Local Supply Of Postdischarge Care Options Tied To Hospital Readmission Rates

Rehospitalizations after an index hospitalization are associated with increased patient morbidity, mortality, and health spending¹ and are increasingly viewed as a quality-of-care indicator.² The high human and financial costs of readmissions spurred Congress to authorize the Hospital Readmissions Reduction Program (HRRP) in 2010.³ Hospitals with thirty-day readmission rates above risk-adjusted national averages may face penalties in the form of reductions in future Medicare payments.

The HRRP’s risk-adjustment algorithm accounts for variation in hospitals’ volume and case-mix, as well as patient-level risk factors for readmission (age, sex, and selected clinical comorbidities).⁴ As of 2019 the Centers for Medicare and Medicaid Services (CMS) stratifies hospitals into five peer groups on the basis of their proportion of patients dually enrolled in Medicare and Medicaid and then calculates hospitals’ risk-adjusted readmission rates using a three-year average.^5,6 However, the HRRP does not account for differences in health system and community characteristics. Approximately seventy-eight million US citizens reside in areas designated by the Health Resources and Services Administration (HRSA) as having shortages of either health care providers or facilities,^7,8 and rural areas are disproportionately represented.⁹ Access to appropriate postdischarge care has been identified as an important determinant of hospital readmissions, and patients with poor access to outpatient medical follow-up may experience a higher risk for readmission.¹⁰ Hospitals in these underserved areas could experience higher thirty-day readmissions if they were unable to make referrals to adequate postdischarge care in their communities. The degree to which readmission rates (and thus HRRP penalties) are driven by differences in community characteristics is unknown.

Prior research has demonstrated a link between patients’ readmission risks and the availability of office-based primary care physicians, ratios of primary care physicians to specialists, and specialist availability.^11–13 Evidence related to the availability of nursing homes and skilled nursing facilities (SNFs) has been mixed.^11,14 Research on the community-level relationships between readmissions and the local availability of home health care and nurse practitioners has not been conducted to our knowledge.

In this study our objectives were to assess whether hospitals’ thirty-day readmission rates were associated with geographic variations in access to postdischarge care supply such as primary care physicians, nurse practitioners, SNFs, licensed nursing home beds, palliative care services, and home health agencies. Previous studies, although valuable, often focused on the local effects of community characteristics on a single hospital or health system,^15,16 a single year of cross-sectional data,¹¹ or a limited range of postdischarge care options^17–19 or did not account for the federal policy environment (for example, pay-for-performance programs), which may affect hospital readmissions.¹³ In contrast, we used nationwide panel data on hospitals to exploit variation over time in these measures, explicating the population-level effects of a wide variety of postdischarge care options on hospital readmission rates. We accounted for multiple factors that may confound the relationship between postdischarge care supply and readmissions such as demographics, hospital characteristics, and pay-for-performance program incentives, which have been associated with both readmissions and access to care.^20–23 If found, a robust association between postdischarge care supply and readmission rates may suggest that pay-for-performance programs such as the HRRP may be rewarding or penalizing hospitals in part based on the characteristics of communities they serve and not solely the quality of care provided.

Study Data And Methods

Data Sources

Data on hospitals’ thirty-day readmission rates were obtained from CMS’s Hospital Compare 2013–19.²⁴ Hospital Compare is publicly available and contains annual hospital-level thirty-day readmission rates for HRRP-eligible conditions. Supplemental hospital institutional and payment characteristics were drawn from the CMS’s hospital inpatient prospective payment system impact files and the American Hospital Association’s (AHA’s) Annual Survey database.²⁵ The CMS impact files contain data for a variety of hospital payment adjustments used by CMS, including readmissions penalties under the HRRP.²⁶ The AHA database contains data on ownership arrangements, safety-net status, number of beds, and availability of palliative care services.

Annual county-level data on demographics, insurance coverage, and postdischarge care supply, including health workforce and health-related infrastructure, were obtained from HRSA’s 2013–19 Area Health Resources Files.²⁷ CMS’s Hospital Service Area files are summaries of calendar year Medicare inpatient hospital fee-for-service claims data and contain the number of discharges by hospital provider number and the ZIP Code Tabulation Area of the Medicare beneficiary;²⁸ these were used to define each hospital’s “catchment area” from which their patients were drawn. We first assigned ZIP Code Tabulation Areas to counties using the Census Bureau’s relationship files.²⁹ We then calculated a weighted average of the Area Health Resources Files variables for each hospital according to the proportion of their Medicare inpatient discharges that were attributable to each county in a given year.

