Currently Viewing:
The American Journal of Managed Care June 2019
Reports of the Demise of Chemotherapy Have Been Greatly Exaggerated
Bruce Feinberg, DO; Jonathan Kish, PhD, MPH; Igoni Dokubo, MD; Jeff Wojtynek, PharmD; and Kevin Lord, PhD, MHS
From the Editorial Board: Patrick H. Conway, MD, MSc
Patrick H. Conway, MD, MSc
Association of Decision Support for Hospital Discharge Disposition With Outcomes
Winthrop F. Whitcomb, MD; Joseph E. Lucas, PhD; Rachel Tornheim, MBA; Jennifer L. Chiu, MPH; and Peter Hayward, PhD
US Care Pathways: Continued Focus on Oncology and Outstanding Challenges
Anita Chawla, PhD; Kimberly Westrich, MA; Angela Dai, BS, BA; Sarah Mantels, MA; and Robert W. Dubois, MD, PhD
Understanding Price Growth in the Market for Targeted Oncology Therapies
Jesse Sussell, PhD; Jacqueline Vanderpuye-Orgle, PhD; Diana Vania, MSc; Hans-Peter Goertz, MPH; and Darius Lakdawalla, PhD
Cancer Care Spending and Use by Site of Provider-Administered Chemotherapy in Medicare
Andrew Shooshtari, BS; Yamini Kalidindi, MHA; and Jeah Jung, PhD
Will 2019 Kick Off a New Era in Person-Centered Care?
Ann Hwang, MD; and Marc A. Cohen, PhD
Enhanced Care Coordination Improves HIV Viral Load Suppression Rates
Ross G. Hewitt, MD; Debra Williams, EdD; Richard Adule; Ira Feldman, MPS; and Moe Alsumidaie, MBA, MSF
Currently Reading
Impact of Care Coordination Based on Insurance and Zip Code
Jennifer N. Goldstein, MD, MSc; Merwah Shinwari, BS; Paul Kolm, PhD; Daniel J. Elliott, MD, MSCE; William S. Weintraub, MD; and LeRoi S. Hicks, MD, MPH
Health Insurance Design and Conservative Therapy for Low Back Pain
Kathleen Carey, PhD; Omid Ameli, MD, MPH; Brigid Garrity, MS, MPH; James Rothendler, MD; Howard Cabral, PhD; Christine McDonough, PhD; Michael Stein, MD; Robert Saper, MD, MPH; and Lewis Kazis, ScD
Improving Quality Measure Maintenance: Navigating the Complexities of Evolving Evidence
Thomas B. Valuck, MD, JD; Sarah Sampsel, MPH; David M. Sloan, PhD; and Jennifer Van Meter, PharmD

Impact of Care Coordination Based on Insurance and Zip Code

Jennifer N. Goldstein, MD, MSc; Merwah Shinwari, BS; Paul Kolm, PhD; Daniel J. Elliott, MD, MSCE; William S. Weintraub, MD; and LeRoi S. Hicks, MD, MPH
A care transitions program for patients who underwent percutaneous coronary intervention appeared to reduce 30-day rehospitalizations for patients with Medicaid who lived in wealthier zip codes.
ABSTRACT

Objectives: To examine whether a care transitions program, Bridges, differentially reduced rehospitalizations among patients who underwent percutaneous coronary intervention (PCI) based on insurance status and zip code poverty level.

Study Design: Retrospective observational cohort.

Methods: We examined data from a single health system in Delaware, collected as part of a care transitions program for patients who underwent PCI from 2012 to 2015 compared with an unmatched historical control cohort from 2010 to 2011. Socioeconomic status was assessed by insurance status and zip code–level poverty data. Patients were divided into tertiles based on the proportion of their zip code of residence living under 100% of the federal poverty level. Rehospitalization rates were analyzed by negative binomial regression and included interaction terms to examine differential effects of Bridges by insurance and poverty level.

Results: There were 4638 patients representing 5710 hospitalizations: 3212 in the historical control and 2498 in the Bridges cohort. Among patients with Medicaid who received the Bridges intervention, those living in the wealthiest zip codes were 15.5% less likely to be rehospitalized than patients with Medicare and 9.4% less likely than patients with commercial insurance (P = .04). However, patients with Medicaid who lived in the poorest zip codes and those with dual Medicare/Medicaid status had higher rates of rehospitalization post intervention.

Conclusions: The Bridges intervention was associated with improved rehospitalization rates for Medicaid patients compared with those with Medicare or commercial insurance within Delaware’s wealthier communities. Care transitions programs may differentially affect Medicaid patients based on the wealth of the communities in which they reside.

