Potentially Preventable Hospitalizations Among Older Adults: 2010–2014

Elham Mahmoudi, PhD, MS; Neil Kamdar, MA; Allison Furgal, MS; Ananda Sen, PhD; Phillip Zazove, MD; Julie Bynum, MD


Ann Fam Med. 2020;18(6):511-519. 

In This Article

Abstract and Introduction


Purpose: We undertook a study to examine national trends in potentially preventable hospitalizations—those for ambulatory care–sensitive conditions that could have been avoided if patients had timely access to primary care—across 3,200 counties and various subpopulations of older adults in the United States.

Methods: We used 2010–2014 Medicare claims data to examine trends in potentially preventable hospitalizations among beneficiaries aged 65 years and older and developed heat maps to examine county-level variation. We used a generalized estimating equation and adjusted the model for demographics, comorbidities, dual eligibility (Medicare and Medicaid), ZIP code–level income, and county-level number of primary care physicians and hospitals.

Results: Across the 3,200 study counties, potentially preventable hospitalizations decreased in 327 counties, increased in 123 counties, and did not change in the rest. At the population level, the adjusted rate of potentially preventable hospitalizations declined by 3.45 percentage points from 19.42% (95% CI, 18.4%-20.5%) in 2010 to 15.97% (95% CI, 15.3%-16.6%) in 2014; it declined by 2.93, 2.87, and 3.33 percentage points among White, Black, and Hispanic patients to 14.96% (95% CI, 14.67%-15.24%), 17.92% (95% CI, 17.27%-18.58%), and 17.10% (95% CI, 16.25%-18.0%), respectively. Similarly, the rate for dually eligible patients fell by 3.71 percentage points from 21.62% (95% CI, 20.5%-22.8%) in 2010 to 17.91% (95% CI, 17.2%-18.7%) in 2014. (P <.001 for all).

Conclusions: During 2010–2014, rates of potentially preventable hospitalization did not change in the majority of counties. At the population level, although the rate declined among all subpopulations, dually eligible patients and Black and Hispanic patients continued to have substantially higher rates compared with non–dually eligible and White patients, respectively.


The Agency for Healthcare Research and Quality (AHRQ) has developed potentially preventable hospitalizations and other prevention quality indicators as measures of access to and quality of primary care.[1] These hospitalizations capture admissions for ambulatory care–sensitive conditions (ACSCs) that could have been avoided if patients had had timely access to primary care.[2] Potentially preventable hospitalizations are costly and negatively affect the health and well-being of individuals, particularly older adults.[3] By 2030, 1 in every 5 Americans will be 65 years of age and older. Changes in potentially preventable hospitalization rates may signal improvement or worsening in access to or quality of primary care among older adults.[4]

Prior research has shown increasing racial and socioeconomic gaps in rates of potentially preventable hospitalizations. For example, examining 2003–2009 trends, Mukamel et al[5] showed that rates declined among White patients but did not change among Black patients. Because of policy efforts in the last decade, hospitalization has consistently declined in the United States.[6,7] Since 2012, the Centers for Medicare & Medicaid Services have initiated a series of incentives to reduce potentially preventable hospitalizations.[8] Examination of trends in these hospitalizations among various subpopulations of older adults who may be at elevated risk for adverse health events and across counties can inform policies that target specific populations.

With use of 2010–2014 Medicare claims data, our primary aim was to examine recent national trends in potentially preventable hospitalization rates among subgroups of older adults in the United States. Our secondary aim was to examine trends in these hospitalizations at the county level across the country. We hypothesized a declining but different potentially preventable hospitalization trend among various socioeconomic and racial/ethnic minorities. Furthermore, we hypothesized that there would be large variations in the change of rates across counties.