Key Outcomes And Explanatory Variables

Our primary outcome was hospitals’ annual crude readmission rates reported in Hospital Compare, stratified by condition. We focused on readmission rates for acute myocardial infarction, heart failure, and pneumonia, as these were the original three conditions targeted by the HRRP and thus are available in Hospital Compare starting in 2009. CMS excludes index stays where the patient does not survive until discharge and uses three previous years of data to generate hospital crude readmission rates, with approximately a one-year lag. For example, crude readmission rates for fiscal year 2018 are calculated using hospital performance data from June 2013 through July 2016.

Our key predictors of interest included several measures of postdischarge care supply within each hospital’s catchment area: per capita counts of home health agencies, nurse practitioners, primary care physicians, licensed nursing home beds, and SNF beds. For primary care physicians, we limited our analysis to nonfederal, office-based general practice physicians. This definition focuses on those physicians who are most likely to provide postdischarge follow-up care and excludes others who are hospital based, employed by the Department of Veterans Affairs, or serve in research or administrative roles. In addition, we controlled for the availability of hospice services within the hospital catchment and included a dummy variable taking on a value of 1 if the hospital had access to a palliative care service either internally, within the same health system, or through a joint venture. Hospice and palliative care may reduce unwanted life-sustaining treatment at the end of life and have been associated with lower readmissions in both observational and experimental studies.^30,31

In addition to postdischarge care supply measures, our adjusted models controlled for demographic variables from the Area Health Resources Files, including the percentage of residents in the hospitals’ catchment who were in poverty, unemployed, uninsured, or enrolled in Medicare.^32,33 Institutional factors included hospital ownership status (private, investor-owned; private, not-for-profit; and government), safety-net status (a binary indicator taking on a value of 1 if the hospital was in the top quartile of disproportionate share hospital patient percentage), number of licensed hospital beds (1–199, 200–399, and 400 or more), and within-hospital crude death rates.^22,34 We included two HRRP performance variables: excess readmission ratios comparing each hospital’s readmission rate relative to a risk-adjusted national average and a dummy variable for whether the hospital received a penalty for any condition in the previous year.^22,34 We controlled for a variety of CMS adjustments to hospital payments: case-mix index (representing the clinical complexity and resource needs in the hospital’s patient population),³⁵ disproportionate share hospital patient percentage (used to adjust payments to hospitals that serve a high volume of low-income patients),³⁶ the percentage of operating payments that are outliers (high-cost cases qualifying for additional reimbursements),³⁷ geographic adjustment factor (reimbursement changes based on regional differences in estimated operating expenses),³⁸ local wage index (used to adjust labor share of reimbursements),³⁹ and value-based purchasing adjustment factor (payment adjustments based on observed inpatient care quality).⁴⁰ Last, we included year fixed effects.

All continuous variables were Z-scored because of their widely varying ranges; coefficients should be interpreted as the expected percentage point change in hospitals’ crude readmission rates resulting from a 1-standard-deviation increase in each predictor.

Statistical Analyses

Our unit of observation was the hospital-condition-year, and our analysis proceeded in four steps. We first characterized the availability of postdischarge care options at the county level, elucidating geographic disparities. We then estimated bivariate linear regressions to assess how hospital readmission rates varied according to local availability of postdischarge care options. Next we estimated an adjusted mixed effects linear regression model with hospital random effects. The postdischarge care supply for a given hospital’s catchment area exhibited limited changes during the study period, precluding the use of hospital fixed effects. Therefore, this analysis exploited variation over time both within and between hospitals. Last, we conducted several sensitivity analyses including alternative weighting of the postdischarge care supply measures by total county-level discharges instead of population, removal of hospital random effects, or limiting our palliative care variable to only measure whether this service was hospital based. For each step, we estimated a pooled model including all three conditions and condition-specific models.

The study was deemed non-human-subjects research by the Boston University Institutional Review Board. All analyses were conducted using Microsoft R Open, version 3.5.3.

