Associations between Medicaid expansion and nurse staffing ratios and hospital readmissions

Abstract

Objective

To examine the associations between Medicaid expansion and nurse staffing ratios and hospital‐wide readmission rates.

Data sources

Secondary data from the 2011‐2016 Healthcare Cost Report Information System, the American Hospital Association Annual Survey, and the Hospital Compare data.

Study design

Difference‐in‐difference models are used to compare outcomes in hospitals located in states that expanded Medicaid with those located in nonexpansion states. The changes in nurse staffing ratios and hospital‐wide readmission rates are calculated in each one of the postexpansion years (2014, 2015, and 2016), compared to pre‐expansion.

Principal findings

Results indicate that nurse staffing ratios increased, whereas hospital‐wide readmission rates declined in expansion states relative to nonexpansion states. Nurse staffing ratios increased by 0.33, 0.42, and 0.46 registered nurses hours per adjusted patient days in 2014, 2015, and 2016 in hospitals located in expansion states, compared with hospitals in nonexpansion states after expansion. This increase was statistically significant (P < .001) in 2015 and 2016, but marginally significant (P = .016) in 2014. Hospital‐wide readmission rates statistically significantly decreased by 9, 16, and 18 per 10 000 patients (P < .001) in 2014, 2015, and 2016, respectively, in expansion vs nonexpansion states hospitals after expansion.

Conclusions

Medicaid expansion was associated with gradually improved hospitals' nurse staffing ratios and hospital‐wide readmission rates from 2014 through 2016. The continued monitoring of quality measures of hospitals can help assess the impact of Medicaid expansion over a longer period of time.

What This Study Adds.

• Medicaid expansion has been associated with several outcomes, including a decline in uncompensated care costs and an increase in Medicaid revenues.

• Yet, it is unclear whether Medicaid expansion is associated with improved number of key staff and hospital performance.

• This is the first study to examine the association between Medicaid expansion and nurse staffing ratios and hospital‐wide readmission rates.

• Findings suggest that Medicaid expansion was associated with improved hospitals' nurse staffing ratios and hospital‐wide readmission rates from 2014 through 2016.

1. INTRODUCTION

Since the passage of the Affordable Care Act (ACA) in 2010, the national uninsured rate decreased substantially from 16% in 2010 to 9.1% in 2017;1 more specifically, between 2013 and 2016, the period relevant to this analysis, the uninsured rate declined from 14.5 to 8.6%, a 5.9 percentage point.2 Expansion of Medicaid, adopted in 33 states and the District of Columbia as of July 2018,3 and the availability of private plans (and for some low‐income individuals premium tax credits and subsidies) through the Heath Insurance Marketplaces were significant factors in reducing the number of uninsured persons.4, 5

Because hospitals provided approximately $37.0 billion in uncompensated care in 2013,6 the year before implementation of these two major coverage expansion provisions in the ACA, hospitals were expected to realize significant reductions in uncompensated care and increases in Medicaid revenues associated with coverage expansions.7, 8 Previous studies reported an association between Medicaid expansion and a decline in uncompensated care costs and increases in Medicaid revenues and margins.9, 10, 11, 12 Yet, it remains unclear how hospitals invested the new financial resources and whether they chose to increase number of key staff and improve performance to avoid potential Medicare readmission penalties.13

There are several ways in which Medicaid expansion may have influenced staffing and meeting quality requirements. One possibility is that hospitals may have used the new financial resources generated from the decline in uncompensated care costs and the increase in Medicaid revenues (since uninsured patients were substituted by Medicaid patients) in improving patients' experiences at a hospital by hiring new staff and/or increasing productive working hours. Further, investing the extra resources in improving quality such as hospital readmission rates saves hospitals form paying penalties. Nevertheless, hospitals are not expected to invest in staffing or improving readmission until they realize the actual presence of additional financial resources. A second possibility is that newly enrolled Medicaid patients may have simply received better quality of services than uninsured patients, resulting in better outcomes (such as lower readmission rates) in Medicaid expansion states. Alternatively, hospitals may have anticipated an increase in number of newly enrolled patients and wanted to prepare for that by adding more staff. However, research showed that admissions remained consistent in expansion and nonexpansion states, but payer mix changed (Medicaid encounters increased while uninsured encounters decreased).14, 15

