Cost variation and revisit rate for adult patients with asthma presenting to the emergency department

Abstract

Background:

Asthma is common, resulting in 53 million emergency department (ED) visits annually. Little is known about variation in cost and quality of ED asthma care.

Study objective:

We sought to describe variation in costs and 7-day ED revisit rates for asthma care across EDs. Our primary objective was to test for an association between ED costs and the likelihood of a 7-day revisit for another asthma exacerbation.

Methods:

We used the 2014 Florida State Emergency Department Database to perform an observational study of ED visits by patients ≥18 years old with a primary diagnosis of asthma that were discharged home. We compared patient and hospital characteristics of index ED discharges with and without 7-day revisits, then tested the association between ED revisits and index ED costs. Multilevel regression was performed to account for hospital-level clustering.

Results:

In 2014, there were 54,060 adult ED visits for asthma resulting in discharge, and 1,667 (3%) were associated with an asthma-related ED revisit within 7 days. Median cost for an episode of ED asthma care was $597 with an interquartile range of $371–980. After adjusting for both patient and hospital characteristics, lack of insurance was associated with higher odds of revisit (OR 1.42, 95% CI 1.18–1.71), while private insurance, female gender, and older age were associated with lower odds of revisit. Hospital costs were not associated with ED revisits (OR = 1.00; 95% CI 1.00–1.00).

Conclusion:

Hospital costs associated with ED asthma visits vary but are not associated with odds of ED revisit.

Introduction

Asthma affects 19 million U.S. adults (1), resulting in an estimated $50.3 billion in annual medical costs as of 2013 (2). One contributor to these costs, and an important marker of morbidity, are the 1.22 million asthma-related visits to emergency departments (EDs) that occur every year (3, 4). Given the disease burden and costs associated with asthma, it is important to optimize transitions of care for this population. Even after appropriate emergency care, patients with asthma may decompensate resulting in a return to the ED. Patient education improves the likelihood of primary care follow-up and treatment adherence (5), which may decrease the risk of ED revisits. Thus, improving transitions of care after ED discharge for asthma patients has been proposed as an area for quality improvement.

The American College of Emergency Physicians (ACEP) has identified ED revisit rate as one indicator of the quality of emergency care for patients with asthma (6). Among all pediatric patients, 4–5% patients will revisit an emergency department within 7 days (7, 8). Among adult ED patients with ‘upper respiratory tract infections’, there is a 4.9% three-day revisit rate (9). Roughly 9.2% of adult asthma patients relapse - as defined by any urgent/unplanned ED, clinic, or physician office visit – within seven days of receiving emergent asthma care in the ED (10).

At the same time, the Merit-Based Incentive Program System (MIPS) track of the Quality Payment Program (QPP) implemented in 2015 requires cost be considered as one factor in determining up to 30% of overall reimbursement from Medicare in 2022. The cost category incorporates episode-based cost measures for targeted conditions, and a cost measure for emergency department visits for asthma is under consideration for development (11). Prior research suggests the average charge for an ED visit for asthma is $1,502 (12); however, variations in costs for asthma across EDs has yet to be described. The relationship between cost and quality remains poorly defined across healthcare settings (ie, hospital or outpatient) and medical specialties (13). When episode-based cost measures are implemented—for example, as part of the QPP regulation mandated by the Centers for Medicare and Medicaid Services comparing the costs of emergency care costs between and within clinician groups—it is possible that clinicians expending fewer resources per episode will be awarded bonus payments regardless of quality (14). However, it is unknown if increased costs are associated with improved outcomes for patients with asthma. We sought to use a statewide ED database to describe variation in costs for asthma care and subsequent revisit rates across emergency departments. We hypothesized that there was wide cost variation among emergency departments. Furthermore, we hypothesized there was no relationship between resources utilized for an initial episode of care and the likelihood of a patient returning to an ED for asthma care.

Materials and Methods

Data and Sample

We performed an observational study using the Healthcare Cost and Utilization Project (HCUP) State Emergency Department Database (SEDD). SEDD includes patient level data on emergency department visits resulting in discharge. Data elements include all listed patient diagnoses, patient demographics (e.g., age, race, and insurance status), total charges and discharge information. We used the 2014 Florida SEDD because it had the largest volume of adult ED visits in the state dataset with a revisit variable verification rate >97% at the time of the analysis (15). This study was determined to be exempt from review by the Institutional Review Board of the Icahn School of Medicine at Mount Sinai.