Limitations

Our study design was observational. Our results should be interpreted as associations, and we were unable to identify the specific causal mechanisms that underlie these relationships. Although we controlled for within-hospital death rates, the geographic distribution of postdischarge care options may partially reflect the illness severity of patients who were discharged in those areas. The postdischarge care supply for a given hospital’s catchment area was generally stable during our study period (online appendix exhibit A1),⁴¹ and thus we could not repeat our models using hospital fixed effects. In addition, some of the effects we found may have been driven by patient-level differences that were unrelated to local postdischarge care supply, including marital or partnership status or availability of a family caregiver. We included postdischarge care supply in our models instead of the use of postdischarge care options, and thus we did not expect that individual differences drove our results. Our postdischarge care supply measures did not include information on the quality of these facilities or providers or their actual availability for discharged patients. Furthermore, it is difficult to identify indicated versus potentially avoidable readmissions.⁴² We could not rule out the possibility that increased readmission rates in areas with a greater supply of nurse practitioners, for example, were due to increased identification of potentially dangerous complications that would otherwise remain untreated. Last, the Area Health Resources Files data allowed us to identify office versus hospital-based physicians, but only a single combined count of nurse practitioners. Thus, this measure did not necessarily reflect the supply of nurse practitioners available for postacute care.

These limitations notwithstanding, our longitudinal design allowed us to estimate how readmission rates related to variation both within and between catchment areas in postdischarge care supply, and it controlled for a wide range of potential confounders including demographics, institutional characteristics, and CMS payment adjustments.

Study Results

Our final sample included 50,592 hospital-condition-years from 3,042 unique hospitals (exhibit 1). The sample’s institutional characteristics were representative of US hospitals in general;^43,44 most hospitals were private, not-for-profit; were located in urban areas; and operated a hospital-based palliative care service. The median hospital had fewer than 200 beds and an average daily census of fewer than 100 patients. Mean hospital-level crude readmission rates during the study period were 18.7 percent for acute myocardial infarction, 22.4 percent for heart failure, and 17.5 percent for pneumonia (data not shown).

Geographic Availability Of Postdischarge Care Options

Access to postdischarge care options varied widely, based on hospitals’ catchment areas. The mean hospital’s catchment area contained 620.2 SNF beds, 24.7 primary care physicians, 48.5 nurse practitioners, 18.8 licensed nursing home beds, and 3.8 home health agencies per 100,000 residents (exhibit 1). Individual postdischarge care supply measures generally exhibited low-to-moderate correlations with each other (appendix exhibit A2).⁴¹ The supplies of SNF beds and home health agencies were the most strongly correlated ($r=0.49$), followed by the supply of home health agencies and licensed nursing home beds ($r=0.26$).

Exhibit 2 displays SNF beds per 100,000 residents within each county for 2019; similar maps for primary care physicians, nurse practitioners, licensed nursing home beds, and home health agencies are in appendix exhibits A3–A6, and maps of postdischarge care supply within hospital catchment areas are in appendix exhibits A7–A11.⁴¹ Major population centers (for example, Los Angeles, California; New York, New York; Chicago, Illinois; Boston, Massachusetts; and San Francisco, California) had greater raw counts of SNF beds but did not exhibit an oversupply of postdischarge care options per 100,000 residents. In contrast, the Great Plains had several counties with postdischarge care options per 100,000 residents that were 2 or 3+ standard deviations above the mean. Despite lower counts of SNF beds, these rural counties also had substantially lower populations; thus, sparsely populated areas throughout the Great Plains may appear to have high SNF availability. However, residents in rural areas may still face postdischarge care supply access barriers as a result of physical distance and transportation issues.

SOURCE Authors’ analysis of data from the Health Resources and Services Administration’s Area Health Resources Files. NOTE The exhibit displays the county-level supply of SNF beds per 100,000 residents.

Regression Results

In general, postdischarge care supply measures were negatively and significantly related to readmission rates in both pooled and condition-specific unadjusted bivariate models (exhibit 3). The results from adjusted random effects regression models are shown in exhibit 4. In adjusted models, nurse practitioner availability was positively associated with overall readmissions (+0.09 percentage points per SD; 95% confidence interval: 0.03, 0.15), whereas licensed nursing home beds (−0.09 pp/SD; 95% CI: −0.15, −0.04) and primary care physicians (−0.16 pp/SD; 95% CI: −0.21, −0.11) were associated with lower readmission rates (exhibit 4 shows point estimates and $p$ value levels; confidence intervals are not shown). Local availability of palliative care, home health agencies, and SNF beds had no significant association with overall readmission rates in pooled models.

