Quality Improvement Effort With Proactive Social Worker Communication to Enhance Palliative Care Access.

Purpose: This study evaluates a quality improvement project to improve access to specialty palliative care (PC) for individuals with advanced cancer.

Methods: A California health care organization implemented a social worker (SW)-led program to improve education about, and referrals to, PC for patients with advanced cancer at one clinic (intervention site), whereas three control sites continued usual care. We used a rigorous intention-to-treat difference-in-difference analysis to detect whether the intervention increased PC referrals (primary outcome) at the intervention versus control sites while accounting for preimplementation trends. We used multivariable logistic and Poisson regressions to compare primary and exploratory outcomes (advance care planning and end-of-life outcomes: PC visits before death, acute care use within 30 days of death, and systemic therapy within 14 days of death) in the preimplementation and postimplementation period.

Results: The analytic cohort included 555 patients (177 preimplementation, 378 postimplementation): 147 (26.5%) at the intervention site and 408 at control sites, with 312 (56.2%) deceased. At the intervention site postimplementation, 52 (51.5%) of 101 patients received the SW-led intervention. At the intervention site, PC referrals increased from 30.4% to 32.7% (v 26.7% to 25.3% at control sites). PC visits increased from 21.7% to 28.7% (v 21.4% to 22%), and among decedents, any PC visits increased from 53.1% to 75% (v 57% to 66.9%). In adjusted multivariable analyses, the only near-significant difference-in-difference PC effect was that decedents at the intervention site had higher odds of receiving PC before death (OR, 3.09 [95% CI, 0.89 to 10.67]; P = .07).

Conclusion: This SW-led intervention was not associated with significantly increased PC referrals. When translating evidence from trials to real-world quality improvement efforts, context and changes to the intervention may affect effectiveness.