Specify Potential Outcomes
Yet another important consideration in this prephase of discovery is determining appropriate outcomes and their measurement for the targeted effect an intervention may induce. Initially, it is helpful to consider a wide range of potential outcomes and/or measures that may be possible from intervening to address an identified problem area. For example, if quality of life is the primary outcome, then it is helpful to consider a range of approaches for measuring this construct to determine which domains are most responsive to the intervention. Similarly, an intervention that is designed to reduce depressive symptoms will obviously include an outcome measure of depression. However, there are various aspects of depression that could be measured including symptom severity or diagnostic category, and there could be various intervention goals including symptom reduction, remission, or an a priori defined clinically significant change in depression score.
On the basis of theory, empirical evidence, and the scope and nature of the problem, outcomes can be categorized in different ways as either proximal, primary, secondary, distal, or mediators of an intervention. Proximal outcomes reflect the immediate effects of an intervention, and that may need to be impacted to achieve a primary outcome of greater interest. For example, if one is testing the role of exercise in reducing hypertension, the first outcome of interest is whether participants engage in the exercise as prescribed. Thus, exercise engagement (dose and intensity) reflects the proximal goal of such an intervention without which other outcomes of interest, such as lowered blood pressure, may not hypothetically be able to be achieved. In this regard, proximal outcomes represent the enactment of strategies provided in an intervention and serve as one indicator of fidelity (see
Chapter 12) and may also represent potential mediators to understand the pathways by which an intervention works. Whereas exercise participation represents the proximal outcomes, of primary interest and public health import is a clinically significant reduction in hypertension. Thus, the reduction of hypertension reflects the primary outcome of interest. There may also be outcomes of secondary interest that may occur as a consequence of an exercise intervention; for example, participants may experience better mood, less fatigue, reduced fall risk, and better sleep. These all represent potential secondary outcomes that could be measured.
The ABLE intervention involved different components, one of which was conducting a home safety assessment and providing recommendations for reducing home hazards. The implementation of these recommendations reflected a proximal, immediate outcome that was expected and measured. This in turn was hypothesized to lead to improved home safety, a primary outcome measure. Thus, home safety was one of several primary outcomes of interest. Additionally, in the ABLE Program, other strategies were introduced to support the safe performance of participants in daily activities of their choice. Thus, another primary outcome was a meaure of functional difficulty; a secondary outcome was self-efficacy managing daily self-care challenges.
Distal outcomes refer to those that the intervention may impact if proximal and primary goals are achieved. These outcomes may occur over time as a consequence of using the intervention strategies. In the ABLE Program, distal outcome measures included quality of life and mortality. It was hypothesized that reduced functional difficulty (primary outcome) and improved home safety (primary outcome) would lead to enhanced self-efficacy and quality of life (secondary outcomes) and reduced risk for mortality (distal outcome).
As one can see, the identification of proximal, distal, and primary and secondary outcomes also suggests potential causal mechanisms that can be tested formally using mediation analyses. So, for example, a mediator of the ABLE effect on mortality may have been a reduction in functional outcomes and the use of positive compensatory strategies (Gitlin et al., 2009).
Specifying outcomes, their respective measurement, and role (primary, secondary, mediator, etc.) for an identified problem is important at this prephase as subsequent evaluations can then evaluate which outcome is most appropriate to consider and which measures are most sensitive to the intervention. Chapters 14 and 15 explore in more depth measurement considerations and intervention outcomes.