limit theorem held, sample data represented accurately the underlying population probabilistic events and sample distribution corresponded to the population probability mass function or the probability density function. Relationships found in samples were inferred to be the same as those in the population and sample data was used to predict population parameters. The validity of inferences based on samples was not questioned for a long time however a few doubts did surface for example extrapolation from sample to the general population was found to be unreliable empirically (3).

Concern about precision and bias was always a nagging problem in sample epidemiology for fear that public health decisions based on sample data might not reflect the reality in the population. Despite all measures taken to ensure that samples accurately represented population experience, epidemiologists were aware that sampling errors and sampling biases were inevitable. Statistical theory and practice therefore developed to characterize and measure the magnitude of sampling errors and sampling biases and thus be able to assess their impact on the conclusions from data analysis. The accuracy of estimators could be expressed as a function of sample size, population size, and probability characteristics. It was therefore obvious that the larger the sample the more precise were the estimates. The problem of precision was addressed by giving effect measures with 95% confidence intervals quoted around them to indicate the degree of precision. The larger the sample size, the narrower the confidence interval, and hence the higher the precision. There reached a point at which further gains in precision were not worth the expense of increasing sample size. Techniques for dealing with bias (confounding, misclassification, and selection biases) were developed to prevent bias at the design stage or cure it at the analysis stage.

Epidemiological studies based on birth cohorts

Birth cohorts were used over the past half-century to provide longitudinal and cross sectional information (at ‘sweeps’ carried out every few years). The primary motivation was mostly from governments that wanted to obtain data for formulating health policies (4, 5). The study of birth cohorts enabled understanding the natural history of morbidity as well as the longitudinal relationship between risk, disease, and health-related behaviors (4). They could have the advantages of being national in representation if recruited from the general population (5-6). They had the advantage of longitudinal data collection (5) which enabled linkage of childhood experience with adult disease outcomes (4). Data quality was high because of trained and experienced researchers (5) who worked on the same study for years. Comparisons among cohorts enabled studying secular changes in risk factors and disease outcomes as well as the relationship between the two. Usually cohorts generated more data than the investigators desired or could analyze. There were therefore a lot of data archives that could be mined by later researchers. Data from birth cohorts was increasingly available to other researchers (7) sometimes on online for free or at a fee (4).

Table #2 shows details of birth cohorts covered in this review. The cohorts were recruited as babies born in a certain week of a year in the whole country (4), a city (6), or a part of the country. The cohorts were followed up until adulthood.  Data was collected either from the whole cohort or from a sample (5). Collection from the whole cohort was preferable to sampling (8). Birth cohorts could also be animals for example a birth cohort of cattle was studied to investigate BSE (9).

Table #3 shows the range of information collected from birth cohorts. Data collection was more frequent in infancy and childhood but less frequent in adulthood (5). Data was collected by postal questionnaires (5), interviews by trained researchers (5) or by telephone. In many cases information was obtained directly from data bases of routinely collected administrative, vital, and health data. This was possible because of data linkage using unique identifying numbers enabled assembling data from population registers, disease registers, pharmacy records, hospital records, conscript data, and death registers (7, 10). Some cohorts were based solely on data linkage for example the Stockholm Birth Cohort Study of 1953. Linkage to parents’ data was also done (10). Record linkage also enabled tracing from anonymized records by matching certain variables (10-11).

The main cause of loss to follow up was change of address by participants who failed to notify the study administration of their new address (4). A second cause was refusal to participate at subsequent sweeps. Death was a minor but expected cause of loss to follow up. A few were lost due to emigration.

Recorded losses to follow up were small. In 2004, 16,078 members were traced; this represented  91% of the 17,634 recruited in 1958 in the British birth cohort (4). At age 53, 82.6% of the original 1946 British birth cohort was contacted and they provided information (5). The Stockholm Birth Cohort of 1953 had an attrition rate of only 4% (10) this being explained by the ability to trace persons using large data bases. The Aberdeen Study was able to trace 99% of the original cohort using government records (11). Losses to follow up due to refusal were also low. In the 1958 British birth cohort refusal rates were 7.1% at age 23, 11.1% at age 33, and 13.2% at age 42 (4). Follow up of children in the Southampton study was 95, 93, 86, and 81% at 6 months, 1 year, 2 years and 3 years respectively (6). Losses due to death in the 1946 British Birth Cohort were 8.7% at age 53 (5). Losses due to emigration were 8.6% (5) and to living abroad were 2.2% (5). Problems of attrition progressively lessened over the past 20 years because of availability of government or health insurance records  about citizens that enabled tracing those who had changes addresses. Some information about those lost to follow up could still be obtained from data bases such as those of health insurance (5), cancer registries (5), and population registers.

In the pre-1980 era fewer variables were collected because the work was manual and too much data could not be handled efficiently. Limited funding sources could also have contributed to limiting the amount of information collected. With availability of information technology and more funding as the value of cohort data was appreciated by funding sources, more data was collected. However not all of the data was collected directly from the cohort participants. Researchers had access to population and health data bases and using various forms of data linkage could obtain information on cohort participants.

Data collection over a long period spanning decades had its own problems. It was difficult to maintain consistency of the data for accurate longitudinal analysis because the type and may be the quality of data collected could change with time. The relevance of some forms of data could also vary with socio-demographic changes and development of biomedical knowledge. Over long periods of follow up of up to 50 years administrative and scientific responsibility for the cohort changed from one institution to another accompanied by changes in procedures (4). The coverage and objectives of the study could also change in response to new scientific knowledge or social and lifestyle changes in the community. In some cases cohorts were abandoned and some were revitalized later when funding became available and new interests developed (11).

The frequency and intensity of follow up varied according to availability of funding (8). Funding sources changed as interest in the cohort waned or grew (4). Funding agencies could develop fatigue in funding a study running over decades (8).

