This is an excerpt from Evidence-Based Practice in Exercise Science eBook by William E. Amonette,Kirk L. English & William J. Kraemer.
As discussed earlier in this book, research design is a critical element to answering questions relevant to mechanistic, practical, and clinical science. One of the key aspects of research design is control of all potential confounding factors to eliminate the possibility that the outcomes arising from the intervention (independent variable) were influenced by an external factor. Although experimental control is a strength of research, and in particular of randomized control trials, it can also create problems when one is implementing research evidence in client, athlete, or patient populations. Suppose a clinical scientist is interested in studying the effect of interval exercise compared to steady-state exercise on hypercholesterolemia in adults who are obese. A logical study design to address such a question would be a randomized controlled trial. Consider the following hypothetical study:
One hundred participants who were obese were selected at random and then randomly assigned to one of four groups: interval exercise, short-duration steady-state exercise, long-duration steady-state exercise, and control with no exercise. In order to eliminate potential confounding factors, the participants were free from comorbidities including hypertension, insulin resistance, diabetes, overt cardiovascular disease, and recent orthopedic injuries. Before the intervention, all groups completed a treadmill O2 maximum test and a test of blood cholesterol; the individuals conducting both of these tests were blinded to subjects' group assignment. The interval exercise group completed a protocol consisting of 1 min of exercise at a velocity equal to 90% O2max with a 2-min recovery between intervals. A total of 10 intervals were performed; thus the total exercise time for the intervention was 30 min. The steady-state short-duration exercise group performed 10 min of exercise at 70% O2max, matching the exercise (work) time of the interval group. The steady-state long-duration exercise group performed 30 min of exercise at 70% O2max, matching the total exercise and rest time of the interval exercise group. The control group did not exercise for the duration of the study. Each exercise group was prohibited from performing any other exercise, including resistance exercise. Moreover, a standardized diet was provided for all groups. The intervention was performed 3 days per week for 16 weeks; at the end of the training period, blood cholesterol and O2max were assessed to determine improvements.
On the surface, this appears to be a sound single-blind randomized controlled trial. It is in fact well-designed and controlled, but there are some potential problems when a practitioner considers implementation. Because the scientists were interested in studying training responses to an intervention in obesity alone, they eliminated subjects with diabetes, hypertension, and other common comorbidities. This was a good decision for internal validity, but not for external validity; practitioners rarely treat individuals who are obese without any of these companion conditions (Castro, Kolka, Kim, & Bergman, 2014; Despres et al., 2008). Thus, practitioners are left to wonder whether the evidence is truly applicable to their patients or clients.
A second possible problem with the evidence is the elimination of secondary interventions. Strengths of a scientific experiment may actually be perceived as weaknesses with regard to practical implementation. In practice, multiple interventions are often implemented in concert with one another. This hypothetical experiment evaluated the effectiveness of interval versus steady-state treadmill exercise alone. An exercise practitioner is unlikely to implement the interval training protocol in isolation; instead, the protocol may be combined with resistance training, a healthy diet, and other behavioral modifications. Implementation of such multifaceted protocols may enhance the effectiveness of either of the interventions. Conversely, the addition of such interventions may also reduce the effectiveness of the protocol. Obviously, it is not scientifically plausible that resistance exercise would cause an increase in cholesterol or biologically negate the effects of the interval training protocol. However, it is possible that implementation of the secondary intervention will lead to overtraining, injury, soreness, and perhaps poor adherence to the protocol due to each of these secondary results of the combined protocol. Again, this is not likely, but it is certainly something the practitioner must consider.
Another potential problem with the implementation of scientific evidence is the possibility that a client may not be able to tolerate the protocol. For example, the initial fitness level of the client may not be sufficient to enable him to tolerate its rigors. The practitioner is then left to decide whether a modified version of the research protocol will be effective in improving blood cholesterol profiles. It is also possible that the protocol used in the scientific experiment will result in undertraining in some individuals; that is, in contrast to the previous scenario, the intensity of the stimulus may be insufficient to induce a training response. The exercise practitioner must then decide how to effectively alter the training loads in order to induce a similar response in a client or patient with a greater initial fitness level. In situations like these, it is often helpful to consult the discussion section of the research paper. Often the authors discuss in great detail the programmatic elements that they believe are responsible for the particular adaptations they observed; for instance, they may note that their subjects completed 10 intervals (with positive outcomes) in contrast to a previous study in the same population that used four intervals with negative findings. Mechanistic evidence, from either the same study or another, would further elucidate the cause(s) for the positive adaptations from a higher volume of exercise.
In medicine, one of the often cited criticisms of the evidence-based philosophy is that it leads to a cookbook approach to practice (Sackett, Rosenberg, Gray, Haynes, & Richardson, 1996). Opponents assert that evidence-based practitioners read scientific papers and then implement the protocols directly from the papers. The example we have presented illustrates that there really is no "recipe" for prescription; that is not the purpose of science. Because people are different, protocols may need to be altered to fit individual patient or client needs. Science influences the "ingredients" of the prescription. This hypothetical experiment demonstrates that interval exercise might be more effective than steady-state treadmill exercise in improving blood cholesterol of patients who are obese. It also provides some basic parameters for intensity, volume, mode, and so on for the implementation of a protocol. Practitioners must then decide whether to implement the protocol. They must also decide how they should implement interval exercise in a manner that is best suited to their client's, patient's, or athlete's individual needs and experience level.
Learn more about Evidence-Based Practice in Exercise Science.