Medicaid expansion has been associated with improved access to outpatient and preventive care and self‐reported quality of care.16, 17 One study reported no association between Medicaid expansion and improvement in quality measures specific for hospitalized patients with heart failure.18 However, no studies have demonstrated associations between Medicaid expansion and nurse staffing or readmission rates. Higher nurse staffing ratios have been associated with shorter length of stay and lower rates of adverse outcomes such as urinary tract infections and upper gastrointestinal bleeding.19 On the other hand, lower nurse staffing ratios were associated with higher rates of 30‐day mortality.20 The 30‐day hospital‐wide all‐cause unplanned Medicare readmission rate (which captures all‐cause unplanned readmissions within 30 days of discharge)21 is one of several measures in the Hospital Inpatient Quality Reporting Program, which aims to improve quality of inpatient hospital care by measuring and publicly displaying data about hospital performance.22 This quality measure has been endorsed by the National Quality Forum (NQF), and its application requires hospitals to expand resources.23 Hospitals may use new revenues (from Medicaid expansion or any other sources) to invest in reducing the 30‐day hospital‐wide all‐cause unplanned Medicare readmission rate.

Hospital performance is incremental and is not limited to financial performance. Hospital performance includes financial measures, clinical measures, and combination measures that attempt to address hospital value.24 This study examines whether improvement in financial measures was translated to hospitals' improved overall performance in ways that align with the objectives of the Department of Health and Human Services of transforming the health care delivery system to improve quality of care.25

In this study, the most recently available data were used and two hypotheses were addressed: (a) Medicaid expansion was associated with increased nurse staffing ratios, reduced 30‐day hospital‐wide Medicare hospital‐wide readmission rates, reduced uncompensated care costs, and increased Medicaid revenues; and (b) improvements in these outcome measures started in the 2014 first postexpansion year and continued during the 2015 second postexpansion year and the 2016 third postexpansion year. Findings on the main outcomes (nurse staffing ratios and readmission rates) are presented in the paper, whereas findings on the secondary outcomes (uncompensated care costs and Medicaid revenues) are presented in the Appendix S1. Although previous studies examined uncompensated care and Medicaid revenues, this paper still includes these outcome measures in the Appendix S1 because it adds an additional recent year of postexpansion (2016), examines the changes in the outcome measures in each year of expansion, and justifies the improvement in nurse staffing ratios and readmission rates based on improvement in the financial measures.

2. METHODS

The 2011‐2016 Healthcare Cost Report Information System (HCRIS) data released by CMS served as the primary source of data.6 All hospitals accepting Medicare beneficiaries are required to submit their Medicare Cost Reports to CMS. Thus, HCRIS contains extensive cost‐related data from a vast majority of hospitals in the United States. HCRIS data contain information on financial performance of hospitals, including Medicaid revenues and uncompensated care costs. Potential outliers in the HCRIS data were addressed by winsorizing the data per case at the 99th percentile. HCRIS data were linked at the hospital level with (a) nurse staffing ratios for 2011‐2016, obtained from the American Hospital Association (AHA) Annual Survey,26 and (b) the hospital's Readmissions and Deaths data, obtained from the Hospital Compare website27 to assess secondary outcomes.

The primary outcome measures were nurse staffing ratios and hospital‐wide readmission rates. Nurse staffing ratio (from AHA) was defined as number of productive hours of full‐time equivalent registered nurses per adjusted patient days (inpatients and outpatients.) The numerator for nurse staffing ratio, the productive nursing hours, was calculated as 80% of total number of hours for a full‐time registered nurse in one year (1768), multiplied by number of full‐time equivalent registered nurses.19, 28 Hospital‐wide readmission rate (obtained from the Hospital Compare data) is a risk‐adjusted 30‐day hospital‐wide all‐cause unplanned Medicare readmission rate that provides a broad assessment of a hospital's quality of care,27, 29 and is used by CMS as one of several measures in the Hospital Inpatient Quality Reporting Program.27, 30 For each hospital, CMS calculates readmission rate based on the weighted mean of the hospital's 30‐day unplanned readmission rates for fee‐for‐service Medicare patients for five clinically similar admission categories: medicine, surgery or gynecology, cardiorespiratory, cardiovascular, and neurology (e.g., rates across all hospitals ranged from 10.8% to 19.4% in 2015).29 For easier interpretation, readmission rates are presented as number of readmissions per 10 000 patients.

