HOW TO TEACH BASICS IN OCCUPATIONAL EPIDEMIOLOGY

1
HOW TO TEACH BASICSINOCCUPATIONAL EPIDEMIOLOGY

• Guido F. MOENS
• Dept. Occupational Environm. Medicine, Catholic
  University of Leuven, Leuven (Belgium)
• Dept. Research Development, Occupational
  Health Services IDEWE, Leuven (Belgium)

2
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

3
1. TEACHING FRAMEWORK

• 1.1. Trainees/students
• Post-academic level students during the 2-year
  theoretical training in OM for MaMed graduates
  (former MD)
• Other groups during the training for agreed O.
  Hygienists
• Previous knowledge of epidemiology poor or
  none BASICS

4
1. TEACHING FRAMEWORK

• 1.2. Teacher
• Professional situation / expertise
• Epidemiologic researcher since 30 years
• University teacher in OE since 15 years
• Epidemiologist in OHS since 18 years
• Director RD in OHS since 6 years
• Experience with / access to practice-linked
  studies/data e.g.
• Standard registration data
• (collected during the periodical medical
  examinations)
• Surveys by OPh
• e.g. stress at work, MSD,

5
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

6
2. TEACHING OBJECTIVES

• Trainees should be able to
• Assess the role of E. for OM and for OM-practice
• Know the
• Objectives of OE
• Strengths and limitations in solving workplace
  health problems
• Key features of OE, distinguishing it from
  other fields of E.

7
2. TEACHING OBJECTIVES

• Objectives of OE
• Describe the occurrence characteristics of
  (work-related) health problems among workers
• Identify hazardous workplace exposures to prevent
  O. disease (examples of successful prevention)
• Provide information for
• Risk assessment
• Prevention in the general population e.g. health
  effect of low exp. levels of pesticides
• --gt estimation from exposed workers
• Assess the effect of workplace interventions
  e.g. replacement of asbestos by MMF

8
2. TEACHING OBJECTIVES

• Trainees should be able to
• Design and perform small OE-study
• Write a study protocol (draft)
• Discuss E. research questions with professionals
  and lay people from OH-practice
• Collect and analyse data properly
• Report OE-findings (communication)
• Interpret and communicate findings from
  OE-literature
• Cooperate with research centre in larger OE-study

9
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

10
3. TEACHING METHOD FORMAT

• 3.1. METHOD
• Practice-oriented / problem-based approach
• application of E. methods in a real setting
• practical problems encountered in setting
  up and performing an E. study
• interactive teaching

11
3. TEACHING METHOD FORMAT

• 3.1. METHOD
• Individual activity of students group activity
• e.g. drafting a study protocol
• (about problem from ones own practice)
• preparation of Master-dissertation
• e.g. critical reading of E-papers
• But, no time left for presentation/ discussion of
  protocols e.g with course coordinator with
  other participants with users (company, union)

12
3. TEACHING METHOD FORMAT

• 3.2. FORMAT
• Incorporated into one course with
• Medical statistics
• Medical data handling (informatics)
• Theoretical lectures and practical exercises
  (individual/group) spread over two years(13
  hours 13 hours 26 hours)
• Course notes
• (containing questions and answers)

13
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

14
4. TEACHING CONTENT

• 4.1. General introduction to E. methodology
• 4.2. Setting up an E. study
• Importance of the study protocol
• Design of a draft study protocolusing a
  practical example
• Prevalence study on low back pain among hospital
  workers

15
4. TEACHING CONTENT

• 4.3. Sources of error in E. studies
• Random error
• Bias - Selection bias
• - Confounding
• - Observation bias
• 4.4. Characteristics and design of questionnaires

16
4. TEACHING CONTENT

• 4.5. Examples of types of E. studies
• Short introduction of each study type
• Real world example for each study type using
  questions to be answered,
• from setting up the study, analyzing the
  data,to the interpretation of the findings.
• (partly based on the K.Steenland book
• Case studies in OE)

17
4. TEACHING CONTENT

• 4.5. Types of E. studies
• Prevalence study
• (Retrospective) cohort study
• Case-control study
• Disease clustering

18
4. TEACHING CONTENT

• 4.5. Types of E. studies
• 1. Prevalence study Low back pain among
  hospital workers(most common study type in OH
  practice)
• 2. (Retrospective) cohort study Risk of
  spontaneous abortion among VDT-users
• 3. Case-control studyOccupational exposure and
  end stage renal disease
• Calculation of association-, and
  impact-parameters,
• sources of bias and error, and the control for
  these
• are discussed and illustrated working through the
• examples.

19
4. TEACHING CONTENT

• 4.5. Types of E. studies
• 4. Disease clustering
• How to investigate a disease cluster using a
  stepwise investigation method
• A cluster is usually a trigger /
  motivation for a formal E. study
• To avoid expensive, large, unnecessary
  (and mostly negative) E. studies
• --gt FIRST
• Cluster investigation through 3 steps / 3
  questions on Outcome / Association / Exposure

20
4. TEACHING CONTENT

• 4.5. Types of E. studies
• 4. Disease clustering
• Cluster investigation through 3 steps / questions
  on Outcome / Association / Exposure
• Orientation
• Quantification
• Quantitative analysis (aggregated level
  control group)
• If positive
• Formal E. study design

21
4. TEACHING CONTENT

• 4. Disease clustering
• How to communicate about risk (notion of risk
  perception)
• Two examples
• Cluster of cancer cases among young nurses in a
  hospital ward
• Cluster of urogenital disorders among production
  workers exposed to chromium
• 4.6. Assessment of E. papers
• Critical assessment of two or three papers from
  peer reviewed journals ( research report) (to
  be prepared by students)

