Rural hospital closures increase mortality

Kritee Gujral 10 June 2020

a

A

About 15% of all US hospitals have closed since 1990, and the rate of closures has accelerated in recent years, in part due to declining inpatient use (Carroll 2019, Lindrooth et al. 2003). Rural closures have increased since 2010 and 25% of all rural hospitals – of which 82% are considered highly essential to their communities – are at high risk of closing (Mosley and Debehnke 2020, North Carolina Rural Health Research Program 2020, Holmes et al. 2017). There have been prominent media outcries regarding the rural hospital crisis, including in 2020 presidential campaign discussions (McCausland 2019, Tribble 2019, Estes 2020, Raffa 2019, Frakt 2018).

Internationally, too, smaller, rural, and remote hospitals are under the threat of closures (Vaughan and Edwards 2020). Low patient volumes, low reimbursement rates, and costly uncompensated care are commonly cited reasons for weakened hospital finances that lead to closures (Carroll 2019, Lindrooth et al. 2018, Kaufman et al. 2016).

The COVID-19 pandemic (or any future economic downturn) is expected to worsen the situation. COVID-19 has led to a 12.1% drop in the US GDP component related to hospitals, and as millions of workers lose their employer-covered insurance, hospitals could be facing unprecedented levels of uncompensated care, drastically increasing the risk of and concerns related to hospital closures (Federation of American Hospitals 2020, Mosley and Debehnke 2020, Ellison 2020).

Tradeoffs in efficiency considerations of closures 

Hospital closures pose a conundrum for economists, policymakers, and hospital administrators. Closures may seem reasonable and inevitable if hospitals are suffering financially, but they can severely fracture patient access and cause enormous harm to patient health (Kaufman et al. 2016).

In a well-functioning market, a firm’s shutdown often signals underlying inefficiency or low product demand. Previous studies have shown this to be the case for the hospital industry, where closures – especially urban ones – were positively associated with inefficiency (Capps et al. 2010, Lindrooth 2003).

A large body of literature suggests that the risk of closure is higher for smaller and non-profit hospitals. Large hospitals can achieve greater efficiency through economies of scale and for-profit hospitals are more responsive to competitive pressures and can ‘cream-skim’ or choose patients that enhance profitability (Allen 2017, Hodgson, et al. 2015, Lindrooth et al. 2003).

The natural divide between rural and urban hospitals

The cost structure, profit motives, and patient mix of hospitals thus engender a natural divide between rural and urban hospitals, one that has continued to widen in recent years (Lindrooth et al. 2018, Thomas et al. 2016, Kaufman et al. 2016).

Rural hospitals may be too small to operate efficiently (Song and Saghafian 2019, Carroll 2019), and economic downturns lead to shrinking populations in rural areas, implying lower patient volumes at hospitals.

Populations left behind may be older, sicker, and more reliant on Medicaid and Medicare, which are financially less favourable payer-mixes for hospitals, making rural hospitals more susceptible to closures and cause for high concern (Kaufman et al. 2016). The same concerns regarding financial viability may also deter new hospitals from entering rural healthcare markets.

Effects of rural hospital closures on patient access and health outcomes

Efficiency considerations of closures must, nonetheless, account for their effect on patient welfare. Urban closures may increase overall patient welfare through by way of cost savings for all patients from the closure of inefficient hospitals (Capps et al. 2010), but patient access and outcomes are also crucial components of patient welfare (Gaynor and Town 2011).

Even if market forces drive hospital shutdowns, closures raise significant concerns about increased transportation time, treatment delay, and adverse health outcomes, especially for emergency care (Miller et al. 2020, Carroll 2019, Sun et al. 2005, Wishner et al. 2016). Patient disutility from access issues can offset utility gains from cost savings, particularly in rural markets (Hsia and Shen 2011, McNamara 1999).

Compared to studies examining reasons for or the financial impact of closures, literature studying the effect of hospital closures, both rural and urban, on patient outcomes is relatively sparse and lacks consensus. Large-scale published studies have found no impact of closures on patient mortality with some evidence of improved outcomes, i.e. a reduction in mortality, for heart attack patients (Joynt et al. 2015). Only one published county-level study (Buchmueller et al. 2006) thus far has shown adverse patient-level impact, i.e. increased deaths due to urban closures.

Given the recent increase in closures, crucial efforts to ascertain the patient-level impact are currently underway; recent working papers (Song and Saghafian 2019, Carroll 2019) do find that closures can increase mortality. However, the emerging and existing work has focused almost exclusively on Medicare patients; there is a need to also study potentially more vulnerable non-Medicare populations.

