Confounder & Bias Lecture Notes PDF

Confounder & Bias By Ashraf Hussain MBChB. PhD/ Com. Med CONFOUNDER What may explain an association between a risk factor and an outcome? 1. Chance 2. Bias 3. Third factor (Confounder) 4. Causal association  Chance is the occurrence of events in the absence of any obvious intention...

Confounder & Bias By Ashraf Hussain MBChB. PhD/ Com. Med CONFOUNDER What may explain an association between a risk factor and an outcome? 1. Chance 2. Bias 3. Third factor (Confounder) 4. Causal association  Chance is the occurrence of events in the absence of any obvious intention or cause. It is, simply, the possibility of something happening. When the chance is defined in Mathematics, it is called probability. CONFOUNDER A part from exposure and outcome, confounder is a third factor which is associated with exposure, and independent of this association it is a risk factor for the outcome. The nuisance introduced by confounding factors May simulate an association May hide an association that does exist May alter the strength of the association – Increased – Decreased It can lead to over or under estimation of the true association and can even change the direction of the observed effect. Confounding factor Characteristics of a third, confounding factor Associated with the exposure – Without being a consequence of exposure Associated with the outcome – Independently from the exposure »/ Exposure Outcome Confounding factor Example of confounding factor Apparent association Exposure 1 Outcome Confounding factor Example of confounding factor (1) Apparent association Ethnicity Pneumonia Crowding To determine if a certain factor is a confounder: Find crude overall estimate of association between exposure and outcome. Find the adjusted estimate, after controlling of the factor. If the two estimates are different, that factor is a confounder. The selection of a potential confounder depends on: Knowledge of the disease Previous evaluation of the variable Investigator’s judgment Statistical significance is not reliable to decide on confounders In evaluation of the role of a confounder, it is necessary to decide: Its presence or absence The magnitude of its effect Its direction BIAS Any systematic error in an epidemiological study that results in an incorrect estimation of association between exposure and outcome risk It represents any deviation from the truth, causing distorted results and wrong conclusions. BIAS Bias is unlike chance or confounder, it can not be measured quantitatively, and almost impossible to be dealt with in the analysis Bias should be considered in the design phase of the study Sources The method of selection of study subjects into the study The ways in which information obtained , reported, or interpreted Types: Selection Bias Error in identification of individuals for inclusion in the study depending on the other axis (exposure in cohort, and outcome in case control) In case control: selection of the cases and control depends on different criteria which are related to exposure status Selection Bias In cohort study :if selection of exposed and non-exposed related to the development of the outcome Selection Bias Selection Bias is a problem in case control and retrospective cohort studies, since the exposure and disease had already occurred  Types of selection bias Selection bias is a general term describing errors arising from factors related to the population being studied, but there are several types of selection bias: Sampling bias or ascertainment bias occurs when some members of the intended population are less likely to be included than others. As a result, sample is not representative of population.(select healthy and sick) Attrition bias occurs when participants who drop out of a study are systematically different from those who remain.  Self-selection bias (or volunteer bias) arises when individuals decide entirely for themselves whether or not they want to participate in the study. Due to this, participants may differ from those who don’t—for example, in terms of motivation. Survivorship bias is a form of logical error that leads researchers who study a group to draw conclusions by only focusing on examples of successful individuals (the “survivors”) rather than the group as a whole. Nonresponse bias is observed when people who don’t respond to a survey are different in significant ways from those who do. Non-respondents may be unwilling or unable to participate, leading to their under- representation in the study.  Asking population about their opinion regarding certain market or restaurants Investigating the socioeconomic States of the patients of certain disease and General Hospital against is the private clinics healthy population are commonly to be recruited in the studies and seen by the researchers than the sick one Individuals who are more interested in their health are more common to be involved in the health related studies the other the time and the place of the study commonly affect the selection of the sample Prevalence of certain disease in a market sample and the type of the market. Information bias 1. OBSERVATIONAL BIAS Results from systematic difference in the way information about exposure or outcome status obtained from study groups Sources of information biases Recall Investigator Data quality Prevarication Information biases Sources of information biases Recall – Cases may recall exposure more than controls Investigator Data quality Prevarication Information biases TYPES: RECALL BIAS It arises when individuals with certain disease reported or remembered previous exposures different from those who are affected, Or those with certain exposure reported subsequent events different from those not exposed RECALL BIAS It is a problem in case control and retrospective cohort since exposure and disease are already occurred Diseased individuals tend to think about possible causes of their disease more than those who are not diseased It can lead to over or under estimation of the true association Ex. Birth defects Sources of information biases Recall Investigator (Interviewer Bias) – Systematic collection of information supporting expected conclusions Unconsciously Consciously Data quality Prevarication Information biases INTERVIEWER BIAS It refers to systematic difference in recording, or interpreting information from study participants It can occur in all studies In case control studies it is a problem in ascertainment of exposure since knowing of disease status may lead to differential probing for previous exposure history Ex. Pain killer assessment  A study can be particularly susceptible to interviewer bias when the same person who provides treatment in a study is also the same person who evaluates outcome at the end of the study. Because clinical personnel have an interest in outcome of a patient, it is likely that person will not be able to fairly evaluate the outcome of the subject. For this reason it is important that outcomes be evaluated by an independent person who has no knowledge of the treatment status of the subject. Sources of exposure and outcome information 1. Using pre-existing record (the most unbiased ). The problem is incomplete information especially of lifestyle variables 2. Use multiple sources of data 3. Use standard uniform criteria for all exposure and outcome to decrease the need for interpretation like the use of WHO criteria Sources of exposure and outcome information 4. In planning phase, practical methods to decrease loss to follow up and obtaining complete assessment of outcome should be incorporated into study design Prevarication Prevarication bias occurs when a study subject over- or under-estimates outcome because of knowledge of the kind of treatment they had received. There can be any number of reasons, for example, a subject may subconsciously feel the need to provide answers pleasing to the interviewer or study coordinator, or the subject may feel that there is some secondary gain from consciously mis-reporting their clinical status.  Level of operation of biases Type of study Case control Cohort Selection biases Inclusion Inclusion on the basis of outcome on the basis of exposure Systematic recruitment Systematic recruitment of cases / controls of exposed / unexposed with specific exposure status with specific risk of outcome Collection of information Collection of information Information biases about exposure about outcome Systematic collection of Systematic collection of information leaning towards information leaning towards specific exposure status specific outcome status Biases