Evaluation of AI Cost Prediction Model to Enroll Patients in Complex Care Management Program

Evaluation of MA Proactive Care Program Using Cost Prediction Model With Randomized Waitlist

Currently, UCLA Health (specifically the Office of Population Health and Accountable Care, or OPHAC) runs a complex care management program called Proactive Care (goal is to reduce care utilization by providing personalized care navigation/case management). Every month, an AI Population Risk tool runs to identify around 250 of the 480,000 or so UCLA primary care patients, and RNs contact these 250 patients to enroll in Proactive Care. Starting in December 2024, OPHAC launched a new method of enrolling UCLA's Medicare Advantage (MA) patients into Proactive Care: an AI Cost Prediction model. The idea is the same-- the top 250 highest predicted cost patients will be enrolled in Proactive Care. The investigators will evaluate this model and subsequent enrollment into the program by randomizing the waitlist of MA patients waiting to enroll in Proactive Care, thereby creating a control group. The top 500 highest predicted cost patients will be identified each month, and following a 1:1 randomization, 250 will be contacted for enrollment and the rest will be put on a wait-list control group for 10 months unless otherwise requested by their provider to be enrolled in the Proactive Care program earlier.

Study Overview

• Status: Recruiting
• Conditions: Chronic Disease
• Intervention / Treatment: Behavioral: Complex care management program

• Study Type: Interventional
• Enrollment (Estimated): 5000
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Richard K Leuchter, MD; Phone Number: 310-267-9111; Email: rleuchter@mednet.ucla.edu
• Study Contact Backup: Name: William Turner, BA; Phone Number: 310-267-9111; Email: wturner@mednet.ucla.edu

Study Locations

• United States
  • California
    • Los Angeles, California, United States, 90095
      • Recruiting
      • UCLA Health
      • Contact: Richard Leuchter, MD; Email: rleuchter@mednet.ucla.edu

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• At least 18 years of age
• Enrolled in a UCLA Managed Care Plan
• Cost prediction model identifies patient as having high predicted costs over the next 12 months

Exclusion Criteria:

• Currently enrolled in any UCLA care management program
• Enrolled in any UCLA care management program in the last 12 months
• Already has an active referral to a care management program

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Prevention
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Double

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Active Comparator: Complex Care Management; Patients randomized to be contacted for enrollment into the complex care management program called ProActive Care.	Behavioral: Complex care management program; Intensive outpatient care management program that includes contact from nurses and case managers to help coordinate care, detect clinical red flags, and reduce overall unplanned acute care utilization.
No Intervention: Care as usual; Patient randomized to be put on a waitlist for being contacted for enrollment into ProActive Care (ie, not enrolled in ProActive Care).

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Days alive and out of hospital (DAOH) at 120 days from randomization	The sum of the number of days that a patient is not hospitalized under inpatient or observation status, and alive, out of a maximum of 120 days post-randomization.	120 days after randomization

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Days alive and out of hospital (DAOH) at 30 days from randomization	The sum of the number of days that a patient is not hospitalized under inpatient or observation status, and alive, out of a maximum of 30 days post-randomization.	30 days after randomization
Days alive and out of hospital (DAOH) at 90 days from randomization	The sum of the number of days that a patient is not hospitalized under inpatient or observation status, and alive, out of a maximum of 90 days post-randomization.	90 days after randomization
Days alive and out of hospital (DAOH) at 10 months from randomization	The sum of the number of days that a patient is not hospitalized under inpatient or observation status, and alive, out of a maximum of 300 days post-randomization.	10 months post-randomization
Total healthcare expenditures at 10 months from randomization	The sum of all healthcare expenditures (inpatient, outpatient, prescription, etc.) as determined by claims data.	10 months post-randomization
All-cause hospitalizations at 10 months from randomization	The total number of hospitalizations under inpatient or observation status for any cause	10 months post-randomization
All-cause emergency department visits at 10 months from randomization	The total number of emergency department visits that did not result in hospitalization, for any cause.	10 months post-randomization
All-cause mortality at 10 months from randomization	The total number of deaths from any cause	10 months post-randomization
Ambulator contact days	The number of days a patient spends outside the home receiving health care, defined as the total number of days with a primary care or specialty care office visit, test, imaging, procedure, or treatment	10 months post-randomization

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Cost prediction model performance over 22 months, as measured by area under the receiver operating characteristic curve	Observed healthcare expenditures for each month's control cohort will be measured over the subsequent 12 months after randomization (ie, 10 months of enrollment + 12 months of follow-up= 22 months), and compared to model-predicted healthcare expenditures. This analysis will only be conducted in the control group, as the intervention is designed to affect healthcare expenditures in the treatment group.	12 months after the final cohort enrolls
Protocol fidelity over the 10 months of enrollment	The sum of two metrics: The proportion of patients randomized to the intervention who enroll in the complex care management program; The proportion of patients randomized to the control who do not enroll in the complex care management program.	14 months post-randomization
Avoidable hospitalizations	The number of hospitalizations that could have been avoided through timely access to ambulatory care	10 months post-randomization
Avoidable emergency department visits	The number of emergency department visits that could have been avoided through timely access to ambulatory care	10 months post-randomization

Collaborators and Investigators

• Sponsor: University of California, Los Angeles
• Investigators: Principal Investigator: Richard K Leuchter, MD, University of California, Los Angeles

Study record dates

Study Major Dates

• Study Start (Actual): August 16, 2024
• Primary Completion (Actual): September 9, 2025
• Study Completion (Estimated): May 12, 2026

Study Registration Dates

• First Submitted: March 21, 2025
• First Submitted That Met QC Criteria: March 31, 2025
• First Posted (Actual): April 8, 2025

Study Record Updates

• Last Update Posted (Actual): January 29, 2026
• Last Update Submitted That Met QC Criteria: January 27, 2026
• Last Verified: January 1, 2026

More Information

Terms related to this study

• Keywords: High utilization; Complex care management; Cost prediction model
• Additional Relevant MeSH Terms: Pathologic Processes; Disease Attributes; Pathological Conditions, Signs and Symptoms; Chronic Disease
• Other Study ID Numbers: IRB-24-0849

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No