Implementation of Collaborative Care for Depression in VA HIV Clinics (HITIDES-H3)

Implementation of Collaborative Care for Depression in VA HIV Clinics: Translating Initiatives for Depression Into Effective Solutions (HITIDES)

HIV Translating Initiatives for Depression into Effective Solutions (HITIDES) is a team-based service to manage depression in Veterans Living with HIV (VLWH). This service is more effective for managing depression than the care VLWH usually receive and saves resources. HITIDES is also liked by HIV care providers and VLWH. Despite this, no VA clinics currently offer this service. This study examines two approaches to engage clinics with HITIDES, the resulting effects on VLWH, and the costs of these approaches. The first approach includes recruiting an HIV care provider at the site to help connect with the service and a network of providers to support this person. The second approach uses an additional external expert to facilitate these connections. Understanding how to connect Veterans to the HITIDES service will allow VA to improve depression care for VLWH and save VA resources.

Background: HIV Translating Initiatives for Depression into Effective Solutions (HITIDES) is a collaborative care intervention that adapts the primary care collaborative care model for depression treatment to HIV clinics. In a randomized controlled trial, HITIDES significantly improved depression symptoms for Veterans Living with HIV (VLWH) and delivered cost savings. However, no VHA HIV clinics have implemented HITIDES. The goal of this study is to support broad implementation of the HITIDES intervention by testing two appropriate implementation strategies: a clinical champion from each site who, with the help of a learning collaborative of peers, will work with local clinicians and leadership to implement the HITIDES intervention at their site with and without the assistance of external facilitation from an implementation expert. Significance/Impact: Preliminary work has been completed to identify implementation strategies acceptable to VLWH and HIV care providers; however, the relative effectiveness and cost of these implementation strategies is unknown. While the HITIDES depression care team (DCT) is housed off-site and can deliver services consistently with high quality and fidelity, the ability of the DCT to interface and engage with HIV care providers at sites is unknown. Additionally, the mediating effect of site-level implementation outcomes such as reach and adoption on effectiveness of the intervention is unknown. Because the DCT can provide services to multiple HIV clinics, a small-scale rollout of the intervention is needed before considering a national roll out. Innovation: This study employs an innovative hybrid study design to concurrently examine implementation and effectiveness outcomes. The use of implementation success as a mediating factor for intervention effectiveness is also novel. The relative ability of implementation activities to impact care for vulnerable populations is an area of research where little is known. VHA HIV clinics are an ideal test case for examining these questions because VLWH are a group where racial minority, low income, sexual minority Veterans are disproportionately represented. Specific Aims: 1) Determine, through a cluster-randomized controlled trial among VHA HIV clinics, the effect of adding external facilitation to an implementation strategy consisting of a site-level clinical champion and learning collaborative. 2) Determine the impact of HITIDES on changes in depression and suicidal ideation among HIV-positive Veterans receiving the intervention. 3) Estimate the budget impact of HITIDES implementation strategies by calculating the costs of each strategy. Methodology: The use of a hybrid type-3 effectiveness-implementation trial to examine the interaction of implementation and intervention effectiveness is an innovative methodology ideal for situations where the lack of robust evidence of effectiveness is coupled with a cost-saving intervention. This hybrid trial will use a cluster randomization of 8 VHA HIV clinics. These clinics will be chosen for balance and diversity of clinic characteristics and randomly assigned to one of the two implementation arms. Evaluation of each aim will use a mix of primary (e.g., QUERI-developed time-tracking tool) and secondary (e.g., clinical data warehouse) data. The investigators expect the clinical champion, learning collaborative, and external facilitation arm to be associated with greater reach and adoption; however, the clinical champion and learning collaborative alone arm is expected to be less costly. Next steps: The findings from implementation of the HITIDES intervention to 8 VHA HIV clinics will be used to inform selection of implementation strategies for a broad roll out in the future. Findings will be presented in cooperation with the investigators operational partner, VA HIV, Hepatitis C, and Related Condition Program to VACO and VISN leadership.

This is a two-arm randomized controlled trial (RCT) with four VA HIV clinics in each arm. Arm 1 receives a learning collaborative and clinical champions at each site Arm 2 receives these plus external facilitation. Learning collaboratives will be hosted separately for sites in Arm 1 and sites in Arm 2 to avoid contamination. Across all aims, there are research activities proposed to evaluate the quality improvement (referred to as "implementation strategy bundles"). All implementation strategy bundles are supported by educational materials about HITIDES and tools to engage Veterans in implementation activities.

A clinical champion is a local provider who serves as a liaison between local clinics and local or regional leadership or national program offices (e.g., HHRC) and advocates for the HITIDES intervention with their local peers. A clinical champion can provide ongoing promotion of and education about HITIDES and remind care providers of its presence and value either formally (e.g., presentations at staff meetings) or informally (e.g., individual conversations about HITIDES benefits). Clinical champions can also engage middle managers or local leadership to buy in to and support intervention uptake. Clinical champions may also work with Veterans in quality improvement efforts to garner their ideas for how to outreach to other Veteran patients, plan quality improvement cycles, or consult in an advisory capacity. Potential champions will be approached about participation after consultation with site HIV care providers and HHRC.

Learning Collaboratives are groups of providers or provider organizations and foster a collaborative learning environment to improve implementation of the clinical innovation (HITIDES). There are several approaches to this in the literature including peer consultation networks, online communities of practice, quality circles, and learning collaboratives. Groups will meet virtually using a wide variety of media. Participation will be monitored by research staff and recorded by participants in a time-tracking log but not controlled. The learning collaboratives will be comprised of a group of site clinical champions, HIV care providers, and other VHA HIV care representatives, potentially including Veterans who are engaged in implementation activities, to foster a learning environment to improve the implementation of HITIDES through consultation.