The impact of cohort studies on policy was profound (4). This is not surprising because this was their raison d’etre. They also influenced health knowledge and practice by their voluminous publications. As of 2006 a total of 900 publications issued out of the 1958 British Birth Cohort (4). As of 2006 a total of 8 books had been published from the 1946 British Birth Cohort (5). The Stockholm 1953 Birth Cohort Study generated more than 100 publications (10). The 1970 British Birth Cohort generated over 300 publications (12).

Epidemiological Studies Based On Defined Groups

Defined groups were used by epidemiologists to study disease consequences of specific exposures. Defined groups were opportunities of getting data from a captive population that was easy to reach. They were identified based on geographical / political units or a defining characteristic of relevance to health. Epidemiologic opportunism was used when participants in a previous study were identified as a defined group for new research (13-14)..

Many studies were based on groups defined on the basis of geography or institution. The Framingham Heart Study based on a middle class cohort in the town of Framingham in Massachusetts USA was one of the most famous geographical cohorts. The Mexico City Prospective Study involved following up 150,000 adult men and women aged 35 years to study risk factors of mortality (15). The Guangzhou Cohort Study followed adults and collected biological samples (16.). Several ways of assembling and studying cohorts were used. Some cohorts were assembled by linkage of databases (17). Some cohorts were recruited at a significant event such as entry into school (18). Geographically defined groups were often rural or urban communities (19-25). Disease outbreaks on isolated islands provided opportunities to study a whole community (26). The information obtained was useful for outbreak control and also for further analysis of other epidemiological hypotheses.

Military groups were studied because of good military record keeping. Studies were made of military recruits, conscripts, volunteers (27-29) and war veterans (30-31). Educational institutions were used because of ease of subject identification, access, and follow up. Research was carried out in schools (32-35) and universities (36). Civil servants were a very stable and a cooperative group (37) liked by researchers. Occupational groups with unique exposures were explored at low cost such as textile factories (38-39) and pesticide workers (40). Research was based on groups that experienced an event of health importance such as birth (41) or travel overseas to disease endemic areas (42). Studies were carried out on population groups with unique characteristics such as homosexuals (43), and members of HIV clinics (44).

Health facilities such as physician clinics provided a good opportunity for recruiting study subjects (45-46). Networks of general practitioners collaborated by providing research data on their patients (47-48). Some of this data was available in databases (49- 50). Data was also obtained from prenatal clinics (51-52) and obstetric practices (53). Expectant mothers provided a stable pool of subjects who could be observed over a period of time and whose children could be recruited into cohort studies. Research was also based on patients on the ward (54-55).

Health insurance organizations (56) and health maintenance organizations (57-58) recruited a large number of participants counted in the thousands and had records on them spanning a long period of time. They had a lot of routinely collected data that could be analyzed to test hypotheses about healthcare delivery systems. Health related data was obtained from hospital admission records (59-61), hospital discharge records (62), and other hospital data (63-64). Hospital medical record departments were a rich source of data that was not exploited because of missing and incomplete information. Use of medical records may need to be supplemented by interviews (65) to obtain the missing information. Biological specimens like blood were collected from visitors to health centers (66), hospitals (67), blood donation centers (68).

Completed or on-going cohort studies have been used as a convenient source of study subjects for new studies. This practice is becoming a regular feature of research (69-70). Recruitment of research subjects from other studies is facilitated by availability of data on socio-demographic and biomedical variables. Even more important is availability of contact information and familiarity of the subjects with being participants in research.

Epidemiological Studies Based On the General Population

Large data studies attempted to collect information from the general population. This process was very daunting in the past when the decennial census was the only population-wide data collection undertaken. With availability of extensive data bases on social, health, and demographic variables about whole cities, districts, or even nations, collection of data from the general population has become an armchair exercise. Population based research could be analysis of data from national health surveys (71). Such data was collected at great expense and was stored with minimal analysis. It was better for a researcher with a new hypothesis to analyze existing data than to go out to collect new data. Data covering several countries was obtained from international organizations such as the United Nations and the World Health Organization (72). Such data enabled study across many countries of death rates (73), cancer incidence rates (74), and morbidity rates (75). Case control studies had been touted as having the advantage of getting information using a few subjects counted in the tens but the new era witnessed population-based case control studies with thousands of subjects (76- 86).

Studies were based on registries of diseases such as stroke registers (87), cancer registers (88, 89, 90, 91, 92), myocardial infarction registers (93), and congenital anomalies registers (94). Prescription data bases (95) could be linked with other data bases to explore many interesting hypotheses. Subjects identified from the electoral roll (96) could be recruited into research projects.

It was a bureaucratic paradox that a lot of socio-demographic and health-related data (census, vital statistics, and routine healthcare data) was collected at great expense with limited benefit. The data was a mine of information that researchers should have used to learn about health and disease in populations. Only a few statistics were usually published for administrative purposes. There were however some attempts to make use of that data. Using vital statistics data, analyses were made of death records (97-104). Health data collected in the general population census contributed to public health (105). Disease notification and surveillance data was analyzed (106-111). With the ease of data access from data bases it was not surprising that one study might obtain data from more than one source for example data from vital statistics could be combined with data from a survey (112). Existing records of previously collected data were exploited with new analyses or repeat analyses using either new techniques or testing novel hypotheses. Analysis of historical data (113-114) provided information on disease and risk factor trends.

Data linkage became an increasingly dominant mode of research. It enabled studying causal relations while controlling for a wide range of potential confounding variables. Vaccination data was linked to hospitalization data (115-116). The population register was linked with the psychiatry register (117), the multiple sclerosis register (118), social insurance data base (119), and mammography screening data (120). Birth data was linked to mortality data (121-122) and health records (123). Census data was linked to mortality data (124-125). Military data was linked to occupational, hospital, and death data (126- 127) as well as to population data (128). Autopsy records were linked to police records (129). Reproductive outcome data was linked to occupational data (130).