The secondary outcome measures (hospital financial measures) included uncompensated care costs and Medicaid revenues. Uncompensated care was defined as the sum of charity care and bad debt costs. This definition included costs attributed to providing services to uninsured and underinsured patients, but excluded Medicaid shortfalls (i.e., underpayment from Medicaid). Medicaid revenues were the revenues from Medicaid and included Medicaid Disproportionate Share Hospital payments.

The sample included hospitals that reported financial data to CMS in any quarter from 2011 through 2016. Hospitals in US territories were excluded from the analysis because they were not affected by Medicaid expansion. Hospitals in five states (California, Massachusetts, Minnesota, Vermont, and Washington) and the District of Columbia were excluded because they had extended coverage to nonelderly, childless, nondisabled adults with incomes up to 133% FPL before January 2014, through various Medicaid programs, including Section 1115 waivers and the ACA early expansion option.8, 31 In addition, four states that expanded Medicaid in late 2014 (New Hampshire) or in 2015 (Alaska, Indiana and Pennsylvania) were excluded from this study. Finally, hospitals that had any missing or negative value for each one of the outcomes of interest were excluded. The primary independent variable was whether hospitals were located in states that expanded Medicaid. During 2014, 20 states expanded Medicaid (not including the states excluded from the study), whereas 21 states had not expanded Medicaid coverage pursuant to the ACA.3

2.1. Statistical methods

Characteristics of hospitals in expansion and nonexpansion states were examined for 2011‐2013 (pre‐expansion years). Changes in means of nurse staffing ratios and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates were examined in expansion and nonexpansion states from 2011 through 2016 (and in the pre‐expansion period from 2011 to 2013). Percent change from 2011 to 2016 was calculated for each outcome measure.

Difference‐in‐difference models were conducted to compare the difference in outcomes in hospitals located in expansion states with hospitals in nonexpansion states. Parallel trends are key assumptions of the difference‐in‐difference model. In this study, it implies that outcome measures in hospitals in both expansion and nonexpansion states would have had similar trends in the absence of Medicaid expansion. The validity of this assumption is tested by comparing the trends in expansion and nonexpansion states as shown in Appendix S1; Figure A1 and by estimating regression models using pre‐expansion data only and a linear time trend for the prior years. (More details in the Appendix S1).

All the analyses controlled for hospital characteristics such as number of total beds, ownership type, location, and teaching status. Each hospital in the analysis had four time points (pre‐expansion years, first postexpansion year, second postexpansion year, and third postexpansion year), allowing for the consideration of the differences in hospitals' reporting periods, which were based on fiscal years. A fiscal year was defined by the calendar year end date. Each fiscal year captured a full calendar year of exposure to Medicaid except 2014, which captured partial exposure to the expansion. To handle this issue, the analyses replaced the full year of 2014 with a fraction of a year (equals to number of months of exposure to Medicaid expansion in 2014) to represent level of exposure to the expansion. For example, hospitals with fiscal years ending in January 2014 had a fraction of 1/12 exposure to Medicaid expansion, whereas hospitals that had fiscal years ending in December 2014 had a 12/12 exposure to the expansion. The analysis controlled for random effects at the hospital level to account for unobserved differences across hospitals that might affect the outcomes. The analysis also controlled for states' fixed effects to account for differences within states. The difference‐in‐difference model captured changes in outcomes of interest in hospitals located in expansion states to changes in the same outcomes in hospitals located in nonexpansion states, after expansion. It also captured the changes in outcomes for hospitals in expansion states during the first postexpansion year, compared with hospitals in nonexpansion states. Various hospital characteristics (number of total beds, ownership type [for‐profit, not‐for‐profit, or government‐owned], location [urban or rural], and teaching status) were controlled for in the analysis. All analyses were conducted using STATA (version 16, Stata Corp).

2.2. Sensitivity analyses

Several sensitivity analyses were conducted. First, the model using nurse staffing ratios as a dependent variable was re‐estimated controlling for case mix index to make sure that severity of disease in each hospital is accounted for. Second, a model using total number of nursing hours (instead of productive hours) for the numerator of nurse staffing ratios was estimated. A standard year of 2080 nursing hours was used for a full‐time registered nurse.19 Third, models controlling for safety net status of hospitals were estimated. Fourth, models that did not control for partial exposure to Medicaid expansion were estimated. Fifth, models that included states with delayed Medicaid expansion (Alaska, Indiana, New Hampshire, and Pennsylvania) in the expansion group were also estimated. Finally, a Wald test was conducted to test the differences between the three coefficients of the postexpansion interaction terms in the primary analysis.

