Objective: Client characteristics prior to entry into treatment, such as socio-demographic, personal, or functional attributes, have been considered extraneous factors that threaten the validity of the intervention causal effects on the intended outcomes. Researchers control these factors, either experimentally or statistically, in an attempt to minimize their influence on outcome achievement. Experimental control is maintained by carefully selecting clients based on a set of rigorous inclusion and exclusion criteria, and/or by randomly assigning clients to the study groups. Statistical control is exerted by residualizing the effects of these factors on the outcomes before determining differences between groups. There is growing concern that the results obtained with such control do not address the question of primary interest to clinicians: Who would most benefit from the intervention. Answers to this question require the adoption of the perspective that client characteristics are substantive factors, and that their influence on the intervention effects is to be determined. This paper will review the role of client characteristics in intervention evaluation research, and will present methods for examining their influence.
Design and Method: The points of discussion are based on a comprehensive review of relevant methodological literature, and are illustrated with examples from published research reports.
Concepts or Issues: The logic underlying the data analysis in intervention evaluation research considers client characteristics as extraneous factors that threaten internal validity. This perspective stems from the uniformity assumptions stating that clients are assumed to be equal on all measured and unmeasured characteristics, and that all clients in the experimental group exhibit the same response to the intervention. These assumptions are simplistic, and do not hold in the real world. Further, there is increasing empirical evidence showing that the clients’ personal and health-related characteristics affect their decision to participate in research and to receive and/or adhere to treatment, as well as their response to treatment. Developing an understanding of which client attributes influence the intervention effectiveness is necessary to guide practice. Clinicians need to be equipped with the knowledge that informs them of who would most likely respond favorably to the intervention. Development of this type of knowledge requires a new perspective that acknowledges and allows examination of client characteristics influence. The perspective calls for the generation of a framework that proposes direct and indirect relationships among client characteristics, intervention, and outcomes, and for testing these relationships empirically, by using statistical approaches such as hierarchical linear models or structural equation models.
Conclusions and Implications: Examining, rather than controlling, the influence of client characteristics has some advantages in terms of enhancing the validity of conclusions, and of developing knowledge that can be useful to clinicians. Incorporating relevant variables in the statistical analysis increases statistical power, delineates their contribution to the observed favorable or unfavorable outcomes, and enhances the applicability of results in everyday practice. Researchers are encouraged to begin to explore the effects of client characteristics on the effectiveness of interventions. They can use different strategies in such investigations.
Back to Methodological Issues in Intervention Research
Back to The Advancing Nursing Practice Excellence: State of the Science