Analytical Framework
Implementation fidelity has been defined as “the degree to which programs are implemented as intended by the program developers”(Dusenbury et al., 2003). The conceptual framework in Figure 1 adapted from Carroll et al. guided the assessment of implementation fidelity and its determinants in this study(Carroll et al., 2007). Implementation fidelity has four constructs of duration, frequency, coverage and content(Ibrahim Muhammad, 2019). The four determinants affecting implementation fidelity according to the framework are: intervention complexity, facilitation strategies, quality of delivery and participant responsiveness(Barat et al., 2001).
Facilitation strategies refer to the delivery mechanism employed towards the effective implementation of an intervention. These strategies include training and the provision of protocols and guidelines. Quality of delivery typically describes the way the health care providers deliver the intervention. In this case, receiving supportive supervision, coaching and mentoring was used as a proxy measure of the quality of delivery of the PMDT management guideline. Intervention complexity describes the sophistication of an intervention, as simple, detailed intervention are very much likely to be delivered with high adherence (implementation fidelity) as compared to complex, vague interventions. Participants' responsiveness refers to the extent of participants' responses to the program or their level of engagement by it (here, PMDT management guideline). This often could be inferred from both the recipient of the intervention and the personnel responsible for delivering the intervention. Table 1 summarizes these analytical constructs. The following characteristics of the healthcare providers were also included in the model: age, sex, work experience and professional care.
Study setting, design and population
The study was conducted at the Infectious Disease Hospital (IDH), Kano, Nigeria. The IDH is a public specialized secondary health care facility serving a population of about 1.5 million people and having a patronage of about 300 patients/day. The hospital is the only infectious diseases hospital in the whole of Northern Nigeria and serves as a referral centre for Tuberculosis (TB) and handles the treatment of a patient with Drug-resistant Tuberculosis. Hearing loss assessment for DR-TB patients is conducted in IDH as part of PMDT guidelines’ implementation. The implementation fidelity assessment was a cross-sectional study. All front-line health care providers that have been involved in the implementation of PMDT guidelines for at least six months before the enquiry in the facility were recruited. Staff that met the inclusion criteria but were unavailable for whatever reason during the period of the study were excluded. A total of 73 health care providers were successfully interviewed.
This study was guided by the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) checklist (Additional file)
Study questionnaire
Data were collected through an interviewer-administered questionnaire. This was developed based on the domains and constructs of the adapted implementation fidelity framework. A 5-point Likert scale of "5-Strongly agree, 4-agree, 3-neutral, 2-disagree and 1-strongly disagree" was used except for the facilitation strategy where Yes/No responses were used. The questionnaire was tested for reliability, as well as face and content validity in a pilot study. Implementation fidelity was measured by ten questions across its four constructs; facilitation strategy by six questions; quality of delivery by six questions; intervention complexity domain by three questions; and participants’ responsiveness by four questions. Other information collected is age, sex, professional cadre and duration of working in PMDT (work experience).
Data management and analysis
The Cronbach-alpha reliability test was used to confirm the internal consistency of the questions used in measuring implementation fidelity and for each of the determinants. The alpha values ranged from 0.50 to 0.86, indicating moderate to high internal consistency. This allows summation of the constructs to derive a composite score for each of the variables. The ten constructs of implementation fidelity had a minimum score of 10 and a maximum score of 50. Likewise, scores were computed for each of the four core determinants. The values were then converted to percentages for analysis. The relationship between implementation fidelity and the determinants was examined using a linear regression model in Stata 14.2 (Inc. USA).