3. RESULTS

3.1. Characteristics of hospitals

Table 1 presents the pre‐expansion (2011‐2013) characteristics of sampled hospitals. The final dataset included 1403 hospitals in expansion states (40.0%) and 2101 hospitals in nonexpansion states (60.0%). Hospitals in expansion and nonexpansion states were statistically different in all characteristics. Expansion state hospitals tended to be larger, not‐for‐profit, urban, and teaching facilities compared to nonexpansion state hospitals. Approximately 73% of hospitals in expansion states were not‐for‐profit, whereas for‐profit and government‐owned hospitals each represented fewer than 20% of hospitals. In nonexpansion states, for‐profit and government‐owned hospitals combined represented approximately 55% of the hospitals. Teaching hospitals constituted 31.9% of hospitals in expansion states, but only 15.8% in nonexpansion states.

Table 1.

Baseline characteristics of hospitals in Medicaid expansion and nonexpansion states, pre‐expansion (2011‐2013)

Characteristic	Hospitals in expansion states	Hospitals in nonexpansion states
Size by number of beds (±SD)	159.3 (±184.2)	134.1a, *** (±217.5)
Ownership status (%)
Not‐for‐profit	72.3	44.9***
For‐profit	10.6	26.5***
Government‐owned	17.1	28.7***
Urban (%)	56.0	50.1***
Teaching (%)	31.9	15.8***
Hospitals (N)	1403	2101

Abbreviation: SD, standard deviation.

^a P‐values from t tests for each hospital characteristic in expansion states, compared with nonexpansion states.

^*** P < .001.

3.2. Changes in mean outcomes over years

Table 2 presents the unadjusted changes of mean and percentage changes of the two primary outcomes from 2011 through 2016 in expansion and nonexpansion states. From 2011 through 2016, nurse staffing ratios increased from 6.1 to 7.2 registered nurses hours per adjusted patient days (18.0% increase) in hospitals located in expansion states while it only increased by 6.2% in hospitals in nonexpansion states. Hospital‐wide readmission rates decreased by 4.9% in hospitals in expansion states and decreased by 3.8% in hospitals in nonexpansion states in 2016, compared with 2011.

Table 2.

Changes in mean of uncompensated care costs, Medicaid revenues, nurse staffing ratios, and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates, 2011‐2016a

Outcome measure	2011	2012	2013	Pre‐expansion (2011‐2013)	2014	2015	2016	Percent change 2011‐2016
Nurse staffing ratiosb (N = 20 770)c
Expansion states	6.1**	6.3**	6.9	6.4**	7.0	7.3	7.2	+18.0%
Nonexpansion states	6.5	6.7	7.0	6.7	6.8	7.0	6.9	+6.2%
30‐day hospital‐wide all‐cause unplanned Medicare readmission ratesd (N = 19 867)
Expansion states	16.2***	15.8***	15.8***	15.9***	15.4***	15.7***	15.4***	–4.9%
Nonexpansion states	15.9	15.5	15.5	15.6	15.2	15.5	15.3	–3.8%

P‐values from t tests for each outcome in each year in expansion vs nonexpansion states.

^aEach outcome measure has its own sample size that excludes its missing values.

^bFull‐time equivalent registered nurses productive hours per adjusted patient days.

^cN is number of observations in each analysis.

^dRates in percent. A readmission rate of 16.2% is equivalent to 1620 readmissions per 10 000 patients.

^** P < .05, ***P < .001

3.3. Parallel trends tests

Table 3 presents the regression results for the parallel trends tests for nurse staffing ratios and readmission rates, comparing hospitals in expansion and nonexpansion states before Medicaid expansion. The coefficients of the interaction term between expansion and linear time trend for both nurse staffing ratios and readmission rates are statistically nonsignificant and close to zero (0.03 registered nurses hours per adjusted patient days, P = .48, and .02 readmissions per 10 000 patients, P = .98). These findings indicate that the pre‐expansion trends for nurse staffing ratios and readmission rates were not statistically different in hospitals in expansion and nonexpansion states, confirming that the parallel trends assumption of the difference‐in‐difference model holds in this study's analyses (Parallel trends tests results for uncompensated care costs and Medicaid revenues in the Appendix S1).