Am J Manag Care. 2019;25(6):e173-e178
Takeaway Points

Among patients who underwent percutaneous coronary intervention who received a care transitions intervention, patients with Medicaid who lived in the wealthiest zip codes were 15.5% less likely to be rehospitalized than their peers with Medicare and 9.4% less likely than their peers with commercial insurance (P = .04). Those who lived in the poorest areas had the highest rates of rehospitalization.
  • Care transitions interventions may be more effective in reducing rehospitalizations if targeted toward patients with Medicaid who live in areas with greater resources.
  • Alternately, rehospitalization rates may increase among patients with Medicaid who live in poorly resourced areas.
Since 2012, CMS has incentivized hospital systems to reduce 30-day rehospitalizations for patients with a history of acute myocardial infarction (AMI).1 Patients experiencing an AMI are medically complex,2 and their high level of medical acuteness is an important driver of high readmission rates. Furthermore, given the additional barriers that patients of low socioeconomic status (SES) face, they are at even greater risk for readmission3,4 and other adverse outcomes.5-8 In an effort to reduce unnecessary rehospitalizations, many health systems have implemented programs to help patients overcome barriers to care, including interventions to improve care transitions from hospital to home.9-12 However, a systematic review of care transitions interventions for patients with AMI found that readmission rates did not differ between those who were enrolled in care transitions programs and those who were not.13 Although patients are often selected for care transitions programs based on variables related to high risk of readmission, variation in the effects of these interventions due to these characteristics is often not examined.14 For example, although patients with a history of AMI are at higher risk of readmission and mortality if they live in high-poverty areas or have Medicaid insurance,4-7 few studies, if any, have examined whether implementation of a care transitions intervention can attenuate these risks. The objective of this study was to examine whether an intensive care transitions program was associated with a reduction in 30-day rehospitalization rates among patients who received percutaneous coronary intervention (PCI) based on individual and combined indicators of low SES.

METHODS

These data were collected as part of a quality improvement program that implemented an intensive care transitions intervention for patients who received PCI at a single regional health system in Delaware from 2012 to 2015. The intervention was called “Bridging the Divides” (“Bridges”) and was developed with a grant from the Center for Medicare and Medicaid Innovation. Details of the intervention have been described previously in the literature.15 Briefly, the intervention consisted of medication education for the patient from a clinical pharmacist and bedside nurse prior to discharge, biweekly phone calls during the first 2 weeks post discharge, and regularly scheduled calls from a care management nurse for up to 1 year if needed. As part of the program, called CareLink, the care management team helped to coordinate many aspects of postdischarge care, including making follow-up appointments, ensuring access to medications, and arranging transportation. During the postdischarge phone calls, the CareLink team assessed clinical issues (eg, clinical symptoms, medication adherence) and nonclinical issues (eg, barriers to obtaining follow-up care or medications). CareLink placed heavy emphasis on directing patients to appropriate sites for regular follow-up or urgent care and proactively identified issues that could lead to the need for emergency care.15 The program and approval to publish results of the study were obtained from the Christiana Care Institutional Review Board.

We collected data for consecutive patients who underwent PCI from 2012 to 2015 who were enrolled in the Bridges program (Bridges cohort); we compared this cohort with historical controls for patients who underwent PCI at the same institution from 2010 to 2011 (control cohort). Low SES was determined by insurance status and zip code–level poverty data obtained from the 2011 to 2015 estimates from the US Census Bureau.4,8,16 Patients were divided into tertiles based on the proportion of individuals within their zip code of residence who live under 100% of the federal poverty level (FPL). The 3 tertiles included the “wealthy” zip code (defined as those living in zip codes where <9.1% of individuals live under the FPL), the “moderate wealth” zip code (where 9.2%-13.1% of individuals live under the FPL), and the “poor” zip code (where >13.1% of individuals live under the FPL) tertiles.17 Insurance status was defined as Medicare, Medicaid, dual eligible (Medicare/Medicaid), commercial, or uninsured/other. Thirty-day rehospitalization was defined as a composite of inpatient, observation, and emergency department (ED) visits to the index site of service. We used a combined metric of rehospitalization because the Bridges intervention sought to reduce the need for all types of emergency care.

We then examined the unadjusted association between poverty level, insurance status, and the Bridges intervention with our outcome of 30-day rehospitalization using Poisson regression or negative binomial regression if the distribution of counts was overdispersed. For Poisson distributions, the mean and variance are equal; overdispersion occurs when the variance is greater than the mean. Analogous to an odds ratio in logistic regression, Poisson regression gives an incident rate ratio (IRR)—that is, a ratio of incident rates between categorical variables or increase (or decrease) for units of a continuous variable. Counts were analyzed because patients could have more than 1 rehospitalization.

We constructed a multivariable count model with poverty tertile as the primary independent variable and 30-day rehospitalization as the dependent variable. We controlled for insurance status, Bridges intervention, age, and all variables that were significantly different (P <.05) between the historical control and Bridges cohorts.3,18,19 To determine if there was a differential improvement in 30-day rehospitalization rates after the Bridges intervention among patients with combined indicators of low SES (Medicaid insurance/poor zip code), we tested the 3-way interaction (control vs Bridges, insurance status, poverty tertile) in the models. We then compared differences in predicted proportions of 30-day rehospitalization rates in the control and Bridges cohorts, stratified by poverty tertile and insurance status. All analyses were conducted in Stata version 14 (StataCorp; College Station, Texas).


 
Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
x
Welcome the the new and improved AJMC.com, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up