The study population included ED visits with a primary diagnosis of asthma as identified by Agency for Healthcare Research and Quality’s (AHRQ) Clinical Classifications Software (CCS) codes (CCS code 128). We restricted our analyses to visits by adults (e.g., ≥ 18 years old) because the pathogenesis of childhood and adult asthma are known to be different, thus requiring different clinical approaches (16, 17). We focused on patients who were discharged because inpatient costs are not attributable to a single provider. In addition, there are established guidelines for the outpatient management of asthma in EDs (18). Given the co-occurrence of respiratory infections and acute asthma exacerbation, visits with a primary diagnosis of “Other Upper Respiratory Infections” (CCS 126), “COPD and Bronchiectasis” (CCS 127), and “Acute Bronchitis” (CCS 125) were included when there was a concurrent secondary diagnosis code for asthma. These CCS codes were selected as they were the most prevalent secondary diagnosis codes when asthma was the primary diagnosis code in a prior national analysis (accounting for 6.12%, 3.79%, and 2.93% of ED visits, respectively) (4). Finally, any visits occurring at hospitals with fewer than 30 ED visits for asthma in a year and subjects with missing charge or demographic data were excluded (0.9%).

Study outcomes and variables

We performed a descriptive analysis to identify patient and hospital predictors of ED revisits and variation in costs. Our primary outcome of interest was the association between index ED visit cost and likelihood of ED revisit within 7 days. Revisits related to asthma were identified using the same diagnostic CCS codes as listed above. We selected 7 days for our primary outcome based on prior evidence suggesting potentially avoidable ED revisits occur 3 to 9 days after the initial visit and regulatory programs include 7 days as a quality measure for unplanned hospital-based encounters (19, 20). In our preliminary analyses, we performed descriptive statistics on the distribution of revisit rates. SEDD provides data on total facility charges (i.e., amount billed by hospitals) for ED visits. Given that charges may not represent the actual expenses incurred, to estimate costs (i.e., actual expenses incurred), we use the hospital charges supplemental file and applied hospital-specific cost-to-charge ratios from HCUP to estimate cost (21, 22).

Patient and Hospital characteristics

Patient demographics included age, gender, median income of the residence zip code, and a combined variable for race and ethnicity (White, Black, Hispanic, other). Charlson comorbidity index scores were calculated from secondary ICD-9 diagnosis codes to adjust for the burden of comorbid chronic conditions among patients. Hospital characteristics included urban vs rural location and annual ED visit volume.

Statistical Methods

We performed descriptive statistics to describe patient and hospital characteristics associated with unadjusted 7-day revisit rates. The distribution of revisits and costs were assessed using histograms and scatter plots. From these graphs, we determined whether any statistical transformations would be necessary prior to performing adjusted multivariable analyses. Descriptive statistics, including median and interquartile range, were used to further characterize variability in costs. We calculated the coefficient of interquartile variation (CQV) by $CQV=\frac{(Q3-Q1)}{(Q3+Q1)}X100$ (23). The CQV was used to assess the extent and dispersion of variability in the distribution of charges and costs.

We then used multivariable regression to assess if there was an association between ED visit costs and the likelihood of a revisit after adjusting for patient and hospital characteristics. We adjusted for race and ethnicity as black and Hispanic patients have been shown to have more severe asthma resulting in higher utilization and worse outcomes (24). Because both costs and revisit rates are likely to be clustered within the same ED, we fit a multilevel regression model using generalized linear models with mixed effects at the level of the hospital given that standard errors are more likely to be similar for patients treated at the same hospitals (vs. patients treated at different hospitals). All analyses were performed in Stata/SE 15.1.

Results

From 8,168,641 total ED visits in Florida in 2014, we identified 54,060 index adult asthma ED visits resulting in discharge across 187 hospitals. The mean age was 41 (SD 16), 42.3% were white, and the subjects were mostly female (65.5%) (Table 1). Most of the cohort had a Charlson comorbidity score of 1–2 (85.8%). Compared to all index asthma visits discharged home, ED revisits were less frequent among female patients (56% vs 65.5%) and privately insured patients (15.4% vs 22.8%), and were more frequent among uninsured (43% vs 32.7%) and Black (40% vs 33.6%) patients. Figure 1 demonstrates the distribution of days to revisit from an index visit among all study subjects. Using a 7-day cutoff, we identified 1,670 patients (3.1% of the initial sample) who experienced an asthma-related ED revisit.

Table 1.

Patient and hospital characteristics among ED asthma discharges by index visit and those with a revisit within 7 days.