Even the random sample had a renaissance with many publications mentioning population-based random samples that were often very large (131). This was because the logistics of data collection were easier with large population-based data bases that supplied the sampling frame (132). Random samples were taken from towns (133), and population registers (134-139), and schools (140). In the age of sophistication reports of convenience samples were published (141) showing that old habits die hard.

The environment became a subject of intense political interest and spawned many studies. Existence of continuous environmental monitoring systems contributed to large data epidemiology. Studies were based on linking routinely collected environmental data with routine health data (142-149).

Future frontiers

We can extrapolate into the future of population epidemiology. Web based data collection will become common. Other data sources like credit card data will be used. Video recording of signs will be possible by video cameras attached to personal or laptop computers. Small hand held laboratories may be mailed to people in their homes and they may put biological samples like urine or saliva for analysis with the results being transmitted online to the data center. Some of these ideas look like science fiction today but could well become the daily reality within a few years.

Using online data collection may open up new frontiers but few epidemiologists so far have experience of virtual world research. A feasibility study of web-base questionnaires was carried out in Sweden (150). It investigated differences in response between a group invited to answer a web based questionnaire and another group invited to answer a paper questionnaire. The web based questionnaire had a higher response. There were no significant differences in socio-demographic and health-related variables between the 2 groups of responders. The investigators concluded that web-based questionnaires were a feasible tool for data collection in large population based epidemiological studies.

TABLE #1: STATISTICAL RESULTS OF THE REVIEW: MEAN NUMBER OF SUBJECTS

		Mean (for no of subjects <100,000)		Mean (for no of subjects >100,000
		Defined groups	Population	Defined groups	Population
Cross sectional	Newly collected	2,627	10,275	6,240,130*	-
	Routinely collected	810	13,963	-	925,704
	Previously collected	1,245*	1,067	-	-
Case Control	Newly collected	1,840	1466		-
	Routinely collected	1,330	1628	1,194,357	-
	Previously collected	-	-	-	-
Non-birth Cohort	Newly collected	4,038	-	246,146*	-
	Routinely collected	28,293	31,164	1,299,177*	-
	Previously collected	56,214*	-	-	-
Randomized	Newly collected	3186*	-	-	-

* Based on a single research report

TABLE #2: STUDIES OF BIRTH COHORTS

Authors and Ref	Years	Place	Title	No of Subjects
1. Wadsworth et al. (5)	1946	UK	1946 National Birth Cohort	16,695
2. 1Leon. (11)	1950-	UK	The Aberdeen Children of the 1950s Study	12,150
3. Osler et al. (7)	1953-	Denmark	The Metropolit 1953 Danish Male Birth Cohort	12,270
4. Stenberg et al. (10)	1953-	Sweden	The Stockholm birth cohort of 1953	15,117
5. Power et al. (4)	1958-	UK	1958 British Birth Cohort	17,000
6. Elliott et al. (12)	1970-	UK	1970 British Birth Cohort	17,287
7. Victoria et al. (8)	1982	Brazil	The 1982 Pelotas (Brazil) Birth Cohort Study	5,914
8. Inskip et al.  (6)	1998-	UK	The Southampton Women’s Study	12,579
Mean				13,614

TABLE #3: DATA COLLECTED FROM BIRTH COHORTS AT VARIOUS PHASES OF THE LIFE CYCLE

Ante-natal: Socio-economic data, socio-demographic data, maternal smoking, maternal hypertension, Labor & delivery, ante-natal care.

Infancy: birth weight, perinatal morbidity, neonatal morbidity

Early childhood: nutrition, immunization, anthropometry, morbidity, development (physical and cognitive), education

Later childhood: morbidity, Behavior, Anthropometry, Vision, Psychological assessment, Development: cognitive, education

Adolescence: morbidity, behavior, anthropometry, vision, development, puberty, education

Young adulthood: morbidity, vision, psychology, anthropometry, smoking, alcohol, physical exercise, education, work

fertility, contraception, sexual practice, health KAP

Middle age: reproductive history, emotional problems, morbidity, nutrition, cardiovascular assessment, respiratory assessment, anthropometric assessment, cognitive assessment, mental health assessment: depression, midlife/menopausal issues, neurological assessment, hearing, life style: alcohol, smoking, drugs, religious practice; health seeking behavior: exercise; vision; hearing; work; partnerships

Others: environment, health services utilization

REFERENCES and NOTES

1. Colson P, Henry M, Motte A et al. Epidemiological and virological features of HBV infection in HIV-2 infected patients living in southeastern France. Eur J Epidemiol 2006; 21(8): 615-618

2.      The Literary Digest Post sent out 10 million ballots, 2.3 million of which were returned and they predicted that Alfred M. Landon would win. It turned out that the democrat Franklin Roosevelt won with a 62% majority. The mistake that the pollsters made was to select their sample from the telephone directory. The sample was therefore a good representative of the higher socio-economic class that would vote republican and not the whole US voting population because in those days telephone ownership was restricted to the rich.