Table 3.

Tests of pre‐expansion trends for nurse staffing ratios and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates, 2011‐2013

Independent variable	Nurse staffing ratios	30‐day hospital‐wide all‐cause unplanned Medicare readmission rates (per 10 000 patients)
Expansion × Pre‐yeara	0.03	0
Pre‐yearb	0.15***	−19***
Expansionc	−0.96	54***
Number of hospital beds	0.00	0
For‐profit	1.14***	20***
Government‐owned	0.28	28***
Urban	1.25***	−9**
Teaching	0.48**	13***
Observations (N)	10 202	9897

Each column presents results of a difference‐in‐difference model, controlling for hospital random effects and states' fixed effects.

^aAn interaction term of expansion and linear trend of pre‐expansion years.

^bA linear trend of pre‐expansion years.

^cExpansion is defined as hospitals located in expansion states.

^** P < .05, ***P < .001.

3.4. Changes associated with Medicaid expansion

Table 4 presents results from the difference‐in‐difference models, using the mean of each outcome of interest as the dependent variable. Nurse staffing ratios increased by 0.33, 0.42, and 0.46 registered nurses hours per adjusted patient days in 2014, 2015, and 2016 in hospitals located in expansion states, compared with nonexpansion states after expansion. This increase was statistically significant (P < .001) in 2015 and 2016, but marginally significant (P = .016) in 2014. Hospital‐wide readmission rates decreased by 9, 16, and 18 per 10 000 patients (P < .001) in 2014, 2015, and 2016, respectively, in expansion vs nonexpansion states hospitals after expansion.

Table 4.

Changes in nurse staffing ratios and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates in expansion states compared to nonexpansion states

Independent variable	Nurse staffing ratios	30‐day hospital‐wide all‐cause unplanned Medicare readmission rates (per 10 000 patients)
Expansion × Post3a	0.46***	−18***
Expansion × Post2b	0.42***	−16***
Expansion × Post1c	0.33**	−9**
Expansiond	–0.82	47***
Postexpansion3e	0.05	−35***
Postexpansion2f	0.09	−8***
Postexpansion1g	−0.01	−57***
Number of hospital beds	0.00	0
For‐profit	0.79***	20***
Government‐owned	0.42**	20***
Urban	0.88***	−5**
Teaching	0.19	14***
Observations (N)	20 770	19 867

Each column presents results of a difference‐in‐difference model, controlling for hospital random effects and states' fixed effects.

^aAn interaction term of expansion and third year of postexpansion (2016). The coefficient represents the change in outcome for hospitals in expansion states, compared with those in nonexpansion states in the 2016 postexpansion year.

^bAn interaction term of expansion and second year of postexpansion (2015). The coefficient represents the change in outcome for hospitals in expansion states, compared with those in nonexpansion states in the 2015 postexpansion year.

^cAn interaction term of expansion and first year of postexpansion. The coefficient represents the change in outcome for hospitals in expansion states, compared with those in nonexpansion states in the 2014 postexpansion year.

^dExpansion is defined as hospitals located in expansion states.

^ePostexpansion3 indicates the 2016 postexpansion year.

^fPostexpansion2 indicates the 2015 postexpansion year.

^gPostexpansion1 indicates the 2014 postexpansion year, adjusted for partial exposure to Medicaid expansion.

^** P < .05, ***P < .001.

3.5. Sensitivity analyses

When analyses controlling for case mix index, using total nursing hours, controlling for safety net status of hospitals, not controlling for partial exposure to Medicaid expansion, or not excluding states with delayed expansion were conducted, results were similar to those in the primary analysis. The results from Wald test indicate that there are statistically significant differences between the coefficients of the three postexpansion interaction terms (except for nurse staffing ratios), indicating that the magnitude of the association between Medicaid expansion and the outcome measures has changed over time after expansion (Results in the Appendix S1 Tables).

4. DISCUSSION

This study examined the association between Medicaid expansion and hospitals' nurse staffing ratios, and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates. Increases in nurse staffing ratios occurred in the three years in an ascending order, although this finding was marginally statistically significant in the first postexpansion year (2014). Medicaid expansion was associated with significant reductions in 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates in all three postexpansion years. The reduction in readmission rates increased over time.

