Selection and information bias.pptx.pptx

This slide is about selection and information bias.

1. Selection and Information Bias Anil Shrestha MPH (CDPH, IOM )

2. Contents Definition of bias Causes of bias Definition of selection bias Types of selection bias and its examples Definition of information bias Types of information bias and its examples Other types of biases References 2

3. Learning Objectives Participants will be able to: Define bias List out the causes of bias Identify different types of bias in different study designs used in epidemiology. Explain about different types of biases with its examples 3

4. Definition of Bias Bias has been defined as “any systematic error in the design, conduct or analysis of study that results in a mistaken estimate of an exposures effect on the risk of disease.” Schlesselman JJ Bias is any systematic error in an epidemiologic study that results in an incorrect estimate of the association between exposure and the health outcome. Bias occurs when an estimated association (risk ratio, rate ratio, odds ratio, difference in means, etc.) deviates from the true measure of association. May be present without investigator being aware. Sources may be difficult to identify. 4

5. Causes of Bias Bias results from systematic errors in the research methodology. The effect of bias will be an estimate either above or below the true value, depending on the direction of the systematic error. The magnitude of bias is generally difficult to quantify , and limited scope exists for the adjustment of most forms of bias at the analysis stage. As a result, careful consideration and control of the ways in which bias may be introduced during the design and conduct of the study is essential in order to limit the effects on the validity of the study results. We have different perspectives based on our race, gender, ethnicity, religion, sexual orientation, socioeconomic status, nationality, and a whole array of other factors. 5

6. Classification of bias More than 50 types of bias have been identified in epidemiological studies, but for simplicity they can be broadly grouped into two categories: information bias and selection bias. Selection bias Differential access to the study population Information bias Inaccuracy in measurement or classification 6

7. Selection bias Selection bias is an error in selecting a study group or groups (exposed or nonexposed , or cases or controls) within the study. It can have a major impact on the internal validity of the study and the legitimacy of the conclusion. 7

8. Cause of selection bias Selection bias occurs when there is a systematic difference between either: Those who participate in the study and those who do not (affecting generalisability) or Those in the treatment arm of a study and those in the control group (affecting comparability between groups). That is, there are differences in the characteristics between study groups, and those characteristics are related to either the exposure or outcome under investigation. Selection bias can occur for a number of reasons. 8

9. Types of selection bias Sampling bias describes the scenario in which some individuals within a target population are more likely to be selected for inclusion than others. For example, if participants are asked to volunteer for a study, it is likely that those who volunteer will not be representative of the general population, threatening the generalisability of the study results. Volunteers tend to be more health conscious than the general population. Allocation bias occurs in controlled trials when there is a systematic difference between participants in study groups (other than the intervention being studied). This can be avoided by randomisation. 9

10. Types of selection bias Loss to follow-up is a particular problem associated with cohort studies. Bias may be introduced if the individuals lost to follow-up differ with respect to the exposure and outcome from those persons who remain in the study. The differential loss of participants from groups of a randomised control trial is known as attrition bias . Exclusion bias: Collective term covering the various potential biases that can result from the post-randomization exclusion of patients from a trial and subsequent analyses. This may also be referred to as attrition bias. 10

11. Types of selection bias Berkson’s bias A form of selection bias that causes hospital cases and controls in a case control study to be systematically different from one another because the combination of exposure to risk and occurrence of disease increases the likelihood of being admitted to the hospital . This produces a systematically higher exposure rate among hospital patients, so it distorts the odds ratio. The bias was described by the American statistician Joseph Berkson (1899–1982). Prevalence/Neyman Bias Selecting completely healthy or very sick participants 11

12. Selection bias in different types of studies Selection bias in case-control studies Selection bias is a particular problem inherent in case-control studies, where it gives rise to non-comparability between cases and controls. In case-control studies, controls should be drawn from the same population as the cases, so they are representative of the population which produced the cases. Controls are used to provide an estimate of the exposure rate in the population. Therefore, selection bias may occur when those individuals selected as controls are unrepresentative of the population that produced the cases. 12

13. Example Results of a Matched-Pairs Analysis of a Case-Control Study of Reserpine Use and Breast Cancer 13

14. Controls Breast cancer cases Used Reserpine Did Not Use Reserpine Used Reserpine 8 45 Did Not Use Reserpine 23 362 Matched-pairs odds ratio = 45/23 = 1.96 Adapted from Heinonen OP, Shapiro S, Tuoominen I., et al: Reserpine use in relation to breast cancer. Lancet 2:675-677, 1974. 14

15. Contd… Exclusion bias results when investigators apply different eligibility criteria to the cases and to the controls. Horwitz and Feinstein tried to replicate the reserpine study in 257 women with breast cancer and 257 controls, calculating odds ratios in two ways: first, including all the women and second after excluding women with cardiovascular disease from the controls where odds ratio was found to be 1.1 and 2.5 respectively. 15