Variables	Index ED asthma discharge		Asthma or Asthma related revisit within 7 days
	(N =54,060)		(N=1,667)
	No.	%	No.	%
Age (SD)	41 (16))	N/A	38 (14)	N/A
Gender
Male	18,673	34.5	734	44
Female	35,387	65.5	933	56
Insurance
Medicare	8,603	15.9	190	11.4
Medicaid	15,452	28.6	500	30
Private insurance	12,321	22.8	256	15.4
Self-Pay/No Charge/Other/Unknown	17,684	32.7	721	43.3
Race
White	22,852	42.3	654	39.2
Black	18,154	33.6	670	40.2
Hispanic	11,657	21.6	308	18.5
Other/Unknown	1,397	2.6	35	2.1
Charlson Score Groups
Score 0	0	0	0	0
Score 1 or 2 (1)	46,373	85.8	1,443	86.6
Score 3 or more (2)	7,687	14.2	224	13.4
Hospital Type
Private Hospital	21,218	39.3	667	40
Urban Hospital	32,211	59.6	969	58.1
Rural Hospital	631	1.2	31	1.9
Median household income national quartile for patient ZIP Code
Quartile 1	27,866	51.6	893	53.6
Quartile 2	16,459	30.5	489	29.3
Quartile 3	6,581	12.2	192	11.5
Quartile 4	2,077	3.8	51	3.1
Missing	1,077	2.0	42	2.5
Average Annual Total ED Visit Volume per Hospital	85,086		85,085
Median charge at index visit	$2,952		$2,839
Median cost at index visitⱡ	$597		$572

^ⱡCosts were calculated as a function of charge using hospital specific cost-to-charge ratios provided by the Healthcare Cost and Utilization Project.

Figure. 1.

Distribution of ED Revisits by Days after Index Discharge

For the index visit, median charge and cost were $2,952 (IQR: $1774–4841) and $597 (IQR: $371–980) respectively. For patients revisiting the ED, median charge and cost at the index visit were $2,839 (IQR: $1704–4598) and $572 (IQR: $351–971) respectively. Slightly over 20% of patients incurred costs more than $1000. The coefficient of quartile variation on the index ED visit was 46 and 45 for charges and cost, respectively.

In the simple logistic model adjusting for patient and hospital characteristics, Black patients (OR = 1.20 ; 95% CI 1.07–1.34), those with Medicaid insurance (OR = 1.23 ; 95% CI 1.00–1.51), self-pay/unknown (OR = 1.46 ; 95% CI: 1.20–1.78), and index ED visits occurring at a rural, non-profit hospital (OR = 1.57 ; 95% CI 1.08–2.27) were associated with higher odds of 7-day ED revisit (Table 2). Female patients (OR = 0.71; 95% CI 0.65–0.79), older age (OR 0.99; 95% CI 0.987–0.995) and those with private insurance (OR = 0.79; 95% CI 0.64–0.98) had lower odds of 7-day revisits. After additionally adjusting for clustering by hospital site, uninsured patients were still more likely to experience 7-day ED revisits (OR = 1.42; 95% CI 1.18–1.71), while female patients (OR = 0.72, 95% CI 0.65–0.80), older age (OR 0.99; 95% CI 0.986–0.994), and privately insured patients (OR = 0.79, 95% CI 0.64–0.97) continued to be have lower odds of revisit. There was no association between initial ED visit cost and revisit rate after adjusting for both patient and hospital characteristics (OR = 1.00 ; 95% CI 1.00–1.00, Table 2).

Table 2.

Visit-level association between ED cost and likelihood of revisit adjusting for patient and hospital characteristics.

	Multivariable logistic regression		Multilevel regression with generalized linear models
Variable	OR	95% CI	OR	95% CI
Cost	1.00	1.00–1.00	1.00	1.00–1.00
Age	0.99*	0.987–0.995	0.99*	0.986—0.994
Female	0.71*	0.65 – 0.79	0.72*	0.65–0.80
Insurance
Medicare	ref	ref	ref	ref
Medicaid	1.23*	1.00–1.51	1.21	0.997–1.47
Private	0.79*	0.64–0.98	0.79*	0.64–0.97
Self-pay, other, unknown	1.46*	1.20–1.78	1.42*	1.18–1.71
Race
White	Ref	ref	ref	ref
Black	1.20*	1.07–1.34	1.12	0.997–1.27
Hispanic	0.90	0.78–1.03	0.89	0.77–1.03
Other / Unknown	0.82	0.58–1.16	0.82	0.58–1.16
Charlson Score Group ⱡ
Score 1–2	Ref	ref	ref	ref
Score 3+	1.02	0.88–1.18	1.03	0.89–1.21
Hospital type
For profit	Ref	ref	ref	ref
Non-profit urban	0.92	0.83–1.03	0.88	0.75–1.04
Non-profit rural	1.57*	1.08–2.27	1.58	0.99–2.52
Median household income *
Quartile 1	Ref	ref	ref	ref
Quartile 2	0.98	0.87–1.09	1.01	0.90–1.14
Quartile 3	1.02	0.87–1.20	1.06	0.90–1.23
Quartile 4	0.94	0.70–1.25	0.94	0.70–1.26
Missing	1.24	0.90–1.70	1.32	0.96–1.81
Hospital volume	1.00	1.00–1.00	1.00	1.00–1.000

^*- p < 0.05

^ⱡ- Patients could not have a Charlson score of 0 as everyone in the study cohort has a designation of chronic pulmonary disease.