. External facilitation will be provided by a trained external facilitator who is given the HITIDES Operations Guide and Consumer Voice tools for Veteran engagement in quality improvement. Facilitation is a process to enable sites to increase uptake of HITIDES through supportive relationships and strategies to navigate the implementation process. Virtual external facilitation is a recognized modality of facilitation delivery in VHA. The external facilitator will work with the clinical champion, site personnel, and Veterans-not conducting research activities but engaging in quality improvement processes with these individuals. The addition of a facilitator allows problem identification and problem-solving, using processes based on supportive relationships.

Reach. Reach will be calculated as a proportion of eligible patients receiving the intervention at 12 months. The primary comparison of interest, and the one on which the study is powered, is the comparison of reach at 12 months in Arm 1 vs. Arm 2.

Reach. Reach will be calculated as a proportion of eligible patients receiving the intervention at 18 months.

Adoption will be assessed at 12 months. The investigators will calculate a proportion of providers adopting intervention recommendations.

The PHQ-9 is the depression module of the PHQ and has high sensitivity and specificity for a major depression diagnosis. Additionally, the PHQ-9, which has a score range of 0 to 27, provides a reliable estimate of depression severity. Scores of 5, 10, 15, and 20 correspond to mild, moderate, moderately severe, and severe depression, respectively. Each of the nine items of the instrument has choices ranging from "0=Not at all" and "3=Nearly every day." These choices apply to the question, "Over the past two weeks, how often have you been bothered by the following problems?" The list of problems includes common depressive symptoms such as "Little interest or pleasure in doing things." The DCM will conduct the PHQ-9 at baseline (PHQ-9 of at least 10 required to continue intervention) and at the 12 month follow-up call for HITIDES patients.

The Columbia-Suicide Severity Rating Scale (C-SSRS) is a valid and reliable measure of suicidal ideation that is sensitive and specific to identifying suicidal ideation and predictive of suicide attempts. (Posner, Brent, Lucas, & al., 2010) Moreover, the C-SSRS ideation and behavioral subscales are both sensitive to change over time. (Posner, Brown, Stanley, & al., 2011). The investigators will consider suicidal ideation reduction as a decrease of one or more in the C-SSRS score and suicidal ideation remission as a move from any nonzero score to zero. Veterans are asked 3 to 6 questions, depending on whether suicidal ideation is endorsed. The current VHA process for comprehensive suicide risk assessment is for providers to administer the PHQ-2 plus the PHQ suicidal ideation question, which, if positive, prompts the administration of C-SSRS. The investigators will code Veterans as not having suicidal ideation if the C-SSRS score equals 0.

The base-case analysis estimates the difference in illness costs before and after the intervention's implementation plus the cost of implementation. Marginal effects (ME) will be calculated for outpatient and pharmacy utilization and cost. Additionally, MEs will be calculated for each outpatient encounter and medication type. To estimate MEs between the implementation arms, the investigators will calculate two predicted values for each patient based on regression of case-mix variables at the site level. Only covariates found to predict dependent variables at the p<0.10 level in a bivariate analysis will be included in the regression. By calculating the difference between the predictions, the effect of the intervention on an individual patient can be estimated. Averaging this value gives an average ME for the intervention.

Adoption will be assessed at 18 months. The investigators will calculate a proportion of providers adopting intervention recommendations.

The PHQ-9 is the depression module of the PHQ and has high sensitivity and specificity for a major depression diagnosis. Additionally, the PHQ-9, which has a score range of 0 to 27, provides a reliable estimate of depression severity. Scores of 5, 10, 15, and 20 correspond to mild, moderate, moderately severe, and severe depression, respectively. Each of the nine items of the instrument has choices ranging from "0=Not at all" and "3=Nearly every day." These choices apply to the question, "Over the past two weeks, how often have you been bothered by the following problems?" The list of problems includes common depressive symptoms such as "Little interest or pleasure in doing things." The DCM will conduct the PHQ-9 at baseline (PHQ-9 of at least 10 required to continue intervention) and at the 18 month follow-up call for HITIDES patients.

Inclusion Criteria: The research team will work with VISN 10 to identify sites that have an adequately sized (i.e. greater than 20) population of Veterans Living with HIV, have adequate PHQ-2 screen data to assess depression prevalence, can identify a clinical champion for implementation activities, are willing to participate, and allow for diversity and balance of clinic characteristics across arms (e.g. rate of referral to specialty mental health for VLWH and presence of HIV-only specialty clinic versus broad infectious disease clinic). For criterion 5, these factors that are important will be determined with VISN 10 clinical leadership and the research advisory council. Randomization at the VAMC level has limited ability to completely balance observed and unobserved health-system factors. However, efforts to balance key site characteristics are important; therefore, site-level characteristics thought to impact implementation efforts were identified by building consensus among HIV, Hepatitis, and Related Conditions Program (HHRC) leadership and the HIV care and implementation experts on the research team. Data on the following observable site-level characteristics will be collected from these sites during Year 1 of the study: baseline PHQ-2 screen rate, clinic size (unique patients, provider full-time equivalents [FTE], change in enrollment over the past 3 years), and current care manager FTE. Based on the five criteria described above and the power calculation below, eight eligible sites will be selected. Sites will be grouped into four couplets based on balance of the identified site-level characteristics. Exclusion Criteria: -None.