16. Selection bias in different types of studies Selection bias in cohort studies Selection bias can be less of problem in cohort studies compared with case-control studies, because exposed and unexposed individuals are enrolled before they develop the outcome of interest. However, selection bias may be introduced when the completeness of follow-up or case ascertainment differs between exposure categories. For example, it may be easier to follow up exposed individuals who all work in the same factory, than unexposed controls selected from the community (loss to follow-up bias). This can be minimised by ensuring that a high level of follow-up is maintained among all study groups. 16

17. Selection bias in different types of studies Selection bias in randomised trials Randomised trials are theoretically less likely to be affected by selection bias, because individuals are randomly allocated to the groups being compared, and steps should be taken to minimise the ability of investigators or participants to influence this allocation process. However, refusals to participate in a study, or subsequent withdrawals, may affect the results if the reasons are related to both exposure and outcome. 17

18. 18 Information bias

19. Definition Information bias, also called measurement bias or observational bias, arises when key study variables (exposure, health outcome, or confounders) are inaccurately measured or classified. It can occur when the means for obtaining information about the subjects are inadequate so that some information gathered regarding exposures and/or disease outcomes is incorrect. Information bias results from systematic differences in the way data on exposure or outcome are obtained from the various study groups. This may mean that individuals are assigned to the wrong outcome category, leading to an incorrect estimate of the association between exposure and outcome. 19

20. Example of information bias Studies of rare or newly discovered diseases that do not have uniform diagnostic criteria are at risk for information bias. In the absence of a common standard, people who do not have a disease may be classified as having it, and vice versa . Another example: exposure data based on interviews where participants may either not be aware of exposure or may erroneously think that it did not occur. Also, when the ascertainment of exposure is based on old records. 20

21. Types of information bias Observer bias may be a result of the investigator’s prior knowledge of the hypothesis under investigation or knowledge of an individual's exposure or disease status. Such information may influence the way information is collected, measured or interpretation by the investigator for each of the study groups. For example, in a trial of a new medication to treat hypertension, if the investigator is aware which treatment arm participants were allocated to, this may influence their reading of blood pressure measurements. Observers may underestimate the blood pressure in those who have been treated, and overestimate it in those in the control group. 21

22. Observer bias can be reduced or eliminated by: Ensuring that observers are well trained. Screening observers for potential biases. Having clear rules and procedures in place for the experiment. Making sure behaviors are clearly defined. Setting a time frame for: collecting data, for the duration of the experiment, and for experimental parts. 22

23. Types of information bias contd… 2. Interviewer bias occurs where an interviewer asks leading questions that may systematically influence the responses given by interviewees. Minimising interviewer bias: Where possible, observers should be blinded to the exposure and disease status of the individual Blind observers to the hypothesis under investigation. In a randomised controlled trial, blind investigators and participants to treatment and control group (double-blinding). Development of a protocol for the collection, measurement and interpretation of information. 23

24. Use of standardized questionnaires or calibrated instruments, such as sphygmomanometers. Training of interviewers. 3. Recall (or response) bias - In a case-control study data on exposure is collected retrospectively. The quality of the data is therefore determined to a large extent on the patient's ability to accurately recall past exposures. Recall bias may occur when the information provided on exposure differs between the cases and controls. For example, an individual with the outcome under investigation (case) may report their exposure experience differently than an individual without the outcome (control) under investigation. 24

25. Recall bias may result in either an underestimate or overestimate of the association between exposure and outcome. Methods to minimise recall bias include: Collecting exposure data from work or medical records. Blinding participants to the study hypothesis. 4. Social desirability bias is a type of response bias where there is the tendency of survey respondents to answer questions in a manner that will be viewed favorably by others. It can take the form of over-reporting "good behavior" or under-reporting "bad", or undesirable behavior. 25

26. Types of information bias 5. In reporting bias , individuals may selectively suppress or reveal information, for similar reasons (for example, around smoking history). Reporting bias can also refer to selective outcome reporting by study authors. 6. Performance bias refers to when study personnel or participants modify their behaviour / responses where they are aware of group allocations. For example, if a weight loss study is investigating if a high protein diet works to reduce weight, participants might up their protein intake. This particular type of bias is also called the set of Hawthorne effects . 26

27. Hawthorne effect The Hawthorne effect refers to people's tendency to behave differently when they become aware that they are being observed. 27

28. 7. Detection bias occurs where the way in which outcome information is collected differs between groups. Blinding of outcome assessors reduces detection bias.  8. Instrument bias refers to where an inadequately calibrated measuring instrument systematically over/underestimates measurement. Blinding of outcome assessors and the use of standardized, calibrated instruments may reduce the risk of this. 28