3. Iwasaki M, Yamamoto S, Otani T et al. Generalizability of relative risk estimates from a well-defined population to a general population. Eur J Epidemiol 2006; 21: 253-262.
4. Power C and Elliot J. Cohort profile: 1958 British birth cohort (National child development study). Int J Epidemiol 2006; 35(1):34-41
5. Wadsworth M, Kuh D, Richards M et al. Cohort Profile: The 1946 National birth cohort (MRC national survey of health and development). Int J Epidemiol 2006; 35(1):49-54
6. Inskip HM, Godfrey KM, Robinson SM. et al. Cohort profile: the Southampton women’s survey. Int J Epidemiol 2006; 35(1):42-48.
7. Osler M, Lund R, Kriegbaum M et al. Cohort Profile: The Metropolit 1953 Danish male birth cohort. Int J Epidemiol 2006; 35(3): 541-545
8. Victoria CG and Barros FC. Cohort Profile: The 1982 Pelotas (Brazil) birth cohort study Int J Epidemiol 2006; 35(2): 237-242.
9. Bohning D and Greiner M. Evaluation of the cumulative evidence for freedom from BSE in birth cohorts. Eur J Epidemiol 2006; 21(1): 47-54.
10. Stenberg S-A and Vagero D. Cohort Profile: The Stockholm birth cohort of 1953. Int J Epidemiol 2006; 35(3):546-548.
11. Leon DA, Lawlor DA, Clark H et al. Cohort profile: the Aberdeen children of the 1950s study. Int J Epidemiol 2006; 35(3):549-552.
12. Elliott J and Shepherd P. Cohort Profile: 1970 British birth cohort (BCS70). Int J Epidemiol 2006; 35(4):836-843.
13. Ogard CG, Petersen J, Jorgensen T et al. Serum ionized calcium and cardiovascular disease in 45-years old men and women followed for 18 years. Eur J Epidemiol 2006; 21(2): 123-128.
14. Mallen CD, Peat G, Thomas E. et al. Is chronic musculoskeletal pain in adulthood related to factors at birth? A population-based case-control study of young adults. Eur J Epidemiol 2006; 21(3): 237-244.
15. Tapia-Conyer  R, Kuri-Morales P, Alegre-Diaz J et al. Cohort Profile: the Mexico city prospective study. Int J Epidemiol 2006; 35(2):243-249.
16. Jiang C, Thomas GN, Lam TH et al. Cohort Profile: The Guangzhou biobank cohort study, a Guangzhou-Hong Kong – Birmingham collaboration. Int J Epidemiol 2006; 35(4):844-852.
17. Hinkula M, Kauppila A, Nayha M et al. Cause-specific mortality of grand multiparous women in Finland. Am J Epidemiol 2006:163(4): 367-373.
18. Reed PL, Storr CL, Anthony JC. Drug Dependence Enviromics: job strain in the work environment and risk of becoming drug-dependent. Am J Epidemiol 2006:163(5): 404-411.
19. Mungala-Odera V, Meehan R, Njuguna P. et al. Prevalence and risk factors of neurological disability and impairment in children living in rural Kenya. Int J Epidemiol 2006; 35(3): 683-688.
20. Bursi F, Rocca WA, Killian JM et al. Heart disease and dementia: a population-based study. Am J Epidemiol 2006; 163(2):135-141.
21. Nguyen VB, Nguyen GK, Phung DC et al. Intra-familial transmission of Helicobacter pylori infection in children of households with multiple generations in Vietnam. Eur J Epidemiol 2006; 21(6): 459-464.
22. Samore MH, Lipsitch M, Alder SC et al.  Mechanisms by which antibiotics promote dissemination of resistant pneumococci in human populations. Am J Epidemiol 2006; 163(2):160-170.
23. Oishi Y, Kiyohara Y, Kubo M et al. The serum pepsinogen test as a predictor of gastric cancer: the Hisayama study. Am J Epidemiol 2006:163(7): 629-637.
24. Hirokawa K, Tsutusmi A, Kayaba K. Impacts of educational level and employment status on mortality for Japanese women and men: the Jichi Medical School cohort study. Eur J Epidemiol 2006; 21(9): 641-651.
25. FEur J Epidemiolr R, Hartvigsen J, Kyvik KO et al. The Funen neck and chest pain study: analyzing non-response bias by using national vital statistic data.  Eur J Epidemiol 2006; 21(3): 171-180.
26. Hyde TB, Dayan GH, Langdrik JR et al. Measles outbreak in the Republic of the Marshall Islands, 2003. Int J Epidemiol 2006; 35(2):299-306.
27. Magnusson PKE, Rasmussen F, Lawlor DA et al. Association of body mass index with suicide mortality: a prospective cohort study of more than one million men. Am J Epidemiol 2006; 163(1):1-8.
28. Hemmingsson T, Melin B, Allebeck P. et al. The association between cognitive ability measured at ages 18-20 and mortality during 30 years of follow-up – a prospective observational study among Swedish males born 1949-51. Int J Epidemiol 2006; 35(3): 665-669.
29. Tomaso H, Mooseder G, Al Dahouk S et al. Seroprevalence of anti-Yersinia antibodies in healthy Austrians. Eur J Epidemiol 2006; 21(1): 77-81
30. Blanchard MS, Eisen SA, Alpern R et al. Chronic multisymptom illness complex in Gulf War 1 Veterans 10 years later.  Am J Epidemiol 2006; 163(1):66-75.
31. Salamon R, Verret C, Jutand MA et al. Health consequences of the first Persian Gulf War on French troops. Int J Epidemiol 2006; 35(2):479-487.
32. Clark C, Martin R, van Kempen E, et al. Exposure-effect relations between aircraft and road traffic noise exposure at school and reading comprehension: The RANCH Project.  Am J Epidemiol 2006; 163(1):27-37.
33. Rauh MJ, Koepsell TD, Rivara FP et al. Epidemiology of musculoskeletal injuries among high school cross-country runners. Am J Epidemiol 2006; 163(2):151-159.
34. Foxman B, Gillespie B, Manning SD, et al. Incidence and duration of Group B Streptococcus by serotype among male and female college students living in a single dormitory. Am J Epidemiol 2006:163(6): 544-551.
35. Dundas R, Leyland AH, Macintyre S et al. Does the primary school attended influence self-reported health or its risk factors in later life? Int J Epidemiol 2006; 35(2):458-465.
36. Papadopoulos FC, Skalkidis I, Parkkari J et al. Doping use among tertiary education students in six developed countries. Eur J Epidemiol 2006; 21(4): 307-314.
