Tailored Motivational Interviewing (TMI): A Pilot Implementation-Effectiveness Trial to Promote MI Competence in Adolescent HIV Clinics

Abstract

This brief report describes results of piloted Tailored Motivational Interviewing (TMI). Tailoring focused on site-specific training needs, target patient behaviors, and implementation facilitators and barriers that staff anticipated. Participating staff (N=31) at two adolescent HIV clinics completed a pre-training qualitative interview (N=27), and MI competency assessments based on three pre- and six post-training standard patient role-plays (N=27). Results included pre- to post-training MI competence improvement (t (153) = −4.13, p ≤ .001) and change in competency category distribution (X² = (2, N=155) = 15.72, p ≤ .001), providing initial support for the implementation of TMI in adolescent HIV clinic settings.

INTRODUCTION

Among youth living with HIV (YLH), motivational interviewing (MI) has contributed to HIV care retention and viral load reductions [1]. MI is included in clinical guidelines for HIV care and risk reduction [2], and it is the only evidence-based practice shown to be effective for youth at each point in the HIV prevention and care cascade [3]. What is not clear is how to deliver training in real-world settings such that it is acceptable to staff and feasible to implement. We pilot-tested Tailored Motivational Interviewing (TMI), a set of implementation strategies integrated into MI training to tailor it for adolescent HIV clinics in preparation for a full-scale effectiveness-implementation hybrid trial [3]. Effective intervention in real-world settings has faced significant barriers [4], and this is the first known implementation science study of MI in this setting [3]. The purpose of this study was to understand factors related to the acceptability of TMI skill development among staff in adolescent HIV clinics and explore the feasibility of the TMI approach in preparation for a large-scale study.

The focus of TMI was on developing key MI skills identified in our prior work to be associated with behavior change in youth with HIV [5]. The tailoring aspect of TMI was both part of the intervention itself, and also part of the training approach. Individual qualitative interviews with participating staff were conducted to identify individual and clinic training needs, patient characteristics and target behaviors related to prevention and treatment of HIV, as well as potential local barriers and facilitators of implementation and sustainment. This information was used to develop customized training strategies and content for a 12-hour TMI training workshop, as well as to inform approaches to staff training and assessment. For example, we offered flexible options for training, including small group training to avoid clinic closures, in-person and remote training options, and day, evening, or weekend training times. We also adapted training materials and skills practice to incorporate patient characteristics described by staff in qualitative interviews, and incorporated staff suggestions to promote training uptake and post-training sustainment (e.g., “We all have to do it. Leaders need to model using MI in the clinic.”).

Throughout this paper, we use “TMI” to refer to MI as we have customized it; that is, standard MI with an emphasis on certain skills associated with behavior change in the target population as well as integration of training strategies that are tailored for each clinic setting based on feedback from staff. Where we refer to MI rather than TMI in the context of competency, our intention is to avoid implying that an assessment measure has been adapted for TMI.

METHODS

This research was reviewed and deemed exempt by the Wayne State University Institutional Review Board. This study used a single-group quasi-experimental design with mixed methods assessments. Thirty-one participants from 2 adolescent HIV clinics enrolled (N=18 at Site A, N=13 at Site B). Of them, and 25 completed a qualitative interview and pre/post role play study measures, 2 participants completed just a qualitative interview, and 2 participants completed just role plays. Across both sites, a total of 27 participants completed a qualitative interview, and 27 completed role plays. Additionally, 2 enrolled and attended the training workshop, but did not complete an interview or pre/post role plays. Eligibility criteria included 4 hours per week of HIV clinical patient contact and ability to attend a 12-hour TMI workshop. Site leaders provided the study team with a list of eligible staff members for recruitment purposes. Study staff provided a project overview and information sheet to each site during a group meeting that was conducted in-person at one site and by conference call at the other site. Study staff members then contacted potential participants by email or phone to offer an opportunity to review study details individually. Participation was voluntary and completion of a qualitative interview or role play was considered consent to participate. Participants were predominantly female (92%), middle-aged (M=43 years, SD=12.5), and Black (38%), White (34.6%), and/or Hispanic/Latino (26.9%). Reported professions were medical providers (44%), support staff (32%), and mental health providers (24%).