Nursing home beds (−0.21 pp/SD; 95% CI: −0.34, −0.08), palliative care (−0.89 pp/SD; 95% CI: −1.41, −0.37), SNF beds (−0.12 pp/SD; 95% CI: −0.24, −0.00), and primary care physicians (−0.21 pp/SD; 95% CI: −0.31, −0.11) were associated with lower acute myocardial infarction readmissions after adjustment. For heart failure, nursing home beds (−0.14 pp/SD; 95% CI: −0.21, −0.06), SNF beds (−0.09 pp/SD; 95% CI: −0.17, −0.01), and primary care physicians (−0.20 pp/SD; 95% CI: −0.27, −0.13) were associated with lower readmissions. Home health agencies (+0.16 pp/SD; 95% CI: 0.08, 0.24) and nurse practitioners (+0.15 pp/SD; 95% CI: 0.07, 0.23) were associated with higher heart failure readmissions. Primary care physicians were associated with fewer pneumonia readmissions (−0.21 pp/SD; 95% CI: −0.27, −0.16), whereas nurse practitioners were associated with higher readmissions (+0.13 pp/SD; 95% CI: 0.07, 0.20).

Sensitivity Analyses

In sensitivity analyses, our results were generally robust to alternative methods for weighting postdischarge care supply to hospitals based on counties’ total discharges instead of population (appendix exhibit A12).⁴¹ When we weighted by counties’ total hospital discharges instead of by population, SNFs were associated with lower readmission rates in pooled models (−0.06 pp/SD; 95% CI: −0.11, −0.01) and in all three condition-specific models. Nurse practitioner supply was no longer associated with readmission rates in both the pooled conditions and heart failure models. The results for palliative care, nursing home beds, home health agencies, and primary care physicians were qualitatively unchanged by alternative weighting methods. Our results were also robust to the removal of hospital random effects and to alternative definitions of our palliative care measure (appendix exhibits A13 and A14).⁴¹

We also examined how postdischarge care supply’s association with readmission may differ according to hospitals’ rurality status, which was identified using the urban-rural codes contained in the CMS impact files. The overall relationship between readmission rates and nurse practitioners appears to be primarily driven by rural counties. For nonrural counties (appendix exhibit 15),⁴¹ palliative care was only associated with higher pneumonia readmissions (+0.37 pp/SD; 95% CI: 0.04, 0.70). Licensed nursing home beds were associated with lower readmissions for pneumonia (−0.11 pp/SD; 95% CI: −0.22, −0.01). Nurse practitioners were no longer associated with higher readmissions in pooled or heart failure models, and the result for primary care physicians was qualitatively similar.

The associations between postdischarge care supply and readmissions were generally smaller in magnitude and less significant in rural than in nonrural counties (appendix exhibit A16).⁴¹ Urban counties appear to drive the association between readmission rates, nursing home beds, and primary care physicians. For rural counties, nursing home beds were associated with lower readmission rates after heart failure discharges (−0.12 pp/SD; 95% CI: −0.23, −0.01). Primary care physician supply was no longer associated with acute myocardial infarction or pneumonia readmission rates. SNF supply was not associated with readmissions in pooled or condition-specific models for rural counties.

Discussion

We found that during the period 2013–19, the population-level supply of certain postdischarge care options varied widely across the US and that this variation was associated with patients’ risk for readmission after a discharge for acute myocardial infarction, heart failure, or pneumonia. After a wide variety of institutional, demographic, and CMS payment incentives were controlled for, hospitals with catchment areas having greater local availability of primary care physicians and licensed nursing home beds experienced lower readmission rates. Availability of palliative care and the supply of SNF beds were associated with lower readmissions for specific conditions, but not overall. A hospital-based palliative care service may reduce readmissions through both improved patient-provider goal setting and increased referrals to hospice care.¹⁸ In contrast, hospitals with greater local availability of home health agencies and nurse practitioners in their catchment areas experienced higher readmission rates, on average. Greater readmissions from home health agencies may be due to frequent staffing changes and associated discontinuities in care.⁴⁵