37. Breeze E, Clarke R, Shipley MJ et al. Cause-specific mortality in old age in relation to body mass index in middle age and in old age: follow-up of the Whitehall cohort of male civil servants. Int J Epidemiol 2006; 35(1):169-178
38. Lee D-H, Ha M-H, Kam S et al. A Strong secular trend in serum gamma-glutamyltransferase from 1996 to 2003 among South Korean men. Am J Epidemiol 2006; 163(1):57-65.
39. Wernli KJ, Fitzgibbons ED, Ray RM, et al. Occupational risk factors for esophageal and stomach cancers among female textile workers in Shanghai, China. Am J Epidemiol 2006:163(8): 717-725.
40. Hoppin JA, Umbach DM, London SJ et al. Pesticides associated with wheeze among commercial pesticide applicators in the Agricultural Health Study. Am J Epidemiol 2006: 163(12): 1129-1137.
41. Lindsay L, Jackson LA, Savitz DA et al. Community influenza activity and risk of acute influenza-like illness episodes among healthy unvaccinated pregnant and postpartum women. Am J Epidemiol 2006:163(9): 838-848
42. Rickettts KD, Slaymaker E, Verlander NQ et al. What is the probability of successive cases of Legionnaire’s disease occurring in European hotels. Int J Epidemiol 2006; 35(2):354-360.
43. Cain EC, Cole SR, Chmiel JS et al. Effect of highly active antiretroviral therapy on multiple AIDS-defining illnesses among male HIV serovonverters. Am J Epidemiol 2006; 163(4):310-315.
44. Lucas GM, Griswold M, Gebo KA et al. Illicit drug use and HIV-1 disease progression: a longitudinal study in the era of highly active antiretroviral therapy. Am J Epidemiol 2006:163(5): 412-420.
45. Victor JC, Surdina TY, Suleimenova SZ et al.  Person-to-person transmission of Hepatitis A virus in an urban area of intermediate endemicity: implications for vaccination strategies. Am J Epidemiol 2006; 163(3): 204-210.
46. Spinelli R, Brandonisio O, Serio G et al. Intestinal parasites in healthy subjects in Albania. Eur J Epidemiol 2006; 21(2): 161-166.
47. Dunn KM, Jordan K, Croft PR. Characterizing the course of low back pain: a latent class analysis. Am J Epidemiol 2006:163(8): 754-761.
48. Thomas SL, Wheeler JG, Hall AJ.  Micronutrient intake and the risk of herpes zoster: a case-control study. Int J Epidemiol 2006; 35(2):307-314.
49. Aggarwal VR, McBeth J, Zakrzewska JM et al. The epidemiology of chronic syndromes that are frequently unexplained: do they have common associated factors?. Int J Epidemiol 2006; 35(2):468-476.
50. Kim J, Evans S, Smeeth L et al. Hormone replacement therapy and acute myocardial infarction: a large observational study exploring the influence of age. Int J Epidemiol 2006; 35(3): 731-737.
51. Kwon HL, Belanger K, Holford TR et al. Effect of fetal sex on airway lability in pregnant women with asthma.  Am J Epidemiol 2006; 163(3): 217- 221.
52. Jaddoe VWV, Mackenbach JP, Moll HA, et al. The Generation R study: design and cohort profile. Eur J Epidemiol 2006; 21(6): 475-484
53. Grosso LM, Triche EW, Belanger K, et al. Caffeine metabolites in umbilical cord blood, cytochrome P-450 1A2 activity, and intrauterine growth restriction. Am J Epidemiol 2006: 163(11): 1035-1041.
54. Sacerdote C, Fiorini L, Rosato R et al. Randomized controlled trial: effect of nutritional counseling in general practice.  Int J Epidemiol 2006; 35(2):409-415.
55. Leblebicioglu H, Yilmaz H, Tasova Y, et al. Characteristics and analysis of risk factors for mortality in infective endocarditis. Eur J Epidemiol 2006; 21(1): 25-31
56. Park HS, Song Y-M, Cho S-I. Obesity has a greater impact on cardiovascular mortality in younger men than in older men among non-smoking Koreans. Int J Epidemiol 2006; 35(1):181-187.
57. Jackson LA, Jackson ML, Nelson JC. et al. Evidence of bias in estimates of influenza vaccine effectiveness in seniors.  Int J Epidemiol 2006; 35(2):337-344
58. Jackson LA, Nelson JC, Benson P et al. Functional status is a confounder of the association of influenza vaccine and risk of all cause mortality in seniors. Int J Epidemiol 2006; 35(2):345-352.
59. Zubaid M, Thalib L, Suresh CG. Incidence of acute myocardial infarction during Islamic holiday seasons. Eur J Epidemiol 2006; 21(3): 191-196.
60. Gerlich M, Gschwend P, Uchtenhagen A et al. Prevalence of hepatitis and HIV infections and vaccination rates in patients entering the heroin-assisted treatment in Switzerland between 1994 and 2002. Eur J Epidemiol 2006; 21(7): 545-550.
61. Chatzipanagiotou S, Maria E, Constatina P, et al. Incidence of bacterial and viral enteric pathogens in children with gastroenteritis over a one year-period, in Attica, Greece. Eur J Epidemiol 2006; 21(8): 613-614.
62. de Pedro-Cuesta J, Bleda MJ, Rabano A. et al: Classification of surgical procedures for epidemiologic assessment of sporadic Creutzfeldt-Jakob Disease transmission by surgery. Eur J Epidemiol 2006; 21(8): 595-604
63. Eveillard M, Lancien E, deLassence A et al.  Impact of the reinforcement of a Methicillin-Resistant Staphylococcus aureus control programme: a 3-year evaluation by several indicators in a French University Hospital. Eur J Epidemiol 2006; 21(7): 551-558
64. Kourbatova EV, Leonard Jr MK, Romero J et al. Risk Factors for mortality among patients with extrapulmonary tuberculosis at an academic inner-city hospital in the US. Eur J Epidemiol 2006; 21(9):715-721
65. Blomgren KJ, Sundstrom A, Steineck G et al. Interviewer variability – quality aspects in a case-control study. Eur J Epidemiol 2006; 21(4): 267-278
66. Saetta AA, Michaloppoulos NV, Malamis G et al. Analysis of PRNP gene codon 129 polymorphism in the Greek population. Eur J Epidemiol 2006; 21(3): 211-216.
67. Minola E, Baldo V, Baldovin T et al. Intrafamilial transmission of hepatitis C virus infection. Eur J Epidemiol 2006; 21(4): 293-298.
