Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
PHC 6000: BiasConf
Introduction to epidemiology: Variability, Bias, Confounding, and Causality
| Question | Answer |
|---|---|
| What is bias? | A systematic, non-random deviation of results and inferences from the truth, or processes leading to such deviation |
| What is the effect of bias? | It creates an association that is not true |
| What is the effect of confounding? | It describes an association that is true, but potentially misleading |
| Can bias and confounding be corrected? | Bias cannot be corrected during the analysis phase, while confounding can be corrected by applying a specific analysis tool given that is planned at the conception phase of the study |
| What are the types of bias? | Selection bias, information bias, and confounding |
| What is selection bias? | Absence of comparability between groups studied |
| What is information bias? | Incorrect determination of exposure or outcome or both; collecting data incorrectly |
| What is confounding? | Distortion of exposure-disease relation by some other factor |
| What type of bias occurs when cases and controls are identified not independently of the exposure? | Selection bias |
| What is detection bias? | Preferential diagnosis or admission of exposed cases (e.g., an exposure leading to search for an outcome) |
| What is non-response bias? | Exposed being more/less likely to return questionnaires |
| What is incidence-prevalence bias? | Occurs when prevalent cases are used |
| What causes information bias? | Imperfect definitions of study variable or flawed data collection procedures |
| What are some types of selection bias? | Detection bias, non-response bias, incidence-prevalence bias |
| What are some types of information bias? | Recall bias, information bias |
| What is recall bias? | Cases are more likely to remember exposure than controls |
| What is interviewer bias? | Interviewer may collect information differently if they knew the disease status |
| How can recall bias be minimized? | Validate self-report (e.g., looking for a biomarker) and encouraging recall (e.g., using memory aids, photos) |
| How can interviewer bias be minimized? | Train & blind interviewers and standardize & monitor interviews |
| Are biological assessments of exposure free from information bias? | Yes |
| How can confounding be minimized? | Restriction, matching, stratification, or adjustment |
| How are confounders controlled in case-control studies? | Matching |
| What characteristics are usually used to match cases and controls? | Major socio-demographic attributes (e.g., sex, age, SES) |
| Can matched variables be considered in the analysis? | No, matching ensures that case-control difference cannot be explained by the difference in the matched variables |
| What is the relationship between matching and power? | Matching allows for more than one control for each case, thus increasing power (especially helpful for studying extremely rare diseases) |
| Matching may increase the precision of case-control comparisons and allow a smaller study. (T/F) | True |
| What are the disadvantages of matching in case-control studies? | -Can be time consuming and expensive -Some potential cases and controls have to be excluded because matches cannot be made -The matched variables cannot be evaluated as risk factors in the study population |
| Should a limit be placed on the number of variables matched? | Matching on too many variables makes it difficult to find controls |
| Is matching effective regardless of variable types? | For continuous or ordinal variables, matching categories may be too broad and residual case-control differences in these variables may persist |
| If it is not possible to blind the interviewers, what can be done instead? | Don't share the study hypothesis with them |
| Is a confounder on the pathway between exposure and disease? | No, it is not but it is related to both and has to be a risk factor on its own |
| Can matched variables be evaluated as risk factors? | No |
| How are results affected if nonparticipants are systematically different from participants? | Can lead to biased results |
| What is precision also known as? | Reliability |
| What is reliability also known as? | Precision |
| What is consistency of measurement of an attribute? | Precision (reliability) |
| What is precision? | Consistency of measurement of an attribute |
| What is reliability? | Consistency of measurement of an attribute |
| What is validity? | The degree to which a measurement or study reaches a correct conclusion |
| What is the degree to which a measurement or study reaches a correct conclusion? | Validity |
| What is internal validity? | Whether the study provides an unbiased estimate of what it claims to estimate |
| What is external validity | Whether the results from the study can be generalized to some other population |
| What is the difference between internal and external validity? | Internal validity is the degree to which a study provides an unbiased estimate of what it claims to be, while external validity is the degree to which study results can be generalized to some other population. |
| What is accuracy also known as? | Validity |
| What is validity also known as? | Accuracy |
| What is the term for reproducibility of the result? | Precision (reliability) |
| What is the term for nearness to the "true" value? | Accuracy (validity) |
| What are the two broad types of error? | Random error and systematic error (bias) |
| What is random error? | Deviation of results and inferences from the truth, occurring as a result of chance. |
| What type of error reflects variability of the estimate? | Random error |
| What is systematic error (bias)? | An error resulting in an incorrect estimate of the population attribute or association between exposure and risk of disease. |
| What is random error also known as? | Chance |
| What is systematic error also known as? | Bias |
| What is chance also known as? | Random error |
| What is bias also known as? | Systematic error |
| What is the impact of random error in analytical studies? | Low precision |
| What type of error increases the variability of an estimate and obscures a real difference (imprecise results)? | Random error (chance) |
| What can be done to compensate for random error? | Larger sample size |
| What is the impact of systematic error in analytical studies? | Low validity |
| Low precision is a result of what? | Random error (chance) |
| Low validity is a result of what? | Systematic error (bias) |
| What type of error results in false estimates or difference (inaccurate result)? | Systematic error (bias) |
| What can be done to compensate for systematic error? | Nothing. Bias cannot be overcome by 'statistics' if present. |
| At what steps in epi studies can errors occur? | -Selection of study participants -Classification and measurement -Comparison and interpretation |
| What is the major concern in analytical studies? | Internal validity |
