7/31/2019 Occupational Epidemiology
1/33
M O H A M A D A Z F A R Z A I N U D D I NP U D 0 0 1 5 / 1 1
Occupational Epidemiology
7/31/2019 Occupational Epidemiology
2/33
List of Content
Introduction
Differences between Clinical & Epidemiologicalapproach
Types of study in OE Problems in OE
Benefits of OE
Conclusion References
7/31/2019 Occupational Epidemiology
3/33
Introduction
Occupational epidemiology involves the application ofepidemiologic methods to populations of workers.
Occupational epidemiologic studies may involve looking atworkers exposed to a variety of chemical, biological orphysical (e.g., noise, heat, radiation) agents to determine if the
exposures result in the risk of adverse health outcomes. Alternatively, epidemiologic studies may involve the
evaluation of workers with a common adverse health outcometo determine if an agent or set of agents may explain theirdisease.
There are currently no specific OSHA standards foroccupational epidemiology. However, a variety of hazards are addressed in specific
standards for OSHA access to employee medical records,recordkeeping, general industry, shipyard employment, andthe construction industry.
7/31/2019 Occupational Epidemiology
4/33
Every year, millions of worker throughout the worldcomplain of ill-health which is caused or aggravated
by work.
Occupational Epidemiology has an important rolenotably in:1) establishing the causes and determinants of this ill-health,
2) ensuring adequate recognition and quantification of (1)
3) determining appropriate occupational exposure limits.
7/31/2019 Occupational Epidemiology
5/33
However, occupational epidemiology may be of valuebeyond the worker and the workplace. E.g. bycontributing to the setting of exposure limits such asair quality guidelines for the population at large.
To gain full benefit from this resource it is importantto have an understanding of epidemiologic concepts,notablycausal associations, and to be aware of issues
which may influence apparent association, notablybias, confounding, and chance.
7/31/2019 Occupational Epidemiology
6/33
Drawing analogies between the clinicaland epidemiologic approach
The logic including data gathering, data processing,interpretation, intervention and hence change has alot in common between the clinical and theepidemiologic method, as illustrated in the followingtable:
7/31/2019 Occupational Epidemiology
7/33
Analogies between dealing with individual patientsand with populations in occupational epidemiology
Consulting with anindividual patient
Contending with a group ofworkers or other people
Process: Taking the individual'ssymptom history*Taking the individual'sexposure history (and gettinginformation about thework/environment)
Carrying out a physicalexamination and testsExercising 'clinicaljudgement'*
Administering a questionnaire,(and collating the replies)As above by a questionnaire, butalso by an objective assessment ofwork/environmental exposuresGathering data from health
surveillance tests etcAnalysis of the data
Note*: The above analogies can be extended to the history taking
process.
7/31/2019 Occupational Epidemiology
8/33
7/31/2019 Occupational Epidemiology
9/33
Analogies between dealing with individual patientsand with populations in occupational epidemiology
(3)
Consulting with anindividual patient
Contending with a group ofworkers or other people
Outcome: Change in patient'scondition
Change in profile of health anddisease
7/31/2019 Occupational Epidemiology
10/33
Types of studies in OE
Case control studies
Cross sectional studies
Cohort studies
7/31/2019 Occupational Epidemiology
11/33
Case control studies
The accompanying image shows anoccupational cancer - of the nasalsinuses - which is fortunately rarernow than it used to be.
In a case-control approach, 'cases'of this disease are compared to acarefully matched referencepopulation of 'controls' (or'referents') to determine
retrospectively what differencesthere may have been in theiroccupational exposure histories.
7/31/2019 Occupational Epidemiology
12/33
For example in one study it was shown that patientswith one type of this tumour (an adenocarcinoma)had a disproportionately greater likelihood of having
worked in the hardwood industry, when compared toreferent patients with other pathology.
Subsequently cohort studies were conducted whichsupported the hypothesis of a causal association by
showing that this type of tumour was muchcommoner in hardwood workers than in otherpeople.
7/31/2019 Occupational Epidemiology
13/33
Cross sectional studies
In a cross-sectional study, the prevalence of aparticular disease or of a set of symptoms or otherindication of ill-health is investigated in a singletime-point (or over a relatively narrow period of
time). Comparisons can then be made in the frequency of
ill-health.
E.g. comparisons between workers exposed to a
particular hazard, and those who are not, or - betterstill - between workers experiencing differentdegrees of exposure.
7/31/2019 Occupational Epidemiology
14/33
A cross sectional study can determine the prevalencerate, which is defined as the number of EXISTINGcases of disease divided by the population at aspecified time point.
