Treatment integrity: Some persistent concerns and some new perspectives.

Abstract

Treatment integrity, or procedural fidelity measures, is a critical yet often overlooked aspect of clinical interventions, research endeavors, and staff training within the field of applied behavior analysis. Yet, a relative paucity of data exists in the literature on treatment integrity. Treatment integrity typically refers to the correct delivery of the independent variable (e.g., therapist prompts or reinforcer delivery) or to staff training issues concerning the delivery of the independent variable (Peterson, Homer, & Wonderlich, 1982). ********** Roughly 20 years ago, two conceptual papers on treatment integrity were published (Peterson et al., 1982; Yeaton & Sechrest, 1981). Gresham, Ganslie, and Noell (1993) more recently reported on treatment integrity issues for articles in the Journal of Applied Behavior Analysis (JABA) with children as the participants. Our paper seeks to update practitioners and researchers on the current status of treatment integrity. The paper is divided into four sections. In Section 1, data from a sample of the first 5 years of JABA and the most recent 5 years of JABA will be reported. In Section 2, a brief review of the literature will be provided with a special emphasis on treatment integrity as a research issue in its own right. In Section 3, an overview of a treatment integrity model developed at Bancroft NeuroHealth will be discussed. Finally, in Section 4, a summary and interpretation will be provided, with a cautionary, yet optimistic prescription offered. REVIEW OF JABA Two observers independently rated a selection of JABA articles. The selection included the first five years (Volume 1 No.1 through Volume 5 No. 4) and the last five years (Volume 29 No.1 through Volume 33 No. 4). The primary observer rated every year and the reliability observer rated 3 years (30%). All research articles and reports were included in the rating. All technical articles, conceptual articles, and book reviews were excluded. Each research article and report were rated for whether the independent variable in the experiment was measured. This included any article where the author mentioned treatment integrity, independent variable integrity, or that observers collected data on the independent variable. Articles were counted regardless of whether independent variable data were actually reported in the article. Table 1 presents the results of the review. While the number of articles reporting treatment integrity has increased in the past 5 years, the total number of articles has also correspondingly increased. The percentage of articles mentioning treatment integrity is presented in Figure 1. The data indicate that over the past three decades there has not been an increase in the percentage of articles that include a measure of the independent variable. These data are slightly different from the results reported by Peterson et al. (1982), possibly due to a change in operational definition. Also, in the second reporting period, JABA distinguished between full-length articles and reports. It is possible that treatment integrity issues are omitted in reports due to space limitations. [FIGURE 1 OMITTED] CURRENT RESEARCH ON TREATMENT INTEGRITY While the data in the previous section suggests that treatment integrity is reported in approximately the same percentage of articles in JABA from 1996-2000 as it was from 1968-1972, significant changes in the approach to treatment integrity have occurred. That is, recent research on treatment integrity represents a new, more pragmatic, and data-based approach. For example, Northup et al. (1994) closely monitored treatment integrity for local school staff responsible for the implementation of behavioral interventions. They found that treatment integrity varied widely across individuals, yet treatment outcomes were quite successful. Others have examined the training of therapists in functional analysis methodology (Iwata et al. …