29. Types of information bias Misclassification bias A systematic error that can occur at any stage in the research process. It occurs when an individual is assigned to a different category than the one to which they should be assigned . Errors in measurement are also known as misclassifications , and the magnitude of the effect of bias depends on the type of misclassification that has occurred. 29

30. Sometimes, people who have the disease (cases) maybe misclassified as controls and some without the disease (controls) maybe misclassified as cases . This may occur due to limited sensitivity and specificity of the diagnostic tests or from inadequacy of information derived from medical or other records . 30

31. Types of misclassification There are two types of misclassification – Differential Nondifferential 31

32. Differential misclassification Rate of misclassification differs in different study groups. Example: unexposed cases are misclassified as being exposed more often than the unexposed controls are misclassified as being exposed. 32

33. Example 1 In a case-control study, women who had a baby with malformation tended to remember more mild infections that occurred during their pregnancies than did the mothers of normal infants due to recall bias. Thus, more unexposed cases were misclassified as exposed than were unexposed controls in regard to prenatal infections. Can lead to an apparent association even if it does not exist or to an apparent lack of association where one does in fact exist. 33

34. Example 2 Emphysema is diagnosed more frequently in smokers than in non-smokers. However, smokers may visit the doctor more often for other conditions (e.g. bronchitis) than non-smokers, which means that a reason smokers could be diagnosed with emphysema more often is simply because they go to the doctor more often — not because they actually have higher odds of getting the disease. 34

35. Nondifferential misclassification Occurs when the probability of individuals being misclassified is equal or approximately equal across all groups in the study. The relative risk or odds ratio tends to be diluted and it is shifted towards 1.0 . Less likely to detect an association even if one really exists. 35

36. Example of non-differential information bias For example, in a case-control study of heart disease and past activity, both cases and controls have difficulty accurately remembering their exercise frequency, duration and intensity over many years. Note that, in this scenario cases and controls are equally likely to report their past exercise levels inaccurately. If one group remembers better, it will cause recall bias, which is differential. 36

37. Due to non-differential misclassification, controls will not have such a low rate of exposure and cases will not have such a high rate of exposure from our data. So, smaller difference in exposure will be found between cases and controls than the actual. 37

38. Other types of biases Researcher bias Publication bias Bias in abstracting records Abstract bias Bias from surrogate interviews Surveillance bias Analytic bias Confounding bias 38

39. Researcher bias Researcher bias occurs when the researcher’s beliefs or expectations influence the research design or data collection process. Example: Researcher bias Suppose you want to study the effects of alcohol on young adults. If you are already convinced that alcohol causes young people to behave in a reckless way, this may influence how you phrase your survey questions. Instead of being neutral and non-judgmental, they run the risk of reflecting your preconceived notions around alcohol consumption. As a result, our survey will be biased. 39

40. Publication bias Publication bias refers to the selective publication of research studies based on their results . Here, studies with positive findings are more likely to be published than studies with negative findings. Positive findings are also likely to be published quicker than negative ones. As a consequence, bias is introduced: results from published studies differ systematically from results of unpublished studies. 40

41. Bias in abstracting records Bias may be introduced in the way that information is abstracted from medical, employment or other records or from manners in which interviewers ask questions. 41

42. Abstract bias The tendency to report only significant results in the abstract, while reporting non-significant results within the main body of the manuscript (not reporting non-significant results altogether would constitute selective reporting). The consequence of abstract bias is that studies reporting non-significant results may not be captured with standard meta-analytic search procedures (which rely on information in the title, abstract and keywords) and thus biasing the results of meta-analyses. 42

43. Bias from surrogate interviews When interviewing a parent of a child or next of kin for a deceased person. They may remember the deceased person better and report fewer risky behaviors than the person would report and that actually happened. In case of diseases with high case fatality rate. 43

44. Surveillance bias Surveillance bias is a type of information bias which occurs when one group of subjects is followed up more closely than others. May lead to biased estimates of exposure/ disease relationship. 44

45. Analytic bias In any study, if the epidemiologists and statisticians who are analyzing the data have strong preconceptions, they may unintentionally introduce their biases into their data analyses and into their interpretation of the study findings. 45

46. Confounding bias A distortions that modifies an association between an exposure and an outcome because a factor is independently associated with the exposure and the outcome. 46

47. References http://evolve.elsevier.com/Gordis/epidemiology/ https://sph.unc.edu/wp-content/uploads/sites/112/2015/07/nciph_ERIC14.pdf https://www.healthknowledge.org.uk/public-health-textbook/research-methods/1a-epidemiology/biases https://www.wallstreetmojo.com/hawthorne-effect-2/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3237868/ https://www.statisticshowto.com/performance-bias/ Information Bias (Observation Bias) (bu.edu) Differential & Non-Differential Misclassification - Statistics How To 47

48. 48