68. Gurol E, Saban C, Oral O et al. Trends in hepatitis B and hepatitis C virus among blood donors over 16 years in Turkey. Eur J Epidemiol 2006; 21(4): 299-306.
69. Klungsoyr O, Nygard JF, Sorensen T et al. Cigarette smoking and incidence of first depressive episode: an 11-year, population-based follow-up Study. Am J Epidemiol 2006:163(5): 421-432.
70. Sowers M, Jannausch ML, Gross M et al. Performance-based physical functioning in African-American and Caucasian women at midlife: considering body composition, quadriceps strength, and knee osteoarthritis. Am J Epidemiol 2006: 163(10): 950-958.
71. Muntner P, DeSalvo KB, Wildman RP et al: Trends in the prevalence, awareness, treatment, and control of cardiovascular disease risk factors among noninstitutionalized patients with a history of myocardial infarction and stroke. Am J Epidemiol 2006: 163(10): 913-920.
72. Moore S.  Peripherality, income inequality, and life expectancy: revisiting the income inequality hypothesis. Int J Epidemiol 2006; 35(3): 623-632.
73. Kesteloot H. Differential evolution of mortality between Denmark and Scotland, period 1970 to 1999. Eur J Epidemiol 2006; 21(1): 3-14.
74. Boffetta P, Casging M, Brennan P. A geographical correlation study of the incidence of pancreatic and other cancers in Whites. Eur J Epidemiol 2006; 21(1): 39-46.
75. Svensson E, Moger TA, Tretli S et al. Frailty modeling of colorectal cancer incidence in Norway: Indications that individual heterogeneity in risk is related to birth cohort. Eur J Epidemiol 2006; 21(8): 587-593
76. Zhang J, Munger RG, West NA et al. Antioxidant intake and risk of osteoporotic hip fracture in Utah: an effect modified by smoking status. Am J Epidemiol 2006: 163(1): 9-17
77. Kasim K, Levallois P, Johnson KC et al.  Chlorination disinfection by-products in drinking water and the risk of adult leukemia in Canada. Am J Epidemiol 2006; 163(2):116-126 
78. Edwards CG, Schwartzbaum, Lonn S et al. Exposure to loud noise and the risk of acoustic neuroma. Am J Epidemiol 2006:163(4): 327-333
79. 79 Schuz J, Bohler E, Berg G, at al. Cellular phones, cordless phones, and the risks of glioma and meningioma (Interphone Study Group, Germany). Am J Epidemiol 2006:163(6): 512-520.
80. Xu WH, Xiang YB, Zheng W et al. Weight history and risk of endometrial cancer among Chinese women. Int J Epidemiol 2006; 35(1):159-166
81. Coen PG, Tully J, Stuart JM et al. Is it exposure to cigarette smoke or to smokers which increases the risk of meningococcal disease in teenagers? Int J Epidemiol 2006; 35(2):330-336 
82. Davis S, Day RW, Kopecky KJ et al.  Childhood leukemia in Belarus, Russia, and Ukraine following the Chernobyl power station accident: results from an international collaborative population-based case control study. Int J Epidemiol 2006; 35(2):386-396.
83. Kerr-Pontes LRS, Barreto ML, Evangelista CMN et al. Socioeconomic, environmental, and behavioral risk factors for leprosy in North-east Brazil: results of a case control study. Int J Epidemiol 2006; 35(4): 994-1000.
84. Okamoto K. Habitual green tea consumption and risk of an aneurismal rupture subarachnoid hemorrhage: a case-control study in Nagoya, Japan Eur J Epidemiol 2006; 21(5): 367-372
85. Portoles O, Sorli JV, Frances F et al. Effect of genetic variation in the leptin gene promoter and the leptin receptor gene on obesity risk in a population-based case-control study in Spain. Eur J Epidemiol 2006; 21(8): 605-612
86. Salameh PR, Waked M, Baldi I et al. Chronic bronchitis and pesticide exposure: a case-control study in Lebanon. Eur J Epidemiol 2006; 21(9): 681-688.
87. Kurth T, Walker AM, Glynn RJ et al. Results of multivariable logistic regression, propensity matching, propensity adjustment, and propensity-based weighting under conditions of nonuniform effect. Am J Epidemiol 2006; 163(3): 262-270.
88. Canchola AJ, Horn-Ross PL, Purdie DM.  Risk of second primary malignancies in women with papillary thyroid cancer. Am J Epidemiol 2006:163(6): 521-527.
89. Trentham-Dietz A, Nichols HB, Hampton JM et al.  Weight change and the risk of endometrial cancer. Int J Epidemiol 2006; 35(1):151-158.
90. Begg CB, Hummer AJ, Mujumdar U et al. A design for cancer case-control studies using only incident cases: experience with the GEM study of melanoma. Int J Epidemiol 2006; 35(3): 756-764.
91. McNally RJQ, Pearce MS, Parker L. Space-time clustering analyses of testicular cancer amongst 15-24-year-olds in Northern England. Eur J Epidemiol 2006; 21(2): 139-144.
92. Houben MPWA, Coebergh JWW, Birch JM et al. Space-time clustering of glioma cannot be attributed to specific histological subgroups. Eur J Epidemiol 2006; 21(3): 197-202.
93. Kark JD, Fink R, Adler B et al. The incidence of coronary heart disease among Palestininans and Israelis in Jerusalem. Int J Epidemiol 2006; 35(2):448-457.
94. Acs N, Banhidy F, Horvath-Puho E et al. Population-based case-control study of the common cold during pregnancy and congenital anomalies. Eur J Epidemiol 2006; 21(1): 65-76
95. Vegni FE, Wilkinson P. Patterns of respiratory drug use in the Lombardy region of Italy, 1995-1997.  Eur J Epidemiol 2006; 21(7): 537-544.
96. Kavanagh AM, Turrell G, Subramanian SV.  Does area-based social capital matter for the health of Australians? A multilevel analysis of self-rated health in Tasmania. Int J Epidemiol 2006; 35(3): 607-613.
97. 97 Greene SK, Ionides EL, Wilson M. Patterns of influenza-associated mortality among US elderly by geographic region and virus subtype, 1968-1998. Am J Epidemiol 2006:163(4): 316-326.
98. Polasek O. Did the 1991-1995 wars in the former Yugoslavia affect sex ratio at birth? Eur J Epidemiol 2006; 21(1): 61-64.
99. Pearce J and Dorling D. Increasing geographical inequalities in health in New Zealand, 1980-2001.  Int J Epidemiol 2006; 35(3):597-603.