Differences between rural vs urban hospital closures

Patient outcome studies either examine rural closures while ignoring concurrent urban closures (making it difficult to disentangle the isolated effect of rural closures) or study the general effect of closures without distinguishing rural from urban. However, there are baseline and post-closure differences between rural and urban hospitals and markets that suggest potentially different effects of rural and urban closures.

At baseline, rural patients face larger transportation times and access barriers, bypass nearby hospitals, and are more likely to use hospitals for primary care needs, compared to urban patients (Wishner et al. 2016, Escarce and Kapur 2009). Rural patients experience a 76% increase in ambulance transportation time post-closure in their ZIP code, whereas urban patients do not experience any such change post-closure in their ZIP code (Troske and Davis 2019).

While urban patients may shift their usual source of care from hospitals to physicians’ offices (Buchmueller et al. 2006), no such favourable substitution is observed for rural patients (Rosenbach and Dayhoff 1995), most likely because there are fewer remaining alternatives in rural areas.

Additionally, rural closures can lead to immediate, substantial, and persistent outmigration of healthcare professionals (Germack et al. 2019, Manlove and Whitacre 2017), dismembering patient access and jeopardising patient outcomes. As rural hospitals are an integral part of the community’s economic foundation, closures also have dire consequences on the local economy (Manlove and Whitacre 2017, Holmes et al. 2006, Probst et al. 1999).

Given such baseline and post-closure differences, the overall adverse effect of a rural hospital closure, which includes transportation delays and related spillover effects, is likely larger than that of an urban closure.

Our work

Our recent working paper (Gujral and Basu 2019) uses a difference-in-difference approach to study the impact of rural and urban hospital closures in California in 1995–2011, on adjusted inpatient mortality for time-sensitive conditions. Outcomes of admissions in hospital service areas with and without closure(s) (‘treatment’ versus ‘control’, respectively) are compared before and after the closure year (Figure 1).

Figure 1 Urban and rural hospital closures in California (1995–2011)

Notes: Map of California showing rural and urban ZIP codes, closure ZIP codes, and hospital service area boundaries that indicate affected patients.

We focus on: (1) the differential effects of rural and urban hospital closures, and (2) the effect on Medicare as well as non-Medicare populations.

Findings

Figure 2 shows trends in unadjusted inpatient mortality before and after closures. The control group and the general (non-differentiated) treatment group follow a similar trajectory prior to closure, and there does not appear to be any change post-closure for the treatment group.

Figure 2 Unadjusted inpatient mortality for admissions in hospital service areas with and without hospital closures

This graphical observation is supported by the difference-in-difference estimation, wherein we find no measurable impact of closures on inpatient mortality, similar to Joynt et al. (2015) and Hsia et al. (2012).

We then differentiate the treatment group by rurality, as shown in Figure 3. The left-hand side shows that in urban areas, there is no change in unadjusted inpatient mortality after the hospital closure, similar to Figure 2 above. However, the right-hand side shows a post-hospital-closure increase in mortality for rural areas.

Figure 3 Unadjusted inpatient mortality for urban and rural areas with and without hospital closures

Notes: Unadjusted inpatient mortality for: control vs urban treatment group (top), and control vs rural treatment group (bottom).

The difference-in-difference estimation shows that rural closures increase inpatient mortality by 8.7%, while urban closures have no measurable impact. This suggests that when no distinction is made between rural or urban closures, the null effect of urban closures have the potential to dominate and mask the significant and detrimental impact of rural closures.

Our subgroup analyses indicate that Medicaid patients and racial minorities are relatively worse affected by rural closures (11.3% and 12.6%, respectively), which supports the hypothesis that closures disproportionately affect vulnerable populations.

We find evidence of spillover effects: rural closures also increase mortality for patients that reside in urban ZIP codes, by 7.6%. Results are thus not driven by rural ZIP codes alone and rural closures have negative implications for neighbouring urban ZIP codes, likely due to overburdening of shared resources in the area.

For each subgroup analysed, the harmful impact of rural closures is accompanied by a significant increase in the length of stay: the overall increase of 5.2% or about 7 hours in length of stay can contribute to overcrowding (Henneman et al. 2010). Similar spillover effects of closures are also noted in other recent work (Song and Saghafian 2019, Liu et al. 2014).

Implications for policy

Despite the increase in rates of hospital closures, mounting concerns for patient health, and prominent media outcries regarding the issue, the research evidence base for its effects on patient outcomes is relatively sparse, with the bulk of large-scale published evidence showing that hospital closures have no adverse patient-level impact.

