Home Engineering Behavioral Intervention Research: Designing, Evaluating, and Implementing
EVOLUTION AND HISTORICAL USES OF THE CONSTRUCT OF FIDELITY
The construct of “fidelity” has been a concern of, and an emphasis in, many fields of study over the past several decades. The construct has been advanced in separate, yet parallel, fashions within respective disciplines. It has long been a topic of discussion in educational research, where efforts have been directed toward translating and assuring scalability and quality of the replication of evidence-based programs in educational settings. Similarly, in program evaluation and implementation sciences, fidelity has been, and continues to be, a primary driver and main focus of research attention. Hence, different models for ensuring replication with fidelity have emerged from each of these respective fields of inquiry (Tomioka & Braun, 2013).
In behavioral intervention research, interest in fidelity has been highly influenced by developments in psychological research. Table 12.1 outlines the ways in which fidelity has been defined and differentially operationalized.
As shown, in the 1970s, researchers began to raise the critical issue of causality as it pertained to the outcomes reported for psychotherapeutic interventions. Variations and anomalies in the implementation of treatments brought to the forefront nagging concerns as to whether positive outcomes could be solely ascribed to the tested therapeutic intervention or to other observed and unobserved factors. Suspected confounding factors included, for example, the skill level of the therapist, dosage, participant readiness, and strength of the therapeutic alliance (Cook, Campbell, & Day, 1979; Sechrest & Yeaton, 1981; Yeaton & Sechrest, 1981). Unfortunately, early trials of psychotherapeutic interventions were not designed to disentangle these and other potential confounders from treatment effects to address these concerns.
Moncher and Prinz (1991) were among the first researchers to formally use the term “fidelity” in reference to treatment integrity or whether a given treatment was delivered as intended. They extended the scope of the construct to include the notion of treatment differentiation, arguing that it was not only important to ensure treatment integrity but also to clearly and demonstrably differentiate multiple treatments from one another. The goal of fidelity was to achieve transparency and to demystify what often was reported as the “black box” of an intervention.
Lichstein, Riedel, and Grieve (1994) followed with further refinements of this construct. They suggested that treatment integrity involved three critical components: delivery (whether the intervention is delivered as intended by interventionists), receipt (whether the study participant receives the intended intervention), and enactment (whether the participant uses or enacts the cognitive or behavioral skills imparted in the intervention). They illustrated these three components with the following scenario: On the basis of clinical guidelines, a nurse practitioner provides a patient with a prescription for hypertension medication; this is the delivery aspect of fidelity. Next, the patient must fill the prescription, which reflects evidence of receipt of treatment. Finally, after receiving the prescription, the patient must self-administer the prescribed medication, reflecting evidence of enactment of the intended intervention. In this scenario, to conclude with confidence that an observed positive change in the patient’s blood pressure is due to the prescribed medication, all three components need to be evaluated affirmatively. A deviation in delivery, receipt, and/or enactment from the protocol could impede benefits (Lichstein et al., 1994).
Alternately, deviations from the protocol may inadvertently contribute to a positive outcome. Consider this scenario. Perhaps the patient takes the prescription from the nurse but forgets or chooses not to have the prescription filled by a pharmacist; or conversely, perhaps the patient has the prescription filled but then decides not to take the medication. Instead, the patient chooses to use alternative strategies such as changing diet, initiating an exercise program, and/or practicing stress reduction techniques. These alternative practices, and not the medication,
may produce lower blood pressure. Without evaluating all three treatment components of fidelity, the nurse might conclude inaccurately that it was the hypertension medication treatment alone that positively lowered the patient’s hypertension level.
For each component of fidelity, Lichstein and colleagues (1994) emphasized several considerations. First, strategies need to be introduced that “induct,” or enhance, the probability of effective delivery, receipt, and enactment. For example, the use of detailed treatment manuals, checklists, and a standardized protocol for training and certifying interventionists can enhance the consistency and integrity of delivering an intervention. Second, both qualitative and quantitative measures can be used to document the extent to which each of the three components of fidelity is achieved.