| Why is internal validity the major concern in analytical studies? | Extrapolation of invalid results (external validity) to the larger population is useless and potentially dangerous |
| What is the main threat to internal validity of an analytical epi study? | Bias |
| What are the types of bias? | Selection bias, information bias, and confounding |
| What is selection bias? | Arises when the relation between exposure and disease is different for those who participate in the study and those who theoretically would be eligible for the study but do not participate |
| What is information bias? | Can be introduced as a result of measurement error in assessment of both exposure and disease |
| What is confounding? | The distortion of the estimate of the effect of an exposure of interest because it is mixed with the effect of an extraneous factor |
| What type of bias creates a systematic difference between the compared populations that is not due to the association under study? | Selection bias |
| When does selection bias occur? | Inclusion in the study |
| How does selection bias occur? | Preferential selection of subjects related to their disease or exposure status |
| What are the consequences of selection bias? | Frequency of disease or exposure are different among those included in the study than those eligible |
| What are some types of selection bias? | -Sampling bias -Diagnostic, referral, admission bias -Participation bias (self-selection, non-response, healthy worker effect, loss to follow up, survival) |
| What are some ways to minimize selection bias? | -Clear def of study pop -Explicit case and control definitions -Cases and controls from same population in ca/co -Selection independent of exposure in ca/co -Selection of exposed & non-exposed that are comparable except for exposure in cohort studies |
| What are some sources of information bias? | -Subject variation -Observer variation -Deficiency of tools -Technical errors in measurement |
| What type of bias creates a systematic difference between the compared populations that is not due to the association under study? | Information bias |
| What is information bias also called by some? | Measurement bias |
| When does information bias occur? | During data collection |
| How does information bias occur? | -Differences in accuracy of exposure data between ca/co -Differences in accuracy of outcome data between exposed and unexposed (cohort study) |
| What are the consequences of information bias? | Misclassification: study subjects are classified in the wrong category (e.g., cases/controls, exposed/unexposed) |
| What type of error leads to assigning wrong exposure or outcome category? | Measurement error. Can be non-differential or differential. |
| What is non-differential measurement error? | Random error |
| Misclassification of exposure is equal between cases and controls in what type of measurement error? | Non-differential |
| Misclassification of outcome is equal between exposed and nonexposed in what type of measurement error? | Non-differential |
| What is the consequence of non-differential measurement error? | Weakens (dilutes) measure of association |
| What is differential measurement error? | Systematic error |
| Misclassification of exposure differs between cases and controls in what type of measurement error? | Differential |
| Misclassification of outcome differs between exposed and nonexposed in what type of measurement error? | Differential |
| What is the consequence of differential measurement error? | Measure of association distorted in any direction |
| What are the two main types of information bias? | -Reporting bias (recall bias, prevarication) -Observer bias (interviewer bias, biased follow-up, missing data) |
| What are some ways to minimize information bias? | -Standardize meas instrums -Train & monitor interviews -Use reminders -Administer instrums equally to ca/co & expos/unexpos -Blind interviews or data collectors to exp or dis status, or at least to hypothesis -Use mult sources of info -Use mult cont |
| What is a confounder? | An extraneous factor that is a risk for the disease, associated with the exposure, and not an intermediate step in the causal path between exposure and disease |
| What are the consequences of confounding? | -May show an association that does not exist -May over estimate an existing association -May under estimate an existing association -May mask an existing association -May inverse the direction of an existing association |
| At what points in a study can confounders be controlled? | Study design and data analysis |
| In the study design, how can confounders be controlled? | -Restriction -Random allocation -Matching |
| In the data analysis, how can confounders be controlled? | -Stratified analysis (Mantel Haenszel) -Implement a matched-design -Restriction (means throwing away data) -Model fitting using regression techniques |
| Can stratification only be used to control confounding in the study design? | No. If the factor was not taken into account in the protocol but is suspected at the time of analysis, can stratify the data analysis according to the potential confounder. |
| What is an effect modifier? | A third factor modifying the association between the risk factor and the disease |
| Is effect modification a bias? | No, but it is useful information |
| If effect modification is present, do strata-specific ORs provide more information than crude ORs? | Yes |
| What is effect modification also called? | Interaction |
| What is interaction also called? | Effect modification |
| Is confounding a bias? | Yes |
| What is the relationship of a third factor that modifies the association between exposure and outcome? | Effect modifier |
| What is the relationship of a third factor that distorts the association between exposure and outcome? | Confounder |
| Is the crude OR wrong or meaningless when effect modification is present? | Meaningless |
| Is the crude OR wrong or meaningless when confounding is present? | Wrong |
| How must results be analyzed when effect modification is present? | Each stratum separately |
| How must results be analyzed when confounding is present? | Adjusted OR |
| What are the types of causes? | -Sufficient cause -Necessary cause -Risk factor |
| What is a sufficient cause? | If the factor (cause) is present, the effect (disease) will always occur (e.e.g, genetic disorder such as Tay-Sach's disease) |
| What is a necessary cause? | If the factor (cause) is absent, the effect (disease) cannot occur (e.g., Tuberculosis) |
| What is a risk factor? | If the factor is present and active, the probability that the disease will occur is increased (e.g., cigarette smoking) |
| What is the key to causation? | Directionality (temporal sequence) |
| To establish causation, what else must be established first? | Directionality (either theoretically, methodologically, or both) |
Created by:
AlneciaPHS
Popular Health & Social Care sets