Our results suggest that hospitals may take a more active role in the development of postdischarge care options in their communities or partner with existing infrastructure to improve continuity of care and clinical outcomes and to avoid penalties under the HRRP.^46,47 Palliative care use may reduce unwanted, potentially unnecessary medical care for seriously ill people,^31,48 whereas patients in areas without sufficient access to primary care or nursing facilities (for example, isolated rural or low-income urban areas) may be forced to return to hospital emergency departments if complications arise.^17,49,50 In theory, increased local postdischarge care supply could correspond to more resources for postdischarge care and reduce readmissions. However, more supply of a particular postdischarge care option may induce hospitals to discharge to this setting because of greater availability but may also reflect higher health care needs or greater local demand for discharge to a particular setting. If not already doing so, hospitals should track readmission performance by discharge site and see whether there are opportunities to improve quality of care and lower readmission rates through reengineered discharge planning.^51,52

Our results have important implications for CMS’s risk-adjustment algorithm under the HRRP.

Our results have important implications for CMS’s risk-adjustment algorithm under the HRRP, which does not account for local postdischarge care supply. Inclusion of postdischarge care supply measures in the HRRP risk-adjustment algorithm would lead to lower readmission targets for hospitals that operate in areas with greater supply of primary care, SNFs, or licensed nursing home beds and higher readmission targets for hospitals with greater availability of home health agencies. Because local postdischarge care supply may have different effects for different condition-specific readmissions, the specific postdischarge care supply measures included in the modified HRRP risk adjustment could vary by condition.

Future research could use Medicare claims data to retroactively simulate the effects of enhanced risk adjustment for postdischarge care supply on the distribution of penalties under the HRRP. CMS should not risk-adjust for the availability of a hospital-based palliative care service, although palliative care is associated with reduced readmissions in both our study and previous research.^19,31 Risk adjustment would dissuade hospitals from initiating a palliative care service and punish hospitals that already have one by decreasing their expected readmission rate and increasing their likelihood of receiving a penalty under the HRRP. In addition, we would not recommend risk adjustment for nurse practitioner supply on the basis of these results. Nurse practitioners have been successfully employed in several interventions to reduce readmissions;^53,54 the mechanisms through which nurse practitioners may affect readmissions at the population level merit further study. Higher levels of nurse staffing may be associated with greater patient acuity,⁵⁵ and the greater role of nurse practitioners in a county may be indicative of being in an area that lacks certain medical resources. Importantly, as our measure of nurse practitioners is a single combined count of both hospital-based and office-based nurse practitioners, this measure does not necessarily reflect the supply available for postacute care, and an association between nurse practitioners and higher readmissions could be driven by higher levels of hospital nurse practitioner staffing. Additional data on the quality and accessibility of postdischarge care supply options (for example, CMS’s Five-Star Quality Rating System for nursing homes) may also provide valuable context beyond simple counts of postdischarge care supply availability.⁵⁶

Conclusion

We observed lower thirty-day readmission rates at hospitals that operated a palliative care service or had a greater local supply of primary care physicians, SNF beds, and licensed nursing home beds. Hospitals with a greater local supply of home health agencies or nurse practitioners were associated with increased readmissions. Our results suggest that hospitals may benefit from work to improve local access to care or hospital-community partnerships to improve continuity of care after discharge. CMS may also consider risk adjustment for postdischarge care supply under the HRRP to avoid penalizing or rewarding hospitals based on the characteristics of the communities they serve instead of the quality of care they provide.

ACKNOWLEDGMENTS

Kevin Griffith received a grant from the Agency for Healthcare Research and Quality to support this work (Grant No. R36HS027306-01).

