Clinical Research Network (CRN) update #13--choosing the appropriate comparison group: considerations from PTClinResNet.

The barriers to conduct randomized clinical trials (RCTs) are not unique in rehabilitation. However, there are unique issues related to the development and implementation of trials in rehabilitation. For example, multidisciplinary or multifocal interventions, which are characteristic of rehabilitation practice, are necessary when dealing with patients with complex problems such as the disabling consequences of cerebral stroke, spinal cord injury, cerebral palsy, or low back pain. Also unique to rehabilitation is the "activity trial"--a trial in which the interaction between clinician and patient and the patient's engagement in the activity are critical. The 4 RCTs hosted by PTClinResNet are defined as "activity trials." In these trials, the behavior, attention, and thought of the individual areof extreme importance. Exercise and learning involve activity and participation, and, therefore, rely on complex interactions between the patient and clinician.

These activity trials pose yet another challenge--choosing the appropriate comparison group. In large part, because of the concept of clinical equipoise, it would be unethical to deny adaptive experiences and activities to participants in a comparison control group. Clinical equipoise means a genuine uncertainty on the part of the expert medical community about the comparative therapeutic merits of each arm of a clinical trial. Clinical equipoise provides a clear moral foundation to the requirement that the health care of subjects not be disadvantaged by research participation. (1)

For the 4 PTClinResNet RCT trials, it was important to provide the strongest possible basis for a causal inference so that the observed results of activity interventions could be attributable to that intervention and not to other factors. Fuhrer suggests that, "particularly strong causal inferences can be drawn from the randomized clinical trial in which participants are assigned on a random basis to either the intervention of interest or to a comparison condition." (2(pS8))

In other words, the only difference between the treatment groups is the activity or intervention, with all other factors (eg, patient characteristics, motivation, etc) being equal across groups. This, in turn, requires knowledge of the factors that are likely to predict the chosen outcomes, regardless of treatment. Whyte and Hart (3) call that the "active ingredient" of a complex intervention. Finally, and consistent with that knowledge, the choice of comparison group must be appropriate for the level of prior evidence and activity-specific research. For example, in the case of the Strength Training Effectiveness Post Stroke (STEPS) trial, a no-treatment control condition would not have met the criteria for clinical equipoise, particularly given prior evidence supporting a benefit from locomotor training after stroke. Instead, prior evidence supported a focus on the specific active ingredients of locomotor training by identifying an appropriate participatory comparison condition (ie, upper-extremity exercise), choosing paired interventions, and using a multiple group design. We outline the characteristics of the comparison/control conditions for each of the four RCTs and highlight the conceptual rationale in the summary Table.

References

(1) World Medical Association. Declaration of Helsinki. 1964, revised 1996, para. A.3. Available at: http://www.wma.net/e/policy/b3.htm.

(2) Fuhrer MJ. Overview of clinical trials in medical rehabilitation: impetuses, challenges, and needed future directions. Am J Phys Med Rehabil. 2003;82(suppl): S8-S15.

(3) Whyte J, Hart T. It's more than a black box: it's a Russian doll. Am J Phys Med Rehabil. 2003;82:639-652.

[DOI: 10.2522/ptj.2007.87.5.619]

Table.
Comparison/Control Group for each PTClinResNet Randomized Controlled
Trial

RCT (a)    Control/Comparison      Activity Intervention
           Condition               Condition(s) (b)

STEPS      Sham condition-upper    BWST + Sham
           extremity ergometry     BWST + MSST
                                   LBST +Sham
                                   BWST + LBST

PEDALS     No exercise             Stationary cycling (LBST +
                                   cardiorespiratory exercise)

MUSSEL     Post-surgical patient   Muscle-specific strength and
           education               endurance training

STOMPS     Patient education       Shoulder muscle strength exercises
                                   + task-specific movement strategies
                                   home program

RCT (a)    Prior                   What Dimension
           Evidence                Was Controlled?

STEPS      Yes                     Time on activity,
                                   specificity/active
                                   ingredients

PEDALS     No                      Specific activity
                                   intervention

MUSSEL     No                      Specific activity
                                   intervention

STOMPS     No                      Specific activity
                                   intervention

(a) STEPS=Strength Training Effectiveness Post-Stroke, PEDALS=Pediatric
Endurance Development and Limb Strengthening, STOMPS=Strengthening and
Optimal Movements for Painful Shoulders in Chronic Spinal Cord Injury,
MUSSEL=Muscle-Specific Strength Training Effectiveness Post-Lumbar
Microdiscectomy.

(b) BWST=Body-weight-supported treadmill walking program,
LBST=locomotor-based, limb-loaded cycling strength training;
MSST=lower-extremity muscle-specific strength training, Sham=upper
extremity ergometry.