E.g. a chest X-ray survey of quarry workers isconducted it might show that workers in quarries
with high exposure to quartz (a crystalline form of
silica) might have a higher prevalence ofpneumoconiosis than those in quarries with little orno such exposure.
7/31/2019 Occupational Epidemiology
15/33
Cohort studies
The term cohort is applied epidemiologically to a clearlydefined population who prospectively share a commonexperience -say an occupational exposure.
Comparisons in the incidence of ill-health, or inmortality can be made between exposed and non-exposed cohorts or between subsets of the same cohort
but with different degrees of exposure. In the context of a cohort study the term 'control' has a
different meaning from the meaning in a case-controlstudy.
In a cohort study, the 'controls' are those people who arenot (or have not been) exposed to the agent underinvestigation.
7/31/2019 Occupational Epidemiology
16/33
Some cohort studies measure mortality, others measureincidence of disease.
The incidence rate is defined as the number of NEWcases of a disease divided by the population at risk over agiven period of time.
In OE, it is important to be able to characterise theexposure of the 'population at risk'.
E.g. a study was conducted in which workers werefollowed up during their employment in a factory bin
which they were exposed to benzene. It showed thatthose categories of workers with a high cumulativeexposure to benzene had a higher mortality from certaintypes of leukaemia than the control population.
7/31/2019 Occupational Epidemiology
17/33
Problems in occupational epidemiology
The Healthy Worker Effect
Other problems
7/31/2019 Occupational Epidemiology
18/33
The healthy worker effect
Workers differ from the general population fromwhich they are drawn, and especially fromunemployed people in many ways.
This differences can result in serious bias inoccupational epidemiology.
Can you consider some of these differences, and howthey can contribute to bias?
E.g. are there socio-economic differences betweenthe employed and the unemployed (can thesedifferences influence health, and if so - in what way?)
7/31/2019 Occupational Epidemiology
19/33
The HWE refers to the observation that employed pop tend tohave lower mortality experience then the general pop. HWE is 1 of the factors that may reduce the validity of exposure
data. The HWE may have an impact on occupational mortality studies
in several ways Example: People whose life expectancy is shortened by dz are less likely to be
employed than healthy person. This phenomenon lead to reduced measure of effect for an exposure that
increases morbidity & mortality. Why? Because general pop includes both employed & unemployed ind,
the mortality rate of that pop may be somewhat elevated compared with apop in which everyone is health enough to work. As a result, any excess mortality a/w a given occupational exposure is
more difficult to detect when the HWE is operative
7/31/2019 Occupational Epidemiology
20/33
The HWE is likely to be stronger for nonmalignant causesof mortality (which usually produce workerattrition/withdrawal during an earlier career phase), thanfor malignant causes of mortality (which typically havelonger latency periods & occur later in life).
In addition, healthier workers may have greater totalexposure to occupational hazards than those who leavework force at an earlier age because of illness.
Trying to understand HWE was not easy considering the
ongoing debate on its nature. Some scientists consider HWE a source of selection bias;
others consider it confounding. A third group considers it a mix of both while some
others look at it as a comparison problem
7/31/2019 Occupational Epidemiology
21/33
Healthy User Bias
The healthy user bias is a bias that can damage the validityof epidemiologic studies testing the efficacy of particulartherapies or interventions.
Specifically, it is a sampling bias: the kind of subjects thatvoluntarily enroll in a clinical trial & actually follow theexperimental regimen are not representative of the generalpop.
They can be expected to on average be healthier as they areconcerned for their health & are predisposed to followmedical advice, both factors that would aid one's health.
In a sense, being healthy/active about one's health is aprecondition for becoming a subject of the study, an effectthat can appear under other conditions such as studyingparticular groups of workers (i.e. someone in ill-health isunlikely to have a job as manual laborer).
7/31/2019 Occupational Epidemiology
22/33
Other problems
Various difficulties can affect the design andinterpretation of studies in occupationalepidemiology.
Can you consider some of these? For example how comparable is an occupational
population to the general public in terms of age, andsex?
Bias, confounding and chance are considered brieflyelsewhere.
7/31/2019 Occupational Epidemiology
23/33
Chance
Example - To determine the frequency of back painamong employees in a particular workplace.
Rather than questioning all the employees, it would beeasier to administer questionnaires to only a sample of
this pop, & from them, estimate the frequency of backpain in the workers.
However, CHANCE may have affected the resultsbecause of random variation in the population
It could be that, by CHANCE, the sample were aparticularly fit & healthy group.
The larger the size of the sample, the smaller the effectthat CHANCE will have on the results.
7/31/2019 Occupational Epidemiology
24/33
Bias
A further important factor to consider is whethersome aspect of the design, or conduct of the studyhas introduced a systematic error or BIAS into theresults.