NOTES

• 1 Ashton CM, Del Junco DJ, Souchek J, Wray NP, Mansyur CL. The association between the quality of inpatient care and early readmission: a meta-analysis of the evidence. Med Care. 1997;35(10):1044–59. Crossref, Medline, Google Scholar
• 2 Joynt KE, Figueroa JE, Oray J, Jha AK. Opinions on the Hospital Readmission Reduction Program: results of a national survey of hospital leaders. Am J Manag Care. 2016;22(8):e287–94. Medline, Google Scholar
• 3 McIlvennan CK, Eapen ZJ, Allen LA. Hospital Readmissions Reduction Program. Circulation. 2015;131(20):1796–803. Crossref, Medline, Google Scholar
• 4 Centers for Medicare and Medicaid Services. Hospital Readmissions Reduction Program (HRRP) [Internet]. Baltimore (MD): CMS; 2021 Dec 1 [cited 2022 May 16]. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program Google Scholar
• 5 Henry J. Kaiser Family Foundation. Aiming for fewer hospital U-turns: the Medicare Hospital Readmission Reduction Program [Internet]. San Francisco (CA): KFF; 2017 Mar 10 [cited 2022 May 16]. Available from: https://www.kff.org/medicare/issue-brief/aiming-for-fewer-hospital-u-turns-the-medicare-hospital-readmission-reduction-program/ Google Scholar
• 6 Joynt Maddox KE, Reidhead M, Qi AC, Nerenz DR. Association of stratification by dual enrollment status with financial penalties in the Hospital Readmissions Reduction Program. JAMA Intern Med. 2019;179(6):769–76. Crossref, Medline, Google Scholar
• 7 Bodenheimer T, Pham HH. Primary care: current problems and proposed solutions. Health Aff (Millwood). 2010;29(5):799–805. Go to the article, Google Scholar
• 8 Henry J. Kaiser Family Foundation. Primary care Health Professional Shortage Areas (HPSAs) [Internet]. San Francisco (CA): KFF; 2021 Sep 30 [cited 2022 May 16]. Available from: https://www.kff.org/other/state-indicator/primary-care-health-professional-shortage-areas-hpsas/ Google Scholar
• 9 Hart LG, Salsberg E, Phillips DM, Lishner DM. Rural health care providers in the United States. J Rural Health. 2002;18(5, Suppl):211–32. Medline, Google Scholar
• 10 Kangovi S, Grande D. Hospital readmissions—not just a measure of quality. JAMA. 2011;306(16):1796–7. Crossref, Medline, Google Scholar
• 11 Herrin J, St Andre J, Kenward K, Joshi MS, Audet AMJ, Hines SC. Community factors and hospital readmission rates. Health Serv Res. 2015;50(1):20–39. Crossref, Medline, Google Scholar
• 12 Barnett ML, Hsu J, McWilliams JM. Patient characteristics and differences in hospital readmission rates. JAMA Intern Med. 2015;175(11):1803–12. Crossref, Medline, Google Scholar
• 13 Aswani MS, Kilgore ML, Becker DJ, Redden DT, Sen B, Blackburn J. Differential impact of hospital and community factors on Medicare readmission penalties. Health Serv Res. 2018;53(6):4416–36. Crossref, Medline, Google Scholar
• 14 Gupta A, Allen LA, Bhatt DL, Cox M, DeVore AD, Heidenreich PAet al. Association of the Hospital Readmissions Reduction Program implementation with readmission and mortality outcomes in heart failure. JAMA Cardiol. 2018;3(1):44–53. Crossref, Medline, Google Scholar
• 15 Tsuchihashi M, Tsutsui H, Kodama K, Kasagi F, Setoguchi S, Mohr Met al. Medical and socioenvironmental predictors of hospital readmission in patients with congestive heart failure. Am Heart J. 2001;142(4):E7. Medline, Google Scholar
• 16 Weinberger M, Oddone EZ, Henderson WGVeterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. Does increased access to primary care reduce hospital readmissions? N Engl J Med. 1996;334(22):1441–7. Crossref, Medline, Google Scholar
• 17 Chetty VK, Culpepper L, Phillips RL, Rankin J, Xierali I, Finnegan Set al. FPs lower hospital readmission rates and costs. Am Fam Physician. 2011;83(9):1054. Medline, Google Scholar
• 18 O’Connor NR, Moyer ME, Behta M, Casarett DJ. The impact of inpatient palliative care consultations on 30-day hospital readmissions. J Palliat Med. 2015;18(11):956–61. Crossref, Medline, Google Scholar
