ISSUES RELATED TO STUDY DESIGN

Types of Studies

To interpret the literature, it is important to understand the strengths and limitations of the various study designs. Prospective cohort studies involve the identification of a group of subjects who are monitored over time for the occurrence of an outcome of interest. The goal is to identify patients who develop the outcome. For studies of perioperative mortality, individual cases can be reviewed to determine the cause of mortality.

895

Figure 24-1 Representation of the influences of various components on poor perioperative outcomes. Surgical, anesthetic, and patient characteristics all contribute to outcome. Anesthesia-related contributions can include issues of judgment and mishaps, as well as characteristics of the provider. The surgical procedure itself affects outcome, as does the location of intraoperative and postoperative care.

Another example of a prospective cohort study is one in which patients with a known disease are studied for the development of predefined outcomes. Such studies provide the natural history of the disease. For example, studies of patients who have sustained a myocardial infarction could be important in determining the optimal time between the infarct and surgery.^{[14] [15] [16]}

Although prospective cohort studies have important value in identifying risk factors for the outcome of interest, there are significant limitations. The selection of the cohort of interest can significantly affect the results obtained. The larger the cohort, the more the results can be generalized. A second bias is that many patients may be lost to follow-up. In perioperative studies, this may not be an important issue for short-term outcomes. The importance of a risk factor depends on the completeness of the data. For example, if the presence of severe angina were not included in the database, it could not be considered as a risk factor, and other factors may appear to be more important.^[13]

A specific example of a prospective cohort study is the randomized clinical trial. Randomized clinical trials represent the gold standard for evidence of causation. They have defined inclusion and exclusion criteria, treatment protocols, and outcomes of interest. They are usually single- or double-blinded (to patient and physician) studies and are designed to test the effect of a new drug or intervention. They rarely are used as a means of identifying risk, but they may be used to determine whether a risk factor is truly linked by a causal relationship or is simply an association. For example, hypothermia in the perioperative period has been associated with an increased incidence of perioperative ischemia, a surrogate marker for morbidity.^[17] In a randomized clinical trial, the use of forced-air warming to maintain normothermia was associated with a significantly lower incidence of perioperative morbid cardiac events.^[18] Randomized clinical trials directed at interventions that reduce the frequency of a given risk factor in the population are frequently performed after the results of a prospective cohort study to confirm the findings of an association between poor outcome and that factor.

Randomized clinical trials derive their strength from an evidence-based perspective because of their high degree of internal validity; the randomization scheme and use of placebo (or accepted alternative treatments) provide strong evidence that the results are related to the intervention. Importantly, these trials have a lower degree of external validity because the intervention may not behave in the same manner when it is diffused in a more heterogeneous population in whom treatment is not defined.

Retrospective studies involve identification of patients who have sustained an outcome and definition of risk factors associated with the outcome. An example of a retrospective design is a case-control study. Case-control studies identify patients with the outcome of interest. Frequently, these patients are included as part of a prospective cohort study. The prevalence of a risk factor in the patients with the outcome (i.e., cases) is compared with the prevalence of the risk factor in matched controls to maximize the efficiency and power of the results. The ratio of cases to controls can be varied, yielding greater power with an increasing number of controls.

Case-control studies are subject to several biases. The exact definitions of cases and controls can influence the analysis. Frequently, patients are matched for age and gender, but other factors may play an important role. It is critical that the exposure precedes the outcome, and a dose-response gradient further confirms the relationship.