Enhancing community capacity to improve cancer care delivery and the effect on patient-reported outcomes, healthcare utilization and total costs of care.

Abstract

6522 Background: To curb rising expenditures and improve patient-reported outcomes (PROs), we designed an intervention with patient, caregiver, provider, and payer input. The intervention is based on prior work using a lay health worker (LHW) to assess advanced cancer patients' symptoms. In this study, we trained the LHW to refer patients to palliative care and/or behavioral health services in response to positive assessments and expanded the intervention to all cancer stages. We implemented the intervention with a health plan and a community oncology group serving elderly racially/ethnically diverse patients to test the effect on symptoms, healthcare use, and total costs. Methods: We enrolled all newly diagnosed health plan beneficiaries with solid and hematologic malignancies from 10/2016 to 11/2017 and compared outcomes to all cancer patients diagnosed in the year prior to the intervention (control arm). Our primary outcome was change in patient-reported symptoms using the Edmonton Symptom Assessment Scale and Personal Health Questionnaire-9 at baseline, 6 and 12 months post-enrollment. Secondarily, we compared 12 month healthcare use and costs. All generalized linear regression models were adjusted for age, stage, comorbidities, diagnosis, and follow-up. Results: 425 patients were in the intervention; 407 in the control. In both groups, mean age was 79 years; 48% were non-Hispanic White, 43% Hispanic, 3% Black, 6% Asian/Pacific Islander; 60% had advanced stages; 28% had breast, 28% had gastrointestinal, and 10% had thoracic cancers. Intervention patients had significantly decreased symptom burden over time as compared with the control (Mean Difference: intervention (-0.77 +/- 0.28 p = 0.01) vs. control: (0.45 +/- 0.25 p = 0.06)); difference in difference: (-0.68 +/- 0.25 p = 0.007)). Depression scores also significantly decreased over time among intervention patients as compared with the control (Mean Difference: intervention (-1.10 +/- 0.38 p = 0.04)) vs control: (1.21 +/- 0.34 p = 0.01); (difference in difference: -2.03 +/- 0.3 p < 0.001)). As compared to the control arm, intervention patients had lower inpatient admissions (0.7 vs. 0.5 p = 0.01) and emergency department visits per thousand patients per year (0.6 vs. 0.42 p = 0.02), and lower median total healthcare costs ($32,270 versus $25,512 p = 0.01). Conclusions: An LHW intervention significantly improved patient-reported outcomes and the value of cancer care delivery and may be a solution to improve burdensome and costly care for patients.