• 19 Ranganathan A, Dougherty M, Waite D, Casarett D. Can palliative home care reduce 30-day readmissions? Results of a propensity score matched cohort study. J Palliat Med. 2013;16(10):1290–3. Crossref, Medline, Google Scholar
• 20 Figueroa JF, Joynt KE, Zhou X, Orav EJ, Jha AK. Safety-net hospitals face more barriers yet use fewer strategies to reduce readmissions. Med Care. 2017;55(3):229–35. Crossref, Medline, Google Scholar
• 21 Joynt KE, Orav EJ, Jha AK. Thirty-day readmission rates for Medicare beneficiaries by race and site of care. JAMA. 2011;305(7):675–81. Crossref, Medline, Google Scholar
• 22 Joynt KE, Jha AK. Characteristics of hospitals receiving penalties under the Hospital Readmissions Reduction Program. JAMA. 2013;309(4):342–3. Crossref, Medline, Google Scholar
• 23 Campbell SE, Seymour DG, Primrose WRACMEPLUS Project. A systematic literature review of factors affecting outcome in older medical patients admitted to hospital. Age Ageing. 2004;33(2):110–5. Medline, Google Scholar
• 24 Centers for Medicare and Medicaid Services. Hospital Compare [Internet]. Baltimore (MD): CMS; 2021 Dec 1 [cited 2022 May 16]. Available from: https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/HospitalCompare Google Scholar
• 25 National Bureau of Economic Research. Centers for Medicare and Medicaid Services (CMS) impact file Hospital IPPS [Internet]. Cambridge (MA): NBER; 2017 Aug 30 [cited 2022 May 16]. Available from: https://www.nber.org/research/data/centers-medicare-medicaid-services-cms-impact-file-hospital-ipps Google Scholar
• 26 Hospital Readmissions Reduction Program (HRRP). NEJM Catalyst [serial on the Internet]. 2018 Apr 26 [cited 2022 May 16]. Available from: https://catalyst.nejm.org/doi/full/10.1056/CAT.18.0194 Google Scholar
• 27 Griffith KN, Feyman Y, Auty SG, Crable EL, Levengood TW. County-level data on U.S. opioid distributions, demographics, healthcare supply, and healthcare access. Data Brief. 2021;35:106779. [Epub ahead of print]. Medline, Google Scholar
• 28 Centers for Medicare and Medicaid Services. Hospital Service Area file [Internet]. Baltimore (MD): CMS; 2021 Dec 1 [cited 2022 May 16]. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Hospital-Service-Area-File Google Scholar
• 29 Census Bureau. Relationship files [Internet]. Washington (DC): Census Bureau; 2022 Feb 28 [cited 2022 May 16]. Available from: https://www.census.gov/geographies/reference-files/2010/geo/relationship-files.html#par_textimage_674173622 Google Scholar
• 30 May P, Garrido MM, Del Fabbro E, Noreika D, Normand C, Skoro Net al. Evaluating hospital readmissions for persons with serious and complex illness: a competing risks approach. Med Care Res Rev. 2020;77(6):574–83. Crossref, Medline, Google Scholar
• 31 Gade G, Venohr I, Conner D, McGrady K, Beane J, Richardson RHet al. Impact of an inpatient palliative care team: a randomized control trial. J Palliat Med. 2008;11(2):180–90. Crossref, Medline, Google Scholar
• 32 Desai NR, Ross JS, Kwon JY, Herrin J, Dharmarajan K, Bernheim SMet al. Association between hospital penalty status under the Hospital Readmission Reduction Program and readmission rates for target and nontarget conditions. JAMA. 2016;316(24):2647–56. Crossref, Medline, Google Scholar
• 33 Ferro EG, Secemsky EA, Wadhera RK, Choi E, Strom JB, Wasfy JHet al. Patient readmission rates for all insurance types after implementation of the Hospital Readmissions Reduction Program. Health Aff (Millwood). 2019;38(4):585–93. Go to the article, Google Scholar
• 34 Mittal M, Wang CE, Goben AH, Boyd AD. Proprietary management and higher readmission rates: a correlation. PLoS One. 2018;13(9):e0204272. Crossref, Medline, Google Scholar
• 35 Centers for Medicare and Medicaid Services. Case mix index [Internet]. Baltimore (MD): CMS; [cited 2022 May 26]. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Acute-Inpatient-Files-for-Download-Items/CMS022630 Google Scholar
• 36 Centers for Medicare and Medicaid Services. Disproportionate share hospital (DSH) [Internet]. Baltimore (MD): CMS; [last modified 2022 Apr 18; cited 2022 May 26]. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/dsh Google Scholar
