The relationship between patient-reported distress and healthcare utilization in mNSCLC.

Abstract

e24138 Background: Distress has a negative impact on the patient experience of cancer. We previously demonstrated that patients with metastatic non-small cell lung cancer (mNSCLC) face significant and sustained distress during first-line treatment. However, the association between distress and unplanned healthcare utilization (HCRU) is not well understood. Here we examine this association. Methods: We conducted secondary analyses of data from 152 adult patients with mNSCLC treated at Duke Cancer Institute who were part of a retrospective chart review study (3/15-6/20). For the original investigation, demographic, clinical, and distress data were abstracted. NCCN Distress Thermometer (DT) scores were recorded at each clinic visit from start of first-line therapy to end of first year, change in therapy or death. The DT is an 11-point ordinal scale, with a 39-item Problem List, assessing overall distress. DT scores of > = 4 indicate actionable distress. Further HCRU data were analyzed, including unplanned hospitalizations, 30-day readmissions, length of stay, clinic visits, infusion center visits, emergency department (ED) visits, and death. Descriptive statistical analyses were performed for sources of distress and HCRU. To examine the association between distress and HCRU, we utilized an adjusted frailty model (for baseline demographics, metastatic site, and National Cancer Institute Index). This Cox proportional hazards model allowed for multiple HCRU events per patient and included actionable distress as a time-dependent covariate. Results: First, the proportion of actionable DT scores (423/1652, 25.6%) was strikingly high. Most DTs included report of at least one physical problem (n = 1593, 96.4%), regardless of distress level. However, emotional problems (e.g., worry, nervousness, depression) were reported much more frequently when the DT score was actionable (61.9% vs. 20.8%). At the patient level (n = 152), actionable distress was associated with increased HCRU. For example, a larger proportion of those with an average DT score > = 4 had at least one hospitalization (n = 22, 53.7%), infusion center visit (n = 11, 26.8%) or ED visit (n = 28, 68.3%) compared to patients with a non-actionable average DT score. Furthermore, HCRU was 54% more likely to occur after a patient reporting actionable distress compared to a patient reporting lower distress levels any time during treatment (adjusted HR 1.54; CI 1.20, 1.97). A one-point increase in distress was associated with an 8% increase in risk of HCRU (HR 1.08; CI 1.03, 1.13) at any point during treatment. Conclusions: Patients with mNSCLC experience high levels of distress, with actionable distress appearing to be driven significantly by emotional problems. Patient-reported distress is also associated with HCRU. Intervention development and testing is needed to address the unmet psychological needs in lung cancer care and their potential impact on unplanned HCRU.