22
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

23
5. SPECIFIC CONTENT ISSUES

• 5.1. Political and ethical issues
• Controversies
• - Strengths and weaknesses of E. methodology
• - Role in causal/risk assessment
• - Role of social, economic, political
  influences (e.g. asbestos debate,
  EM-fields-debate)
• Warning
• (Undue) emphasis on criticism of publications
• --gt desillusion on the value of OE
• versus
• practical problems in setting up a study
  (writing of protocol)

24
5. SPECIFIC CONTENT ISSUES

• Practical problems in setting up a study
  (writing of protocol)
• Practical compromises
• No study can be perfect / definitive
• Contribution to the pool of knowledge
• BALANCE OF EVIDENCE
• Pressure from interested parties to- set up a
  study quickly, or- delay a study or action
  e.g. further studies needed ? delay
  of prevention
• Disclosing conflict of interest / sources of
  funding

25
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 1. Healthy Worker Effect (HWE) How to
  handle?
• 2. Exposure data Problem of historical
  exposure Markers of internal dose /biological
  effect 3. Confounding4. Specific study
  designs (see Content)5. Use of routine data

26
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

27
6. ASSESSMENT / EVALUATION / TIMETABLE

• 6.1. Assessment of the students
• Year 1 Classical examination (oral with
  written preparation)
• Year 2 Discussion with course coordinatoron a
  draft of a study protocol
• From 2006-2007 on also self teaching exercises
  in year 2

28
6. ASSESSMENT / EVALUATION / TIMETABLE

• 6.2. Assessment of the course content /
  format (quality assessment)
• Each yearanonymous questionnaire filled in
  by students on - content
• - format
• - course notes

29
Overview

• Teaching framework
• Teaching objectives
• Teaching method format
• Teaching content
• Specific content issues
• Political ethical issues
• Major methodological issues
• Evaluation / timetable
• Limitations

30
7. LIMITATIONS

• Very poor basic knowledge of E. or statistics
  among starters (MaMed)
• More time needed e.g.
• Public presentations / discussions e.g. of
  protocol
• More exercises

31
7. LIMITATIONS

• Co-ordination / timing with other subcourses
  (statistics, informatics) is not always easy in
  practice e.g. the statistics course has to be
  finished before the analysis aspects of E.
  studiese.g. discussion about the content of
  medical informatics subcourse

32
REFERENCES

• Olsen J, Saracci R, Trichopoulos D.
• Teaching epidemiology Chpt 11.
• IEA, New York Oxford Univ. Press, 2000.
• Steenland K.
• Case studies in occupational epidemiology.
• New York Oxford Univ. Press, 1993.

33
THANK YOU FOR YOUR ATTENTION!ANY QUESTIONS?
34
(No Transcript)
35
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 1. Healthy Worker Effect (HWE)
• type of selection bias in comparing workers
  outcome with general population outcome
  because
• Workers are healthier at hiring
• Selection of unhealthy workers out of the
  workforce.
• Mostly discussed in the context of cohort
  (follow-up) studies, but also present in other
  types (e.g. prevalence studies)

36
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• How to handle the HWE?
• Use of appropriate comparison groups(within the
  cohort internal comparison)
• but does not affect those lost to the workforce.
• Limitation of the extrapolation of the findings.

37
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 2. Exposure data
• Estimation of historical exposure difficult
• Personal information/ interview/ recordson work
  history, causes - Recall bias - Other
  forms of reporting bias
• Job exposure matrices (but also based on
  personal job information)
• Real life examples to show the problems involved

38
5. SPECIFIC CONTENT ISSUES

• 2. Exposure data
• Molecular E. methods markers of internal dose
  / biological effect
• but
• major limitations for historical exposure(only
  last weeks or months)
• Choice of most appropriate method
• varies from study to study

39
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 3. Confounding
• Could be present in all E. studies but the
  association with exposure and outcome must be
  strong in OE e.g. exposed (manual) workers
  usually will have the same distribution of
  lifestyle factors (e.g. smoking) as non-exposed
  (manual) workers
• Complex mixtures (e.g. welding fumes)
• --gt identification problem

40
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 4. Study designs
• Specific methodological issues and
  characteristics e.g. due to the
  OH-situation, cohort studies are more common
  (most occupational exposures are rare in general
  population).
• 5. Use of routine data
• Mortality data
• - Occupation is not reliably registered and
  coded on the Belgian death certificate
• - Only the last occupation
• Registers occupation is mostly not registered
• Thus standardization techniques are not in the
  course

41
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 6. Cohort studies
• Relatively frequent
• rare exposure historical personal hygiene
  records
• --gt retrospective / historical cohort studies
• Calculation of
• - Descriptive parameters
• Incidence risk
• Incidence rate (person-time)
• - Associative parameter
• Relative risk
• I risk ratio
• or I rate ratio
• - Impact parameter
• Attributable risk (etiologic fraction)

42
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 7. Case-control studies
• Nested within a cohort
• More efficient than analysis of the whole cohort
• Combination of
• Rare exposure (cohort)
• Rare outcomes (case-control)
• For - common exposure
• - dispersed groups (e.g. farmers)
• Population-based case-control studies

43
5. SPECIFIC CONTENT ISSUES

• 5.2. Major methodological issues
• 8. Prevalence studies
• Frequent use in practice
• (description of a health problem)
• Mostly for non-fatal O. disorders (asthma, MSD,
  )
• HWE is of concern (prevalent cases selected
  sample)

44
STEPS IN CLUSTER INVESTIGATION