BIAS is most easily understood if you think in terms
of the danger of not comparing 'like with like'. 'BIAS' in an epidemiologic context has a specific
meaning which does not necessarily imply bad faith. IOW, a criticism of a study as having been biased
does not necessarily nor usually mean that the
investigators set out with the intent of swaying theresults or interpretation one way or the other.
Types of Bias: Selection & Participation Bias,Observation Bias & Recall Bias
7/31/2019 Occupational Epidemiology
25/33
Bias Types: Selection & Participation Bias
Occurs if the study pop being compared are notstrictly comparable ('not comparing like withlike)
Example: In a study to determine the effect of aWorkplace Health Promotion (WHP) program on'sickness absence', the rate of subsequentsickness absence might have been comparedbetween those who participated in the WHP
program & those who did not.
Results might appeared to show that the WHPgroup had lower rates of sickness absence?
7/31/2019 Occupational Epidemiology
26/33
Bias Types: Selection & Participation Bias (2)
However, bias may well have been present in thisstudy because those who took part in the WHPmay, for other reasons, such as their smokinghabits, diet, or psychological factors, have been at
a lower likelihood of sickness absence evenbefore joining the WHP.
IOW, bias would have risen through notcomparing like with like.
Selection bias also may occur if some workers areexcluded because their records have been purgedfrom the companys database.
7/31/2019 Occupational Epidemiology
27/33
Bias Types: Observation Bias
Occurs if non-comparable info is obtained from eachstudy group.
Example In a case-control study to determine whetherscleroderma (systemic sclerosis) is a/w occupational
exposure to certain hazards (e.g trichloroethylene orsilica)
If the interviewer knew (or could tell) which people werethe cases & which were the controls then interviewer
might seek more detail about exposures from the casesthan the controls (referents)
Thus observer bias would have influenced the results.
7/31/2019 Occupational Epidemiology
28/33
Bias Types: Recall Bias
Recall bias might arise if the cases (suffering from thedisease) having previously pondered about possiblecauses of their misfortune, were to recollect more detailabout their past exposures, than the controls (who may
have no real motivation to reflect at length on their pastoccupations).
E.g. some retrospective studies of human exposure(Exposure Assessment) rely on surveys & the recall of the
exposed persons. This latter method, is the least reliable& 1 reason that data from some epidemiologic studiescannot always be used for quantitative risk assessment.
7/31/2019 Occupational Epidemiology
29/33
Confounding
Results from multiple associations bet theexposure, the dz, & a 3rd factor (i.e. theCONFOUNDING variable') which is a/w boththe exposure, & independently affects the riskof developing the dz.
Th li ti f ti l
7/31/2019 Occupational Epidemiology
30/33
The application of occupationalepidemiology
Benefit for the workers Occupational epidemiology has a great deal to contribute
to the reduction of risks to health from work, throughreducing exposure, and in other ways.
Benefit for the community at large Various direct & indirect benefits can accrue to the
population at large. E.g. through the derivation and application of exposure
limits the recommendations of the Expert Panel on AirQuality Standards in relation to benzene were largelybased on occupational epidemiology.
7/31/2019 Occupational Epidemiology
31/33
Conclusion
Occupational epidemiology is an important aspect ofclinical epidemiology and of occupational hygienesince it provides:
powerful and practical information to understand the causesand determinants of work related ill-health,
to help establish what steps should be taken to reduce thoserisks, and
to evaluate interventions for the benefits of workers, and of the
community at large.
7/31/2019 Occupational Epidemiology
32/33
References
Friss, RH (2009). Epidemiology for Public Health Practice. Fourthedition. Jones and Bartlett.
Rothman, KJ (1998). Modern Epidemiology. Second edition.Lippincott Raven.
Ladou, J (2007). Current Occupational & Environmental Medicine.Fourth edition. Mc Graw Hill.
Agius, R (2006). Association and Cause. [Online]www.agius.com/hew/resource/assoc.htm accessed on 6 Disember2011.
Shah, D (2009). Healthy worker effect phenomenon. PubMedCentral. [Online]
www.ncbi,nlm.nih.gov/pmc/articles/PMC2847330 accessed on 7Disember 2011. Li, CY (1999). A review of the healthy worker effect in occupational
epidemiology. Occup Med 49: 225-229.
http://www.agius.com/hew/resource/assoc.htmhttp://www.ncbi%2Cnlm.nih.gov/pmc/articles/PMC2847330http://www.ncbi%2Cnlm.nih.gov/pmc/articles/PMC2847330http://www.agius.com/hew/resource/assoc.htm7/31/2019 Occupational Epidemiology
33/33
THANK YOU