Objective: To evaluate the Care Assessment Need (CAN) score as an accurate mortality prediction tool in Parkinson’s Disease (PD) to facilitate timely end-of-life care needs and hospice enrollment.

Background: PD imposes substantial physical, cognitive, and financial burden on patients and healthcare systems.^1-2 Despite palliative care needs comparable to or exceeding those of advanced cancer patients, individuals with PD are significantly less likely to receive timely hospice referrals.³ The heterogeneous course of PD limits prognostic accuracy, contributing to delayed goal-concordant care, higher hospice disenrollment rates, and lower enrollment during the final month of life when benefits are most effective.⁴ The CAN score, developed by the Veterans Health Administration (VHA), stratifies patients into mortality risk percentiles (0–99).⁵ We aimed to evaluate the CAN score’s predictive performance for 90-day mortality in PD to evaluate its effectiveness as a tool for timely palliative care and hospice enrollment.

Method: We performed a retrospective cohort study using the Corporate Data Warehouse (CDW), a nationwide 20+ year and 20+ million person VHA database. The population included a pre-validated cohort of patients with PD identified through 1)presence of ICD-10 code for PD, 2) two or more levodopa prescriptions, and 3) age ≥40 years. 37,306 patients were included in the study. Data collected included baseline patient characteristics (e.g. age), CAN 3.0 scores, predicted 90-day mortality, and logged death within 90 days. Analysis was done using Rstudio and differences between predicted and actual mortality were evaluated using a Chi-square test of proportions.

Results: CAN scores consistently underestimate mortality in patients with PD across all scores, with the greatest discrepancy among those with scores between 70–90. Among patients within five years of PD diagnosis and a CAN score of 70–90, there was a 34.3% under-estimation of predicted 90-day mortality. CAN score accuracy decreases as time from PD diagnosis increases, regardless of CAN score used.

Conclusion: These findings highlight the need for a revised scoring system incorporating PD–specific variables to improve mortality prediction and establish feasibility of a prognostic tool to support timely palliative care discussions and hospice referral. Further research is needed to understand why the CAN score under-estimates mortality in PD patients.

References: 1.Akbar U, McQueen RB, Bemski J, et al. Prognostic predictors relevant to end-of-life palliative care in Parkinson’s disease and related disorders: a systematic review. J Neurol Neurosurg Psychiatry. 2021;92(6):629-636. doi:10.1136/jnnp-2020-323939

2.Maclagan LC, Emdin A, Fu L, et al. Net Health System Costs of Parkinson Disease. Neurology Clinical Practice. 2025;15(1): e200371. doi:10.1212/CPJ.0000000000200371

3.Goy ER, Bohlig A, Carter J, Ganzini L. Identifying Predictors of Hospice Eligibility in Patients with Parkinson Disease. Am J Hosp Palliat Care. 2015;32(1):29-33. doi:10.1177/1049909113502119

4.Bock M, Gan S, Aldridge M, et al. Hospice Use Among Medicare Beneficiaries with Parkinson Disease and Dementia with Lewy Bodies. JAMA Netw Open. 2025;8(3): e250014. doi:10.1001/jamanetworkopen.2025.0014

5. Veterans Health Administration. (2024). Care Assessment Needs (CAN) Primer. Clinical Assessment Reporting, and Tracking Program. FY2024.

« Back to 2026 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/the-care-assessment-need-score-underestimates-90-day-mortality-in-veterans-with-parkinsons-disease/