100.  Chaix B, Rosvau M, Lynch J, et al. Disentangling contextual effects on cause-specific mortality in a longitudinal 23-year follow-up study: impact of population density or socioeconomic environment?.  Int J Epidemiol 2006; 35(3): 633-642.

101.   Sartorius B, Jacobsen H, Torner A et al. Description of  a new all cause mortality surveillance system in Sweden as a warning system using threshold detection algorithms. Eur J Epidemiol 2006; 21(3): 181-190.

102.   Borreli C, Mari-Dell’Olmo M, Rodriguez-Sanz M et al.  Socioeconomic position and excess mortality during the heat wave of 2003 in Bercelona. Eur J Epidemiol 2006; 21(9): 633)

103.   Harding S, Boroujerdi M, Santana P et al. Decline in, and lack of difference between, average birth weights among African and Portuguese babies in Portugal. Int J Epidemiol 2006; 35(2):270-276.

104.   Arntzen A, Samuelsen SO, Daltveit AK et al Post-neonatal mortality in Norway 1969-95: a cause-specific analysis. Int J Epidemiol 2006; 35(4): 1083-1089.

105.   Silviken A, Haldorsen T, Kvernmo S. Suicide among indigenous Sami in Arctic Norway, 1970-1998. Eur J Epidemiol 2006; 21(9): 707-713.

106.   de Greeff SC, Spanjaar L, Dankert J et al. Underreporting of meningococcal disease incidence in the Netherlands: Results from a capture-recapture analysis based on three registration sources with correction for false positive diagnoses. Eur J Epidemiol 2006; 21(4): 315-322.

107.   Montagna MT, Napoli C, Tato D et al. Clinical-environmental surveillance of legionellosis: an experience in Southern Italy. Eur J Epidemiol 2006; 21(4): 325-332.

108.   Link MW, Ahluwalia IB, Euler GL et al. Racial and ethnic disparities in influenza vaccination coverage among adults during the 2004-2005 Season. Am J Epidemiol 2006:163(6): 571-578.

109.   Massari V, Viboud C, Dorleans Y et al. Decline in HCV testing and compliance with guidelines of Sentinelles general practitioners, 1996-2002. Eur J Epidemiol 2006; 21(5): 397-406.

110.   van Everbroeck B, Michotte A, Sciot R et al. Increased incidence of sporadic Creutzfeldt-Jakob disease in age groups between 70 and 90 years in Belgium. Eur J Epidemiol 2006; 21(6): 443-448.

111.    Jensen OC. Injury risk at the work processes in fishing: a case referent study. Eur J Epidemiol 2006; 21(7): 521-528.

112.    Singh GK, Hiatt RA. Trends and disparities in socioeconomic and behavioural characteristics, life expectancy, and cause-specific mortality of native-born and foreign-born populations in the United States, 1979-2003. Int J Epidemiol 2006; 35(4): 903-918.

113.    Nishiura H. Epidemiology of a primary pneumonic plague in Kantoshu, Manchuria, from 1910 to 1911: statistical analysis of individual records collected by the Japanese Empire. Int J Epidemiol 2006; 35(4): 1059-1065.

114.    Sonnenberg A. Causes underlying the birth-cohort phenomenon of peptic ulcer: analysis of mortality data 1911-2000.  Int J Epidemiol 2006; 35(4): 1090-1096.

115.    Cameron JC, Walsh D, Finlayson AR et al. Oral Polio Vaccine and intussusception: A Data linkage study using records for vaccination and hospitalization. Am J Epidemiol 2006:163(6): 528-533.

116.    Emch M, Ali M, Park J-K, et al. Relationship between neighborhood-level killed oral cholera vaccine coverage and protective efficacy: evidence of herd immunity. Int J Epidemiol 2006; 35(4): 1044-1050.

117.    Pedersen CB, Mortensen PB. Are the cause(s) responsible for urban-rural differences in schizophrenia risk rooted in families or in individuals? Am J Epidemiol 2006; 163(11):971-978.

118.    Bager P, Nielsen NM, Bihrmann K, et al. Sibship characteristics and risk of multiple sclerosis: a nationwide cohort study in Denmark. Am J Epidemiol 2006: 163(12): 1112-1117.

119.    Artama M, Ritvanen A, Gissler M et al. Congenital structural anomalies in offspring of women with epilepsy - a population-based cohort study in Finland. Int J Epidemiol 2006; 35(2):280-287

120.   van Euler-Chelpin M. Olsen AH, Njor S et al. Women’s patterns of participation in mammography screening in Denmark. Eur J Epidemiol 2006; 21(3): 203-210.

121.    Yang Q, Wen SW, Smith GN et al. Maternal cigarette smoking and the risk of pregnancy-induced hypertension and eclampsia. Int J Epidemiol 2006; 35(2):288-293.

122.   Morley R, McCalman J, Carlin JB. Birth weight and coronary heart disease in a cohort born in 1857-1900 in Melbourne, Australia. Int J Epidemiol 2006; 35(4): 880-885.