This study includes the pre- and post-training periods of an TMI implementation intervention, focusing on the feasibility and acceptability of the intervention. Participants completed a 1-hour semi-structured interview based on the Exploration, Preparation, Implementation, and Sustainment (EPIS) framework [6]. Participants were asked to describe their organizational culture, prior experience with MI, and expected barriers and facilitators of using it. Interviewers recorded their impressions of interview themes immediately following each interview. Two coders conducted multi-step thematic analysis as outlined by Braun and Clarke [7] and resolved discrepancies to consensus.

The MI competency assessment protocol included three monthly role plays as baseline pre-training assessments of participant MI skills, and three monthly role plays followed by three quarterly role plays as post-training assessments of MI skills. We used a standard patient (“SP”) model with actors to conduct role plays by telephone in order to overcome TMI practice and assessment barriers [8]. For example, it was not practical to use audio recordings of real patient-provider interactions in clinic [8]. Additionally, using SPs allowed us to ensure that each interaction at each time point presented participants with the opportunity to demonstrate their MI skills in a context similar to that experienced by other participants at the same time point. Role plays were audio recorded and duration was expected to be approximately 15 minutes. Each of the SPs was presented to participants in two sequential role plays, once as a first visit (either a newly diagnosed or a newly transferred patient) and once as a follow-up to that visit. Participants were provided with guidance at the beginning of each role play, which included information about the time available to complete their role play, and SP profile information (i.e., name, age, race, gender, sexual orientation, medical history, reason for visit, medications, current CD4 and viral load, and target behaviors for the interaction). Target behaviors included the following: medication adherence/viral load reduction, clinic appointment attendance, reduction of risky behavior. Other patient characteristics, such as readiness for change, were discoverable by participants during the role play interaction and varied by SP and by visit (first visit or follow-up) for the same SP. MI information or coaching was not included in the role play or the guidance portion of the call.

MI competence was assessed with the MI Coach Rating Scale (MI-CRS), a 12-item scale developed using Item Response Theory methods [9, 10]. The instrument was designed to measure MI competence within community-based agencies and clinics for research or supervision-related purposes. In the original evaluation [10], the MI-CRS measured a single dimension of MI competence, the four-point ordered categorical rating scale (Poor, Fair, Good, Excellent) performed as intended, the items were well-targeted to the sample, and the resulting measurements had strong reliability (Rasch reliability = .88-.89, Cronbach alpha-equivalent reliability = .91-.93, reliability of session-level scores = .75-.83). Examples of MI-CRS items include, “The counselor supports autonomy of the client(s),” and “The counselor works to evoke client(s)’ ideas and motivations for change.” Role play sessions were rated by 2 trained coders with 10% co-coded to confirm interrater reliability based on total scores of the scale using a single-measurement, absolute-agreement, two-way mixed-effects ICC model (ICC >70%).

RESULTS

Twenty-seven participants were interviewed (N=15 at Site A; N=12 at Site B). Interviews were M=48 minutes, 21 seconds in length (SD=8 minutes, 46 seconds)¹. Coders identified barriers and facilitators of using TMI that were categorized into four major themes: clinic operations, organization and clinic culture, patient considerations, and training considerations. The themes are described in Table 1.

Table 1.

Thematic Analysis Descriptions of Facilitator and Barrier Categories

Themes	Descriptions
Caseload/Clinic Operations	• Ability to prioritize patients and maintain clinic flow with minimal interruptions, which may be related to caseload, facility hours of operation, staffing shortages, and appointment frequency, type, and duration. • Contributions of MI to efficient use of clinic time. • Resource availability to support MI use with patients, which may include time, finances, human capital (i.e., supervisors, internal facilitators, staff), physical resources (i.e., space, books, videos), or ongoing training and skill development resources.
Organizational/Clinic Culture	• Leadership/key stakeholder buy-in. • Level of institutionalized use and support for MI, which may include use of MI by others, encouragement and expectations for MI use, beliefs about who should use MI and its effectiveness, accommodating study participation time in work schedules. • Perceptions about the value of MI training and its impact on employee evaluations.
Patient Considerations	• Patient characteristics, acceptance/rejection, and response to MI, including responses related to starting to use it with established patients. • Expectations for patient outcomes related to using MI, including engagement, adherence, and health outcomes. • Individual patient needs and treatment priorities.
Training Considerations	• Staff perceptions and expectations of the MI training process, and responses to MI training, including timing and scheduling of training and other study activities, and whether training examples are realistic and applicable. • Trainer characteristics, efforts to understand staff needs and preferences, training logistics and format (especially related to training approach, training materials, and ability to model MI). • Staff buy-in vs resistance, prior experience with MI or other EBTs, level of comfort with skill development processes, and view of whether training is needed or will be helpful relative to other skills already used.

