Internal and External Validity in Medicine - Essay Example

Clinical research deals with studying either the therapeutic, harmful, diagnostic or prognostic potential of a particular management or condition of a patient. Three studies are most usually conducted, randomized controlled trials (RCTs), cohort and case-control studies. Briefly, RCTs randomize the population to receive different amounts of a particular exposure (such as treatment) and observe particular parameters on these participants in order to determine the outcome of exposure. On the other hand, cohort studies recruit individuals who have experienced the particular exposure being looked into. The participants are then observed for certain outcomes. In contrast, case-control studies gather individuals that manifest a particular outcome. Researchers then determine through the participants’ histories whether they have exposure or not (Dans, Dans & Silvestre, 2008; Jepsen, Johnsen, Gillman, & Sorensen, 2004). Merits of these studies are based on their internal validity, confounding, feasibility and external validity. This report looks into the merits of RCT, although it should be noted that poorness in one or two should not be the basis for ignoring the findings of the study. Internal Validity According to Taylor and Asmundson (2008), internal validity refers to the degree to which the changes in the parameters observed are caused by the identified and controlled changes in the independent variable. Those that tend to decrease internal validity include extraneous event pre- and post-test that can affect the outcome, maturation, reactivity of testing, instrumentation, statistical regression, attrition, selection, imitation of treatments, compensatory equalization of treatments, and experimenter expectancy. The effect of extraneous events and effects of selection are most easily prevented by RCTs through randomization, especially. In contrast, aside from representing the population beings studies, samples of case cohort and case-control studies should be included based on exposure and outcome, respectively. If these studies include other parameters in their inclusion criteria, it will be much more difficult to recruit subjects (Dans, Dans & Silvestre, 2008). Meanwhile, maturation and reactivity of testing can be easily addressed by comparing the exposed group with a control group, especially when the groups have similar baseline characteristics (Taylor & Asmundson, 2008). Imitation and compensatory equalization of treatments, as well as experimenter expectancy, although not possible in case-control studies, can occur in RCTs and cohorts. To avoid these, blinding of the subjects and outcome assessors can be done (Dans, Dans & Silvestre, 2008). However, it is much more difficult in the latter, because the subjects already had exposure even before the study was conducted. As well, instrumentation, statistical regression and attrition are potential issues for RCTs and cohorts. Confounding Similar to extraneous events pre- or post-test affecting the outcome described above is the concept of confounders, described by Jepsen, Johnsen, Gillman, & Sorensen (2004) as the characteristic of subjects that affect the outcome without it being a connecting link between the exposure and the outcome. As mentioned by Dans, Dans & Silvestre (2008) and as verified by Jepsen, Johnsen, Gillman, & Sorensen (2004), this can be avoided by ensuring that the baseline characteristics of the exposure and control groups are similar, best done through randomization. In addition, confounders can be prevented by restricting the inclusion criteria to those with a specific value of the confounding variable. These two can be best achieved through RCT. In cases wherein baseline characteristics are not comparable, stratified analysis may be performed. Feasibility Despite the advantages of RCTs over case cohort and case control studies in terms of ensuring internal validity and addressing confounders, this study is more treacherous to conduct. Logically, if we assign a work amount value of 1 to each of exposure and outcome observation or measurement, then RCT scores two, cohort gets one, and case-control has zero. Because of this, higher investment is needed for RCT. Moreover, the conduct of this study becomes more difficult when the outcome looked for is rare, since the experimenters will need a bigger sample size for the study. Meanwhile, if prolonged exposure is necessary, such as in the case of smoking, then follow-up is necessary. In addition, one cannot randomize people to a certain exposure, especially if the outcome to be studied is negative or harmful (Dans, Dans & Slivestre, 2008). External Validity External validity of a study determines whether or not its findings can be generalized to and across populations and settings different from those used in the study. Usually, further studies are needed to determine whether or not the findings of an earlier study are seen as well in studies using other populations. At times, stringent inclusion criteria, although increasing the internal validity of the study, it may limit usability by populations that do not meet certain qualifications of the criteria. In such case, external validity may be compromised. For example, sampling for patients who do not take medications for their condition might not be representative of the patients who seek consult in clinical practice. Aside from the difference between the samples and the general population, the setting or context of the study may also be different from the setting to which it will be applied. Similarly, whether or not the study is still externally valid warrants conduct of this study in different settings. These benchmarking studies compare the results of tightly controlled experiments that have high internal but low external validity with that of field studies (Taylor & Asmundson, 2008). If benchmarking is indeed necessary to assess external validity, it may be difficult in RCTs, because of the resources needed to conduct the study again.. However, it should be noted that external validity of a study is only of utmost importance if the research question asks, “What typically does happen?” instead of “What can happen?” Thus, in what cases external validity is considered depends on the research question, objectives and hypothesis of the study (Taylor & Asmundson, 2008). Conclusion RCT is a study design that gets a sample of the population, subjects them to a certain exposure in question, as well as observes and/or measures the outcome. Compared to the cohort and case-control studies, RCTs are more internally valid, partly by decreasing the effects of confounders. However, feasibility and external validity are compromised. References Dans, A. L., Dans, L. F., & Silvestre, M. A. A. (2008). Painless Evidence-Based Medicine. England: John Wiley and Sons, Ltd. Jepsen, P., Johnsen, S. P., Gillman, M. W., & Sorensen, H. T. (2004). Interpretation of Observational Studies. Heart, 90, 956-960. Taylor, S. & Asmundson, G. J. G. (2008). Internal and external validity in clinical research. In D. McKay (Ed.), Handbook of Research Methods in Abnormal and Clinical Psychology. California: SAGE Publications, Inc. Read More