Results concerning uncompensated care costs and Medicaid revenues build on previous findings that Medicaid expansion was associated with improvement in these two hospitals' financial measures in 2014 and 2015.9, 10, 12 This study finds that uncompensated care declined gradually between 2014 and 2016, but the declining rate was the same in 2015 and 2016 (each compared with pre‐expansion years). Growth in Medicaid revenues also began in the first postexpansion year and continued during the following two years, indicating that more patients receiving care in the hospital had Medicaid coverage and fewer patients received charity care or resulted in bad debt.

This is the first study that examines the association between Medicaid expansion and nurse staffing ratios and readmission rates. A significant association was observed between Medicaid expansion and an increase in nurse staffing ratios and a reduction in 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates, indicating improvement in quality of care provided in hospitals located in states with Medicaid expansion. Simply put, Medicaid expansion was independently associated with two indicators of hospitals' quality and two indicators of increased financial resources. These additional financial resources may have been used to improve quality of care. Previous research had reported that hospitals under financial pressures were more likely to have adverse patient safety events and might choose to reduce investment in processes that support delivery of care, including compliance with selected quality measures of the Joint Commission on Accreditation of Healthcare Organization.32, 33 Although Medicaid expansion was associated with increases in nurse staffing ratios in the three years, the association was only statistically significant in the second and third postexpansion years, suggesting that the impact was delayed until the end of the first postexpansion year of Medicaid expansion.

This study has a few limitations. HCRIS is the only dataset that contains information on hospitals' financial measures; however, the data have quality issues. Despite CMS tries to provide up‐to‐date, accurate HCRIS data, CMS is not responsible for any misrepresented, misinterpreted, or altered data. Further, potential outliers in the Cost Report were addressed by winsorizing the data per case at the 99th percentile. In addition, it is worth noting here that Medicaid revenues in expansion states were double the revenues in nonexpansion states and even the hospitals' characteristics in expansion and nonexpansion states were significantly different at baseline. However, the difference‐in‐difference model with hospitals' random effects and states' fixed effects controls for any unobservable differences between the characteristics of the two groups. Hospitals' random effects accounted for any unobserved heterogeneity among hospitals while states' fixed effects accounted for unobserved characteristics within states. Including random effects at the hospital level and fixed effects at the state level in the models accounts also for other unobservable characteristics such as coding, sociodemographic characteristics, or any potential quality improvement related to Medicaid expansion.34, 35 In 2017, CMS issued updates to the Cost Report Worksheet S‐10, allowing hospitals to have the option to resubmit an amended cost report for Fiscal Year (FY) 14 and FY 15 if they had additional data for lines 20, 22, 25, and 26. The data used in this paper should reflect the latest Cost Report, including all the updates.36 Another limitation is that the available data included only three postexpansion years. Monitoring the trends in finances, staffing, and quality over a longer period will be useful. This study only examined nurse staffing ratios and 30‐day hospital‐wide all‐cause unplanned Medicare readmission rates. Further studies should examine a broader array of quality measures as data on these indicators become available. Despite findings on nurse staffing ratios and readmission rates are statistically significant, some may argue that the magnitude is small. Nevertheless, an increase of 0.33‐0.46 productive nursing hours (19‐28 minutes) relative to an average of 6.8 hours is considered meaningful, especially for patients. The magnitude of the findings on readmission rate is comparable to other studies.37, 38 Findings concerning the association between Medicaid expansion and improvement in hospitals' financial measures and quality are early findings, and further studies need to examine the long‐term effects of Medicaid expansion on hospitals' financial performance and quality of care.

One of the goals of expanding health insurance coverage was to provide access to care for persons and protection from financial adversity because of illness.39 A potential consequence of Medicaid expansion was the improved financial status of hospitals which otherwise would be delivering a significant amount of uncompensated care.40 Hospitals provided about $28 billion of such care in 2015. These results indicate that at least initially, hospitals benefiting from Medicaid expansion also experienced improvement in nurse staffing ratios and readmission rates. It is important to continue monitoring the impact of Medicaid expansion over a longer period of time and for an expanded set of staffing and quality measures.

CONFLICT OF INTEREST

The author has no conflict of interest to declare.

DISCLAIMER

The findings and conclusions in this paper are those of the author and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Supporting information

 

 

Tarazi WW. Associations between Medicaid expansion and nurse staffing ratios and hospital readmissions. Health Serv Res. 2020;55:375–382. 10.1111/1475-6773.13273