One of the first large-scale, multisite behavioral trials to formally employ the Lichstein et al. (1994) model was the National Institutes of Health Resources for Enhancing Alzheimer’s Caregiver Health initiatives (REACH I, 1995-2001; and REACH II, 2001-2006). In REACH I, six sites tested a different novel caregiver intervention and developed specific strategies for enhancing and tracking the delivery, receipt, and enactment (Burgio et al., 2001). In REACH II, one complex, multilevel intervention was tested across five sites utilizing a shared fidelity plan. For both REACH I and REACH II, strategies to induct fidelity included, but were not limited to, the development and use of well-constructed manuals of procedures and detailed treatment manuals, training and certification of interventionists using active learning techniques (including role-play, demonstrations, monitoring delivery through direct observation and audiotaping, and coding of treatment sessions for level of protocol adherence), and supervisory sessions for course corrections and prevention of drift. The concentrated level of attention to fidelity efforts in the REACH initiatives set a high scientific bar for the conduct of caregiver intervention studies in particular that had not previously been achieved in this area (Burgio et al., 2001; Chee, Gitlin, Dennis, & Hauck, 2007; Gitlin et al., 2003). The level of fidelity rigor achieved in the REACH initiatives also brought into question whether the inconsistent findings reported previously for the initial wave of caregiver intervention research were due, in part, to inconsistencies in treatment implementation across sites and a consequence, at least in part, to the lack of attention to treatment fidelity (Callahan, Kales, Gitlin, & Lyketsos, 2013).
Building on these previous efforts, the Health Psychology Workgroup for the National Institutes of Health (NIH)-sponsored Behavior Change Consortium (Ory, Jordan, & Bazzarre, 2002) developed what is now considered the classic explication of fidelity (Bellg et al., 2004). This workgroup defined fidelity as reflecting two interrelated components: “methodological strategies used to monitor and enhance the reliability and validity of behavioral interventions” and “methodological practices used to ensure that a research study reliably and validly tests a clinical intervention” (Bellg et al., 2004, p. 443). The NIH workgroup definition integrated the approach defined by Lichstein and colleagues (1994) with trial design considerations. Thus, four integral aspects of fidelity are emphasized: (a) adherence to trial design protocols, (b) treatment delivery, (c) treatment receipt, and (d) treatment enactment. Attention to all four components has become the recommended approach in behavioral intervention research.
Since the NIH guidelines were proposed, additional fidelity definitions have been suggested, often reflecting specific research contexts (e.g., public health, nursing, worksite evaluation) (Prohaska & Peters, 2007; Santacroce, Maccarelli, & Grey, 2004; Strijk, Proper, van der Beek, & van Mechelen, 2011). For instance, Carroll and colleagues (2007) conducted a critical review of existing fidelity models and proposed a new conceptual framework. Their model suggested that intervention outcomes are dependent primarily upon adherence, and its subcomponents including intervention content, coverage, frequency, and duration can be moderated by such factors as intervention complexity, facilitation strategies, and quality of delivery. However, adherence alone may not capture the complexity of the intervention and interactions among core components of fidelity. This model was formally tested by Hasson (2010) using a multiple case study method. In an investigation of the implementation processes of three intervention studies conducted in complex health or social care environments, the framework was modified to include two additional moderating factors: context and recruitment (Hasson, 2010).
Gearing and colleagues (2011) evaluated 24 meta-analyses and review articles on fidelity published over the past 30 years and concluded that attention should be given to the four aspects suggested by the NIH workgroup: design, training, monitoring of intervention delivery, and intervention receipt. More recently, Tomioka and Braun (2013), in conjunction with Hawaii’s Healthy Aging Partnership, developed a four-step protocol for assuring replication with fidelity as follows: (a) deconstruct the program into its components and prepare a step-by-step plan for program replication; (b) identify agencies ready to replicate the program and sponsor excellent training to local staff who will deliver and coordinate it; (c) monitor the fidelity of program delivery using standardized checklists; and (d) track participant outcomes to assure achievement of expected outcomes. As the need for implementing evidence-based practices for health promotion continues, guidelines and protocols will become even more critical for the successful adaptation and replication of evidence-based approaches in diverse communities.
Although there is general agreement concerning the importance of fidelity in behavioral intervention research, there remains a lack of consensus concerning its definition, essential elements, assessment, and scope. Nevertheless, an important take-home point is that, as one develops an intervention, it is critical to consider fidelity. Behavioral intervention researchers have numerous emerging conceptual models and approaches to consider, with the models of Bellg and colleagues (2004) and Lichstein and colleagues (1999) now considered the classic approaches.
|< Prev||CONTENTS||Next >|