Our work calls the attention of policymakers, hospital administrators, and healthcare researchers to the significant adverse effects that rural hospital closures have. These effects can be overlooked (and may have been missed in prior work) if no distinction is made between rural and urban hospitals.

Policies to mitigate the effects of closures should be targeted and prioritised to account for structural differences across rural and urban hospitals and healthcare markets – differences that we show lead to vastly different health outcomes.

Not all closures are necessarily problematic. Urban closures may reflect the local communities’ preferences for nearby substitutes and may even include mechanisms that improve patient outcomes such as improved care quality due to consolidation of services at neighbouring open hospitals.

Other scholars (Song and Saghafian 2019, Frakt 2019, Lindrooth et al. 2018) advocate for similar targeting and prioritising for a range of policy instruments – such as hospital bailouts, subsidies or changes in reimbursement policies – to harness the benefits of certain types of closures, while mitigating the considerable harm caused by others.

Additionally, while policy instruments that tackle the root causes of closures appear less straightforward, it is essential to ensure emergency transportation and healthcare accessibility post-closures, especially for rural areas and for vulnerable populations.

Research also indicates that spillover effects on surrounding hospitals – resulting from potentially different healthcare demands – need to be managed, particularly if there are simultaneous reductions in the long-term supply of healthcare professionals.

References

Allen, Jacob (2017), “Medicaid expansion and hospital closures: Examining hospital, county, and state effects in the wake of the Affordable Care Act”, public health theses, Georgia State University.

Buchmueller, Thomas C, Mireille Jacobson and Cheryl Wold (2006), “How far to the hospital? The effect of hospital closures on access to care”, Journal of Health Economics 25(4): 740–61.

Capps, Cory, David Dranove and Richard C Lindrooth (2010), “Hospital closure and economic efficiency”, Journal of Health Economics 29(1): 87–109.

Carroll, Caitlin (2019), “Impeding access or promoting efficiency? Effects of rural hospital closure on the cost and quality of care”, NBER Working Paper Series.

Ellison, Ayla (2020), “Rural America braces for COVID-19 as more hospitals close”, Becker’s Healthcare, 13 April.

Escarce, Jose J, and Kanika Kapur (2009), “Do patients bypass rural hospitals?Journal of Health Care for the Poor and Underserved 20(3): 625–44.

Estes, Clary (2020), “1 in 4 rural hospitals are at risk of closure and the problem is getting worse”, Forbes, 24 February.

Federation of American Hospitals (2020), “COVID-19 alert: US hospital finances in critical condition”.

Frakt, Austin B (2019), “The rural hospital problem”, JAMA Health Forum, 1 May.

Frakt, Austin (2018), “A sense of alarm as rural hospitals keep closing”, The New York Times, 29 October.

Gaynor, Martin, and Robert J Town (2011), “Competition in health care markets”, in Mark V Pauly, Thomas G Mcguire and Pedro P Barros (eds.), Handbook of Health Economics, 499-637.

Germack, Hayley D, Ryan Kandrack and Grant R Martsolf (2019), “When rural hospitals close, the physician workforce goes”, Health Affairs 28(12): 2086–94.

Gujral, Kritee, and Anirban Basu (2019), “Impact of rural and urban hospital closures on inpatient mortality”, NBER Working Paper 26182.

Henneman, P L, B H Nathanson, H Li, H A Smithline, F S Blank, J P Santoro, A M Maynard, D A Provost and E A Henneman (2010), “Emergency department patients who stay more than 6 hours contribute to crowding”, The Journal of Emergency Medicine.

Hodgson, Ashley, Paul Roback, Andrew Hartman, Erin Kelly, and Yujie Li (2015), “The financial impact of hospital closures on surrounding hospitals”, Journal of Hospital Administration 4(3).

Holmes, George M, Brystana G Kaufman and George H Pink (2017), “Predicting financial distress and closure in rural hospitals”, The Journal of Rural Health.

Holmes, George M, Rebecca T Slifkin, Randy K Randolph and Stephanie Poley (2006), “The effect of rural hospital closures on community economic health”, Health Services Research.

Hsia, Renee Y, Hemal K Kanzaria, Tanja Srebotnjak, Judy Maselli, Charles McCulloch and Andrew D Auerbach (2012), “Is emergency department closure resulting in increased distance to nearest emergency department associated with inpatient mortality?” Health Policy.

Hsia, Renee Yuen-Jan, and Yu-Chu Shen (2011), “Rising closures of hospital trauma centers disproportionately burden vulnerable populations”, Health Affairs.

Joynt, Karen E, Paula Chatterjee, E John Orav and Ashish K Jha (2015), “Hospital closures had no measurable impact on local hospitalization rates or mortality rates, 2003–11”, Health Affairs.