• 37 Centers for Medicare and Medicaid Services. Outlier payments [Internet]. Baltimore (MD): CMS; [last modified 2021 Dec 1; cited 2022 May 16]. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/outlier Google Scholar
• 38 Edmunds M, Sloan FA, editors. Geographic adjustment in Medicare payment: phase I: improving accuracy. Second edition. Washington (DC): National Academies Press; 2012. Google Scholar
• 39 Centers for Medicare and Medicaid Services. Wage index [Internet]. Baltimore (MD): CMS; [last modified 2022 Apr 25; cited 2022 May 26]. Available from: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/wageindex Google Scholar
• 40 Centers for Medicare and Medicaid Services. Hospital Value-Based Purchasing (HVBP) Program: payments [Internet]. Baltimore (MD): CMS; [cited 2022 May 26]. Available from: https://qualitynet.cms.gov/inpatient/hvbp/payment Google Scholar
• 41 To access the appendix, click on the Details tab of the article online.
• 42 Boozary AS, Manchin J, Wicker RF. The Medicare Hospital Readmissions Reduction Program: time for reform. JAMA. 2015;314(4):347–8. Crossref, Medline, Google Scholar
• 43 Dumanovsky T, Augustin R, Rogers M, Lettang K, Meier DE, Morrison RS. The growth of palliative care in U.S. hospitals: a status report. J Palliat Med. 2016;19(1):8–15. Crossref, Medline, Google Scholar
• 44 American Hospital Association. Fast facts on U.S. hospitals, 2022 [Internet]. Chicago (IL): AHA; [cited 2022 May 16]. Available from: https://www.aha.org/statistics/fast-facts-us-hospitals Google Scholar
• 45 Werner RM, Coe NB, Qi M, Konetzka RT. Patient outcomes after hospital discharge to home with home health care vs to a skilled nursing facility. JAMA Intern Med. 2019;179(5):617–23. Crossref, Medline, Google Scholar
• 46 Kurtzman JH. A community hospital-county health department partnership to reduce preventable readmissions: lessons learned for population health management. J Healthc Manag. 2015;60(4):258–67. Medline, Google Scholar
• 47 Cheon O, Baek J, Kash BA, Jones SL. An exploration of community partnerships, safety-net hospitals, and readmission rates. Health Serv Res. 2020;55(4):531–40. Crossref, Medline, Google Scholar
• 48 Tangeman JC, Rudra CB, Kerr CW, Grant PC. A hospice-hospital partnership: reducing hospitalization costs and 30-day readmissions among seriously ill adults. J Palliat Med. 2014;17(9):1005–10. Crossref, Medline, Google Scholar
• 49 Toth M, Holmes M, Van Houtven C, Toles M, Weinberger M, Silberman P. Rural Medicare beneficiaries have fewer follow-up visits and greater emergency department use postdischarge. Med Care. 2015;53(9):800–8. Crossref, Medline, Google Scholar
• 50 van Dis J. MSJAMA. Where we live: health care in rural vs urban America. JAMA. 2002;287(1):108–13. Medline, Google Scholar
• 51 Berkowitz RE, Fang Z, Helfand BK, Jones RN, Schreiber R, Paasche-Orlow MK. Project ReEngineered Discharge (RED) lowers hospital readmissions of patients discharged from a skilled nursing facility. J Am Med Dir Assoc. 2013;14(10):736–40. Crossref, Medline, Google Scholar
• 52 Bowles KH, Holmes JH, Ratcliffe SJ, Liberatore M, Nydick R, Naylor MD. Factors identified by experts to support decision making for post acute referral. Nurs Res. 2009;58(2):115–22. Medline, Google Scholar
• 53 Smith J, Pan D, Novelli M. A nurse practitioner–led intervention to reduce hospital readmissions. J Nurse Pract. 2016;12(5):311–6. Google Scholar
• 54 Delgado-Passler P, McCaffrey R. The influences of postdischarge management by nurse practitioners on hospital readmission for heart failure. J Am Acad Nurse Pract. 2006;18(4):154–60. Medline, Google Scholar
• 55 Mark BA, Harless DW. Adjusting for patient acuity in measurement of nurse staffing: two approaches. Nurs Res. 2011;60(2):107–14. Medline, Google Scholar
• 56 Williams A, Straker JK, Applebaum R. The nursing home five star rating: how does it compare to resident and family views of care? Gerontologist. 2016;56(2):234–42. Crossref, Medline, Google Scholar