Confounding can be addressed either at the study design stage, or adjusted for at the analysis stage providing sufficient relevant data have been collected. A number of methods can be applied to control for potential confounding factors and the aim of all of them is to make the groups as similar as possible with respect to the confounder s. Potential confounding factors may be identified at the design stage based on previous studies or because a link between the factor and outcome may be considered as biologically plausible.
Methods to limit confounding at the design stage include randomisation, restriction and matching. This is the ideal method of controlling for confounding because all potential confounding variables, both known and unknown, should be equally distributed between the study groups.
It involves the random allocation e. However, this method can only be used in experimental clinical trials. Restriction limits participation in the study to individuals who are similar in relation to the confounder. For example, if participation in a study is restricted to non-smokers only, any potential confounding effect of smoking will be eliminated. However, a disadvantage of restriction is that it may be difficult to generalise the results of the study to the wider population if the study group is homogenous.
Matching involves selecting controls so that the distribution of potential confounders e. In practice this is only utilised in case-control studies, but it can be done in two ways:. The presence or magnitude of confounding in epidemiological studies is evaluated by observing the degree of discrepancy between the crude estimate without controlling for confounding and the adjusted estimate after accounting for the potential confounder s.
If the estimate has changed and there is little variation between the stratum specific ratios see below , then there is evidence of confounding. It is inappropriate to use statistical tests to assess the presence of confounding, but the following methods may be used to minimise its effect.
Stratification allows the association between exposure and outcome to be examined within different strata of the confounding variable, for example by age or sex. The strength of the association is initially measured separately within each stratum of the confounding variable. Assuming the stratum specific rates are relatively uniform, they may then be pooled to give a summary estimate as adjusted or controlled for the potential confounder. An example is the Mantel-Haenszel method.
One drawback of this method is that the more the original sample is stratified, the smaller each stratum will become, and the power to detect associations is reduced. Statistical modelling e. It is the most commonly used method for dealing with confounding at the analysis stage. Standardisation accounts for confounders generally age and sex by using a standard reference population to negate the effect of differences in the distribution of confounding factors between study populations.
It is only possible to control for confounders at the analysis stage if data on confounders were accurately collected. Residual confounding occurs when all confounders have not been adequately adjusted for, either because they have been inaccurately measured, or because they have not been measured for example, unknown confounders.
An example would be socioeconomic status, because it influences multiple health outcomes but is difficult to measure accurately. Interaction occurs when the direction or magnitude of an association between two variables varies according to the level of a third variable the effect modifier.
For example, aspirin can be used to manage the symptoms of viral illnesses, such as influenza. Where interaction exists, calculating an overall estimate of an association may be misleading. Unlike confounding, interaction is a biological phenomenon and should not be statistically adjusted for. A common method of dealing with interaction is to analyse and present the associations for each level of the third variable.
Interaction can be confirmed statistically, for example using a chi-squared test to assess for heterogeneity in the stratum-specific estimates. However, such tests are known to have a low power for detecting interaction 5 and a visual inspection of stratum-specific estimates is also recommended.
Skip to main content. Create new account Request new password. You are here 1a - Epidemiology. Bias in Epidemiological Studies While the results of an epidemiological study may reflect the true effect of an exposure s on the development of the outcome under investigation, it should always be considered that the findings may in fact be due to an alternative explanation 1. Bias Bias may be defined as any systematic error in an epidemiological study that results in an incorrect estimate of the true effect of an exposure on the outcome of interest.
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.
Common types of bias in epidemiological studies 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. Information bias Information bias results from systematic differences in the way data on exposure or outcome are obtained from the various study groups.
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. Use of standardised questionnaires or calibrated instruments, such as sphygmomanometers.
Training of interviewers. Methods to minimise recall bias include: Collecting exposure data from work or medical records. Blinding participants to the study hypothesis. 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.
Confounding Confounding, interaction and effect modification Confounding provides an alternative explanation for an association between an exposure X and an outcome. The variable must also be associated with the exposure under study in the source population. The variable should not lie on the causal pathway between exposure and disease. Examples of confounding A study found alcohol consumption to be associated with the risk of coronary heart disease CHD.
Effects of confounding Confounding factors, if not controlled for, cause bias in the estimate of the impact of the exposure being studied. The effects of confounding may result in: An observed association when no real association exists.
No observed association when a true association does exist. An underestimate of the association negative confounding. An overestimate of the association positive confounding. Controlling for confounding Confounding can be addressed either at the study design stage, or adjusted for at the analysis stage providing sufficient relevant data have been collected. Controlling for confounding at the design stage Potential confounding factors may be identified at the design stage based on previous studies or because a link between the factor and outcome may be considered as biologically plausible.
In practice this is only utilised in case-control studies, but it can be done in two ways: Pair matching - selecting for each case one or more controls with similar characteristics e. Residual confounding It is only possible to control for confounders at the analysis stage if data on confounders were accurately collected.
Carneiro I, Howard N. Introduction to Epidemiology. Open University Press, BMJ ; Marshall SW. Power for tests of interaction: effect of raising the type 1 error rate. Epidemiological perspectives and innovations ; Our most popular content Public Health Textbook. Identifying and managing internal and external stakeholder interests. Management models and theories associated with motivation, leadership and change management, and their application to practical situations and problems.
Dietary Reference Values DRVs , current dietary goals, recommendations, guidelines and the evidence for them. Section 1: The theoretical perspectives and methods of enquiry of the sciences concerned with human behaviour.
Inequalities in health e. Stratification Analyse the data in subgroups for each potential confounding factor. References Sackett, D. Bias in analytic research. Journal of Chronic Diseases 32 1—2 : 51— PS Myles, T Gin. Statistical methods for anaesthesia and intensive care.
Oxford: Butterworth-Heinemann, Stats notes from my MPh University of Sydney. Probably a Timothy Schlub lecture, circa Last updated
0コメント