Two main themes emerged. First, predicted facilitators of program implementation were primarily related to organization and clinic culture. Support (“buy-in”) and skill modeling by peers and clinic leaders were viewed as necessary. Participants identified a need for strong leadership support of TMI that would translate into action, such as structuring activities to include TMI [i.e., team meetings, case conceptualizations, “lunch and learns” (brief training activities often conducted during lunch periods, also known as brown bag learning sessions)]. Several participants also commented on the importance of seeing their peers using it when communicating with patients and other staff. Institutionalized use of TMI was prospectively identified by participants to be a key facilitator. For example, when asked what factors might facilitate program implementation, participant responses included, “Encouragement from fellow colleagues, talking about [MI] among each other, and practicing it,” “Management has to set an expectation. It would be helpful to provide data on progress and foster a competitive spirit [among staff and clinics],” and “The biggest thing is making sure that there is a supervisory aspect to prevent drift over time.”

Second, predicted barriers were primarily related to clinic operations. Time was the most prevalent implementation concern, particularly with regard to scheduling study activities and perceived potential for disruptions to patient care as a result. Participants also identified a need to maintain their skills with access to training boosters and other resources, but expressed concerns about the availability of funding to sustain such access. For example, when asked to identify potential barriers to program implementation, participant responses included, “Scheduling and timing for some of the steps. [Our] previous experiences were focused on not disrupting clinic flow and ensuring that it could be implemented without disrupting time. Folks are getting a little nervous about time,” “[It] seems like a lot. [Staff are] busy and worried about fitting the timing in,” “Anything that has a major disruption to clinic hours will be a big deal,” and “Scheduling. The clinic recently went from two days a week to five days a week.”

Twenty-seven participants (N=16 at Site A and N=11 at Site B) completed MI competence assessments, resulting in N=155 MI-CRS ratings (66 pre and 89 post). Overall competency ratings increased from M=1.12 (SD=0.33) at pre- to M=1.46 (SD=0.60) post-workshop, t (153) = −4.13, p ≤ .001 Sites did not differ in post-workshop score, but Site A did have higher pre-competency (M=1.19, SD=0.40; M=1.03, SD=0.18, respectively), t (64) = 2.03, p < .05. Scores were categorized as “novice” (<2.5), “intermediate” (≥2.5 and <3.5) and “advanced” (≥3.5). Both sites showed significant change in distribution across competency categories, X² = (2, N=155) = 15.72, p ≤ .001. Pre-workshop, 87.9% of ratings were novice, 12.1% intermediate, and none advanced. Post-workshop, 59.6% of ratings were novice, 34.8% intermediate, and 5.6% advanced.

DISCUSSION

Qualitative interviews were used to explore staff expectations of TMI uptake prior to training and were used for tailoring purposes in combination with our prior work identifying key MI skills associated with behavior change in youth with HIV [5]. Interview participants expressed interest in the study activities and were helpful in providing specific information about their respective clinic settings to aid the tailoring process. This included information about anticipated facilitators and barriers of program implementation, which centered primarily on factors such as clinic operations, clinic culture, and logistics. The interviews were also intended to assess the initial feasibility and acceptability of a TMI implementation intervention prior to the completion of other study activities.

A strong majority (82%) of participants completed the full workshop. Pre- and post-training competency ratings using the MI-CRS indicated improvement in MI competence. A significant proportion of participants moved from “novice” to “intermediate” MI skill level following the workshop, providing initial support for the implementation of this training program in busy clinic settings.

This study is among the first to assess the feasibility and acceptability of implementing TMI training in HIV clinics, but is limited to only two participating sites. A comparison of TMI to established MI training approaches to evaluate relative advantages of the tailoring aspect of TMI over standard intervention approaches should be a focus of future research and a larger trial is underway to test the efficacy of TMI training.