CLINICAL INVESTIGATION Cost-Effectiveness of a Care Transitions Program in a Multimorbid Older Adult Cohort Gregory J. Hanson, MD,* Bijan J. Borah, PhD, †‡ James P. Moriarty, MS, ‡ Jeanine E. Ransom, BA, § James M. Naessens, ScD, †‡ and Paul Y. Takahashi, MD* BACKGROUND/OBJECTIVES: Facing penalties for pre- ventable 30-day hospital readmissions, many provider groups have implemented programs to remedy this prob- lem, but the cost efficacy and value of such programs are not well delineated. The objective was to compare total cost of care over 30 days of individuals enrolled in the Mayo Clinic Care Transitions (MCCT) program and indi- viduals not enrolled. DESIGN: Retrospective cohort study using secondary data analysis of a previously published cohort study. SETTING: Mayo Clinic, Rochester, Minnesota. PARTICIPANTS: MCCT participants (n = 363) and indi- viduals in a propensity-matched referent cohort (n = 365). INTERVENTION: MCCT program enrollment. MEASUREMENTS: The primary outcome was total cost of care over 30 days after hospital discharge. A 2-part modeling strategy was used to analyze 30-day costs: whether individuals had non-zero costs during the 30 days after discharge and a generalized linear model for individu- als who incurred costs. Potential heterogeneous effects of the MCCT program were examined according to decile of 30-day costs using quantile regression. RESULTS: Mean age was 83 in both groups. Adjusted mean 30-day cost after hospitalization was $3,363 (95% confidence interval (CI) = $2,512À4,213) in the MCCT group and $4,161 (95% CI = $3,096À5,226) in the con- trol group (P = .25). Cost savings of $2,744 (P = .008) at the eighth decile and $3,388 (P = .20) at the ninth decile were demonstrated. Thus, the only statistically significant differences were in the post hoc subgroup analysis in the highest-cost subgroups. CONCLUSION: We did not find a difference in overall mean costs between the MCCT group and the control group, although intervention participants in the upper dec- iles of costs appeared to experience lower costs than con- trols. A larger study cohort might better determine the value of the intervention. J Am Geriatr Soc 2017. Key words: care transitions; costs; readmissions S ince the publication of the seminal article on Medicare hospital readmissions in 2009, 1 the Centers for Medi- care and Medicaid Services (CMS) has placed clinical focus on preventable 30-day hospital readmissions for targeted diagnoses through the Hospital Readmissions Reduction Program. As healthcare organizations prepare for payers’ increased focus on value-based purchasing, comprehensive and accurate cost analysis will become a necessary compo- nent of defining the value of care provided, specifically in terms of person-centered health outcomes per dollar cost. 2 Given the high costs of inpatient care, reduction of pre- ventable hospital admissions and readmissions can lead to significant cost savings. 3,4 Numerous interventions have been developed to address this for CMS beneficiaries. The efficacy of transition programs has been mixed, with some studies showing fewer readmissions and others showing no effect. 5–7 Varied program elements, duration, intensity, and target populations might explain the differ- ences in outcomes and measured costs. 6–8 There are rela- tively few recent total-cost-of-care analyses of programs in the United States, although the Transitional Care Model 9,10 and the Care Transitions Intervention 11 have demonstrated total-cost-of-care savings in older adults at high risk for readmission. A recent metaanalysis demon- strated net cost savings of $8,282 per person with inter- ventions that engage individuals and caregivers in general populations, but cost savings were not demonstrated in heart failure populations and general populations when the intervention did not include engagement. 12 These mixed results have not been fully explained. In our previ- ous work 13 with the Mayo Clinic Care Transitions (MCCT) program, we found a 38% relative risk reduction Frome the *Division of Primary Care Internal Medicine; † Division of Health Care Policy and Research; ‡ Robert D. and Patricia E. Kern Center for Science of Health Care Delivery; and § Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota. Address correspondence to Gregory J. Hanson, Division of Primary Care Internal Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905. E-mail: hanson.gregory@mayo.edu DOI: 10.1111/jgs.15203 JAGS 2017 © 2017, Copyright the Authors Journal compilation © 2017, The American Geriatrics Society 0002-8614/17/$15.00