ISSUES IN INTERVENTION RESEARCH - MAINTAINING INTEGRITY

The need for intervention studies has been noted (Oberst, 1992). As nursing interventions are developed through qualitative or quantitative descriptive research, studies to test interventions need to be conducted with meticulous care (Stewart & Archbold, 1992). Researchers need to attend to maintaining integrity of the intervention in all phases of intervention studies. Without serious questioning at every juncture, and without close attention to detail, erroneous conclusions may be drawn. When an intervention is tested and no difference is found between groups, the first conclusion could be that the intervention was not effective. Other potential explanations include (a) an insufficiently operationalized intervention, (b) insufficient power from small sample sizes or less powerful analytic techniques, (c) too much heterogeneity within each group on the dependent variable, (d) lack of sensitivity of the measurement of the outcome variables, or (e) incorrect timing of the outcome measurements (Lipsey, 1990). If power, heterogeneity, sensitivity, and timing questions can be addressed readily, then the intervention itself needs to be evaluated. The first question to be addressed is whether the intervention was indeed administered. This question is easy to answer in studies where there is one intervention administered one time by one investigator. When interventions are repeated, administered in different settings over time, and involve more than one investigator, assessment must be made of the integrity of the intervention (Gilliss & Kulkin, 1991). A computer search of the Cumulative Index of Nursing and Allied Health Literature from 1983 to 1991 yielded no articles on intervention research in general with the use of key words as "intervention", "research", "integrity"; only specific intervention studies were found. Although research textbooks discuss the implementation of experiments, the testing of interventions in clinical settings is a special case of an experiment. When the test is conducted in the clinical setting, the complexities of care are added to the difficulties of conducting research. Few textbooks (Mateo & Kirchhoff, 1991) or articles (Tyler, Clark, Winslow, & White, 1990) are written on the research process allowing for the complications of fitting into the clinical setting. In this article, the authors will discuss issues in implementing an intervention study, using an example of one in which the entire nursing staff of a unit administered the intervention. The planning and problem solving needed to conduct a successful study will be detailed. This information can be used for (a) the clinical research process and (b) clinical evaluations after implementation of a change in practice. © 1994.