Kaufman, Brystana G, Kristin L Reiter, George H Pink and George M Holmes (2016), “Medicaid expansion affects rural and urban hospitals differently”, Health Affairs.

Kaufman, Brystana G, Sharita R Thomas, Randy K Randolph, Julie R Perry, Kristie W Thompson, George M Holmes and George H Pink (2016), “The rising rate of rural hospital closures”, The Journal of Rural Health.

Lindrooth, Richard C, Anthony T Lo Sasso, and Gloria J Bazzoli (2003), “The effect of urban hospital closure on markets”, Journal of Health Economics 691–712.

Lindrooth, Richard C, Marcelo C Perraillon, Rose Y Hardy, and Gregory J Tung (2018), “Understanding the relationship between Medicaid expansions and hospital closures”, Health Affairs.

Liu, Charles, Tanja Srebotnjak, and Renee Y Hsia (2014), “California emergency department closures are associated with increased inpatient mortality at nearby hospitals”, Health Affairs.

Manlove, Jacob, and Brian Whitacre (2017), “Short-term economic impact of rural hospital closures”, 2017 Annual Meeting, Southern Agricultural Economics Association.

McCausland, Phil (2019), “Rural hospital closings cause mortality rates to rise, study finds”, NBC News, 6 September.

McNamara, Paul E (1999), “Welfare effects of rural hospital closures: A nested logit analysis of the demand for rural hospital services”, American Journal of Agricultural Economics.

Miller, Katherine E M, Hailey J James, George M Holmes, and Courtney H Van Houtven (2020), “The effect of rural hospital closures on emergency medical service response and transport times”, Health Services Research.

Mosley, Dave, and Daniel Debehnke (2020), “2020 rural hospital sustainability index”, Guidehouse.com, 8 April.

North Carolina Rural Health Research Program (2020), “172 rural hospital closures: January 2005–present (130 since 2010)”, The Cecil G Sheps Center for Health Services Research, University of North Carolina at Chapel Hill.

Premkumar, Deepak, Dave Jones and Peter Orazem (2016), “Hospital closure and hospital choice: How hospital quality and availability will affect rural residents”, Iowa State University Economics Working Papers.

Probst, Janice C, Michael E Samuels, James R Hussey, David E Berry, and Thomas C Ricketts (1999), “Economic impact of hospital closure on small rural counties, 1984 to 1988: Demonstration of a comparative analysis approach”, The Journal of Rural Health.

Raffa, Allie (2019), “America’s rural hospital crisis becomes major 2020 campaign issue”, Fox News, 8 December.

Rosenbach, Margo L, and Debra A Dayhoff (1995), “Access to care in rural America: Impact of hospital closures”, Health Care Financing Review.

Song, Lina D, and Soroush Saghafian (2019), “The Spillover effects of hospital closures on the efficiency and quality of other hospitals”, John F Kennedy School of Government Faculty Research Working Paper Series.

Sun, Benjamin C, Sarita A Mohanty, Robert Weiss, Richard Tadeo, Maureen Hasbrouck, William Koenig, Carol Meyer, and Steven Asch (2005), “Effects of hospital closures and hospital characteristics on emergency department ambulance diversion, Los Angeles County, 1998 to 2004”, Health Policy and Clinical Practice.

Thomas, Sharita R, George M Holmes and George H Pink (2016), “2012–14 profitability of urban and rural hospitals by Medicare payment classification”, Findings Brief, North Carolina Rural Health Research Program, University of North Carolina at Chapel Hill.

Tribble, Sarah Jane (2019), “No mercy: After the hospital closes, how do people get emergency care?” National Public Radio, 18 August.

Troske, SuZanne, and Alison Davis (2019), “Do hospital closures affect patient time in an ambulance?”, Policy Brief, Rural and Underserved Health Research Center, University of Kentucky.

Vaughan, Louella, and Nigel Edwards (2020), “The problems of smaller, rural and remote hospitals: Separating facts from fiction”, Future Healthcare Journal 7(1): 38–45.

Wishner, Jane, Patricia Solleveld, Robin Rudowitz, Julia Paradise and Larisa Antonisse (2016), “A look at rural hospital closures and implications for access to care: Three case studies”, Issue Brief, The Kaiser Commission on Medicaid and the Uninsured.

a

A

Topics:  Health economics Poverty and income inequality

Tags:  health inequality, health outcomes, healthcare, healthcare quality, hospital closures, Hospitals, mortality, Rural, urban, US, vulnerable groups

Health Economist and Senior Fellow, CHOICE Institute, University of Washington

Events

CEPR Policy Research