Discussion

In this analysis, we observed variation in costs for an ED asthma discharges and no association between cost and ED revisit rate. Rather we observed associations between revisit rates and patient characteristics, including age, gender, and insurance status, that persisted after adjusting for confounders.

The lack of an association between cost and ED revisit rate suggests an opportunity to improve value among the subset of facilities with high charges and no difference in revisit outcomes. Cost variation among ED asthma discharges is likely driven by variation in hospital facility fees, which have been shown to vary widely across institutions and have been implicated as a major driver of ED charges (25, 26). Prior studies have demonstrated wide variation in costs and charges related to care for emergency medical conditions (26–28). Our findings are consistent with prior work demonstrating little association between overall health care costs and outcomes (13, 29, 30) with one recent exception relevant to emergency care, which suggested that Medicare patients transported by EMS to higher cost hospitals have lower mortality (31). Payment reform programs like episode-based measures mandated by the QPP are intended to drive value by improving quality while reducing costs. Clinician quality and cost scores achieved in the QPP are used to adjust their total Medicare reimbursement as mandated by federal law. Initially, QPP scores were based primarily on quality; however, the quality category receives lower weight over time, while the cost category increases so that by 2022, each is weighted at 30% (32). Therefore, while clinicians would maximize their reimbursement by achieving high quality and low cost, it is possible for some clinicians to have significantly lower costs and average or below-average quality and still receive QPP scores that quality for financial incentives.

Our findings demonstrate that lack of insurance had the largest magnitude association with ED revisits after adjusting for patient and hospital factors. While this finding is unsurprising given the barriers to obtaining prescribed medications and follow-up care among the uninsured, it suggests that ED revisits for asthma may be less of an indicator of quality and more of an indicator of the patient population served by the ED. Thus, payment programs based on outcomes that do not consider socioeconomic drivers of health outside of health care providers’ control may result in unintended consequences, as observed in the disproportionately higher financial penalties experienced by safety-net facilities as a consequence of the Hospital Readmissions Reduction Program (which then further reduce the resources and capacity for safety-net facilities to improve quality) (33, 34).

We also found that younger, male patients had higher odds of revisit, consistent with prior studies (9, 35, 36). While female asthma patients tend to have higher severity disease, they may also be more likely to report a usual source of care and follow up (37), which may reduce the likelihood of an ED revisit. While older patients tend to be sicker and have higher acute care needs; however, older patients are also more likely to have Medicare, which likely provides improved access to prescriptions and outpatient care compared to Medicaid or no insurance. Targeted interventions to improve care transitions for these higher risk subgroups - such as appropriate use of steroids, providing patients with inhalers on ED discharge or engaging in patient education – may improve care transitions and long-term asthma management (38–40).

Our findings have several limitations common to administrative datasets. First, we were not able to access information on the severity of clinical presentation, such as vital signs and respiratory status, which might drive both revisits and resource expenditures; however, this is mitigated in part by selecting a subset of patients who were discharged to home. Secondly, we could not measure patient-centered outcomes such as symptom improvement. Further some ED revisits are due to natural progression of disease and adherence to appropriate return precautions, and thus do not necessarily represent adverse outcomes. However, there are no validated measures to capture symptom improvement and revisit rate has been identified as a patient-centered adverse outcome (41). Our dataset may have limited national generalizability given that we restricted our analysis to the state of Florida in 2014; however, Florida is the third most populous state in the US and has a high diversity index (42). Further, as previously discussed, Florida is the only state with a high verification (>97%) of the ED revisit variable at the time of the analysis (15). Finally, the charge-to-cost ratio only accounts for facility fees and excludes any professional fees involved in an episode of care. While we likely underestimated the costs for an episode of care, the facility fees have historically exceeded the professional fees and likely represent most of the costs.

Conclusion

There is variability in the costs and revisits for asthma across emergency departments. Increased costs do not appear to have a meaningful impact on short-term future healthcare utilization. However, female sex, older age, and lack of private insurance were associated with a higher adjusted odds of a subsequent ED revisit. While reducing variation in costs is an opportunity to improve value, the implementation of value-based payment programs should consider the unintended consequences these policies may have on EDs who disproportionately serve high-cost, high-need populations.