123.   Frisch M, Pedersen B V, Wohlfart J et al. Reproductive patterns and non-Hodgkin lymphoma risk in Danish women and men. Eur J Epidemiol 2006; 21(9): 673-679.

124.   Singh GK, Siahpush M. Widening socioeconomic inequalities in US life expectancy, 1980-2000. Int J Epidemiol 2006; 35(4): 969-979.

125.   Blakely T, Atkinson J, Ivory V et al.  No association of neighborhood volunteerism with mortality in New Zealand. Int J Epidemiol 2006; 35(4): 981-988.

126.   Hemmingsson T and Lundberg I. Is the association between low job control and coronary heart disease confounded by risk factors measured in childhood and adolescence among Swedish males 40-53 years of age? Int J Epidemiol 2006; 35(3): 616-622.

127.    Yang G, Rao C, Ma J et al. Validation of verbal autopsy procedures for adult deaths in China. Int J Epidemiol 2006; 35(3): 741-747.

128.   Magnusson PKE, Rasmussen F, Gyllensten UB. Height at age 18 years is a strong predictor of attained education later in life: cohort study of over 950,000 Swedish males. Int J Epidemiol 2006; 35(3): 658-662.

129.   Christensen PB, Kringsholm B, Banner J et al. Surveillance of HIV and viral hepatitis by analysis of samples from drug related deaths. Eur J Epidemiol 2006; 21(5): 383-388.

130.   Mjoen et al: Linkage of reproductive outcome data with occupation data. Eur J Epidemiol 2006; 21(7): 529

131.    Fezeu L, Minkoulou E, Balkau B, et al. Association between socioeconomic status and adiposity in urban Cameroon.  Int J Epidemiol 2006; 35(1):105-111.

132.   Russel MB, Levi N, Saltyte-Benth J-S, et al. Tension-type headaches in adolescents and adults: a population-based study of 33,764 twins. Eur J Epidemiol 2006; 21(2): 153-160.

133.   Van der Pols, Xu C, Boyle GM et al. Expression of p53 tumor suppressor protein in sun-exposed skin and associations with sunscreen use and time spent outdoors: a community-based study. Am J Epidemiol 2006: 163(11): 982-988.

134.   Tikkinen KAO, Auvinen A, Huhtala H et al. Nocturia and obesity: a population-based study in Finland. Am J Epidemiol 2006: 163(11): 1003-1011.

135.   Aro P, Storskrubb T, Ronkainen J et al. Peptic Ulcer disease in a general adult population. The Kalixandra Study: a random population-based study. Am J Epidemiol 2006: 163(11):  1025-1033.

136.   Heyworth JS, Glonek G, Maynard EJ et al. Consumption of untreated tank rainwater and gastroenteritis among young children in South Australia. Int J Epidemiol 2006; 35(4): 1051-1058.

137.    Fan AZ, Russel M, Dorn J et al. Lifetime alcohol drinking pattern is related to the prevalence of metabolic syndrome. The Western New York Health Study (WNYHS) Eur J Epidemiol 2006; 21(2): 129-138.

138.   Lagerros YT, Mucci LA, Bellocco R et al. Validity and reliability of self-reported total energy expenditure using a novel instrument. Eur J Epidemiol 2006; 21(3): 227-236.

139.   Rathmann W, Haastert B, Giani G et al. Is Inflammation a causal chain between low socioeconomic status and type 2 diabetes? Results from the KORA survey 2000. Eur J Epidemiol 2006; 21(1): 55-60.  

140.   Kuehni CE, Strippoli M-P F, Zwahlen M et al. Association between reported exposure to road traffic and respiratory symptoms in children: evidence of bias. Int J Epidemiol 2006; 35(3): 779-786.

141.    Chiolero A, Gervasoni P-J, Rwebogora A et al. Difference in blood pressure readings with mercury and automated devices: impact on hypertension prevalence estimates in Dar es Salam, Tanzania. Eur J Epidemiol 2006; 21(6): 427-434.

142.   Medina-Ramon M, Zanobetti A, Schwartz J. The effect of ozone and PM₁₀ on hospital admissions for pneumonia and chronic obstructive pulmonary disease: a national multicity study. Am J Epidemiol 2006:163(6): 579-588.

143.   Zeka A, Zanobetti A, Schwartz J. individual-level modifiers of the effects of particulate matter on daily mortality. Am J Epidemiol 2006:163(9): 849-859.

144.   Jackson LE, Hilborn ED, Thomas JC. Towards landscape design guidelines for reducing Lyme disease risk. Int J Epidemiol 2006; 35(2):315-322.

145.   Cheng AC, Jacups SP, Gal D et al. Extreme weather events and environmental contamination are associated with case-clusters of melioidosis in the Northern Territory of Australia. Int J Epidemiol 2006; 35(2):323-329

146.   Muntoni S, Cocco P, Muntoni S et al. Nitrate in community water supplies and risk of childhood type 1 diabetes in Sardinia, Italy. Eur J Epidemiol 2006; 21(3): 245-247.

147.    Boldo E, Medina S, Le Tertre A et al. Apheis: Health impact assessment of long-term exposure to PM_2.5 in 23 European cities. Eur J Epidemiol 2006; 21(6): 449-458.

148.   Michele M, Alberto M, Liana S et al. Do environmental factors influence the occurrence of acute meningitis in industrialized countries? An epidemic of varying aetiology in Northern Italy. Eur J Epidemiol 2006; 21(6): 465-468.

149.   Villeneuve PJ, Chen LI, Stieb D et al. Associations between outdoor air pollution and emergency department visits for stroke in Edmonton, Canada. Eur J Epidemiol 2006; 21(9): 689-700.

150.   Ekman A, Dickman PW, Klint A, et al. Feasibility of using web-based questionnaires in large population-based epidemiological studies. Eur J Epidemiol 2006; 21(2): 103-112.