Setting
The Richard Fine People’s Clinic (RFPC) is an academic primary care safety-net clinic at Zuckerberg San Francisco General Hospital and Trauma Center affiliated with the University of California, San Francisco. RFPC serves over 9,000 adults with high medical and social complexity. It serves a racially/ethnically diverse population: 42% Latinx, 23% Asian/Pacific Islander, 16% white, 12% Black/African-American, 1% Native Hawaiian/Other Pacific Islander and 7% other. Forty-four percent of RFPC patients speak a primary language other than English (including 25% Spanish, 7% Cantonese, 1.9% Tagalog, 1.3% Vietnamese and 1.3% Russian). Only 1% of RFPC patients have commercial insurance; the vast majority have Medi-Cal and/or Medicare, or Healthy San Francisco, an innovative city health plan designed to make health care available and affordable to uninsured San Francisco residents.
Provider Participants & Training
All PCPs at RFPC were invited to participate in the BIU pilot. Seven of 40 PCPs chose to participate: three MDs and four NPs. Each of these providers received a two-hour in-person training on evidence-based depression care, a depression care algorithm, the BIU pilot model of care, and the BIU referral process. After this training, PCPs could refer eligible patients to the BIU team.
Patient Participants
BIU participants were enrolled as part of a clinic quality improvement pilot between April 1, 2018, and September 30, 2018. Eligible patients were: (1) ≥ 18 years old; (2) diagnosed with major depression with a current PHQ-9 score ≥ 10; and (3) active patients of the seven RFPC PCPs trained for BIU. We excluded patients (1) with bipolar disorder, schizophrenia, or schizoaffective disorder, and (2) receiving hospice or palliative care, or (3) permanent nursing home residents. The principal investigator (LOF) reviewed referred patients to ensure they met inclusion criteria before formal BIU enrollment.
Control participants were drawn from the panel of the same PCPs who participated in the BIU intervention, but before BIU was implemented and before they received the BIU-depression care education. We selected this retrospective control group to avoid spillover effect. Using a random number generator, we selected patients who received usual care from the PCP in 2017 and met the remainder of the BIU participant inclusion criteria.
Development of the Intervention
Prior quality improvement efforts identified that baseline depression remission rates in RFPC were only 3% per year, which was below the 25th percentile for the HEDIS26 measure of depression remission. As a result, a multidisciplinary working group (including a PCP mental health champion, psychiatrist, clinic panel manager, and behavioral health clinician) was formed in 2018 to develop and pilot BIU. This model was based on four of the five core principles of collaborative care: 1) patient-centered team care with primary care and behavioral health collaboration; 2) population-based care using a registry to track enrolled patients with defined outreach intervals; 3) measurement-based treatment-to-target [using the Patient Health Questionnaire- 9 items (PHQ-9) with target of depression remission (PHQ-9 < 5)]; and 4) evidence-based care.17 Accountable care, the fifth core principle, was not incorporated because RFPC’s payment model does not allow for reimbursement to individual PCPs based on quality of care.
Intervention: The BIU Model
We developed and implemented an algorithm with timelines for appropriate patient follow-up and depression management (see Fig. 1). Key intervention components included:
-
Patient-centered team care: BIU team members (see Fig. 2) underwent a two-hour training on the BIU model and their expected roles. This team met every two weeks for one hour. During meetings, the team tracked depressive symptoms (using PHQ-9), discussed patient progress and treatment plans, reviewed adverse effects and treatment adherence, facilitated scheduling of additional visits as needed, and provided psychiatrist-generated depression treatment recommendations to PCPs via messages sent through the electronic health record (EHR) to support evidence-based depression care. The BIU team provided outreach to enrolled patients as clinically appropriate, to help with appointment attendance, treatment adherence and linkage to behavioral therapy. Notably, instead of hiring a new depression care manager, existing staff adopted CCM roles and responsibilities.
-
Population-based care: The BIU team tracked enrolled patients using a validated and free patient registry from the University of Washington Advancing Integrated Mental Health Solutions (AIMS) center, the academic center that developed and champions CCM.27 In preparation for the bi-monthly team meetings, the existing clinic panel/QI manager updated the patient registry with recent data about interim visits, medication changes, and PHQ-9 scores.
-
Measurement-based treatment-to-target: PHQ-9 scores were tracked for intervention participants at each provider visit. The clinic panel manager issued reminders to the primary care team to repeat PHQ-9 administration during upcoming primary care visits. The PCP mental health champion and psychiatrist issued treatment guidance via EHR messages with a goal of reaching depression remission (PHQ-9 < 5).
-
Evidence-based care: All referring PCPs received a treatment algorithm which included evidence-based guidelines for depression care, such as timing of follow-up, formulary medications available for patients and guidelines for behavioral therapy referral. This algorithm also included evidence-based guidance on measurement-based care, anti-depressant dosing, titration, augmentation and the importance of reaching the minimum effective antidepressant dose (referred to as target dose in our intervention) as defined by the American Psychiatric Association’s clinical practice guideline.28
Data Collection
Baseline and six-month follow-up data for all participants were extracted from the clinic EHR.
Outcome Measures
The primary outcome was depression remission, as defined by PHQ-9 < 5 within six months. Additional secondary outcomes included depression response (reduction of PHQ-9 score by ≥ 50%) and PHQ-9 score change (defined as the difference between baseline and follow-up PHQ-9 scores within six months of initial PCP visit).
We also collected information about adherence to pharmacologic treatment guidelines, including 1) antidepressant prescription; 2) medication adjustments (increased or augmented); and 3) time (in weeks) from initial visit until target dose was reached. Finally, we collected process metrics about care coordination by the team (see Fig. 1).
Statistical Analyses
We compared baseline characteristics, process outcomes, and depression care outcomes between control and intervention groups using two-sample t-test with unequal variances or Wilcoxon rank-sum test for continuous variables and Chi-square test or Fisher's exact test for categorical variables. Baseline and follow-up PHQ-9 scores were used as assessment points for score-based analyses. Score change was calculated by subtracting the follow-up score from the baseline score. An improved score was defined as having at least a one-point reduction in PHQ-9 from baseline. Patients who did not have a follow-up score were excluded from score-based analyses.
Because almost half (46%) of controls lacked follow-up scores, we compared baseline characteristics and scores between controls with and without depression care outcomes. Due to the limited sample size, we did not test the interaction effect between demographic variables and process outcomes. All tests were conducted at the 0.05 level of statistical significance. Due to a small number of patients for certain categories, we limited the analysis to patients’ gender (female or male) and language (English, Spanish, or other).
Analyses were performed using Stata/IC 15.1 (StataCorp, College Station, Texas). The study protocol was approved by University of California, San Francisco Institutional Review Board.