Prices for Common Services at Quaternary vs Nonquaternary Hospitals

Abstract

This study uses commercial claims data to assess whether quaternary hospitals charge higher prices for common, unspecialized services also offered by nonquaternary hospitals.

Hospital spending, the largest share of US national health expenditures, accounted for $1.3 trillion in 2021.^1,2 Price variation is a major driver of spending variation among the privately insured.³ Quaternary hospitals, which offer highly specialized, uncommon services, have anecdotally faced scrutiny for charging higher commercial prices than nonquaternary hospitals.^4,5 Besides their typically larger size and academic affiliations, their reputational value deters their exclusion from payer networks, a key negotiating strategy to lower prices.^5,6 We assessed whether quaternary hospitals charged higher prices for common, unspecialized services also offered by nonquaternary hospitals.

Methods

Using commercial claims from the Health Care Cost Institute (HCCI) 2.0 data set and hospital data from the Healthcare Cost Report Information System (HCRIS) from 2017 to 2019, we compared in-network prices (inpatient or outpatient facility plus physician fees) for common services offered by quaternary and nonquaternary hospitals. HCCI data contain more than 1 billion annual commercial claims from more than 60 million US beneficiaries covered by large insurers; HCRIS provides hospital data reported to the Centers for Medicare & Medicaid Services. Quaternary hospitals were defined as those with 5 or more solid organ transplants from 2017 to 2019 captured in the HCCI data. Chosen for their high frequency and relative homogeneity, we defined common services as uncomplicated inpatient admissions for hip replacement, knee replacement, pneumonia, and diabetes, as well as outpatient lower joint magnetic resonance imaging, pulmonary spirometry, screening colonoscopy, and screening mammography (eTable in Supplement 1). Differences in characteristics of patients using these services in quaternary vs nonquaternary hospitals were assessed using standardized mean differences. Because prices vary widely among services, we focused on the percentage differences in prices between the hospital groups. The quaternary price premiums for each service individually and as pooled inpatient and outpatient categories were estimated using linear regression, both unadjusted and adjusted for patient characteristics (sex, age, Charlson Comorbidity Index score), insurance plan type, hospital market share, hospital teaching status, market (commuting zone), and year. Statistical significance was defined at the 2-sided P < .05 level. Analyses were conducted in Stata version 17.0 (StataCorp). Detailed methods are in the eMethods in Supplement 1.

Results

We included 4 893 907 care episodes in 152 quaternary and 4278 nonquaternary hospitals. Patient characteristics were similar between the 2 hospital groups (all standardized mean differences <0.1) (Table 1). Unadjusted and adjusted quaternary price premiums were qualitatively similar (Table 2). The adjusted pooled inpatient price was 8.2% (95% CI, 2.8%-13.6%; P = .003) higher for common admissions in quaternary vs nonquaternary hospitals. Adjusted prices for admissions for hip and knee replacements were, respectively, 10.2% (95% CI, 4.2%-16.1%; P = .001; adjusted mean price difference, $3450; unadjusted means, $37 683 vs $33 924) and 7.6% (95% CI, 1.1%-14.1%; P = .02; adjusted difference, $2433; unadjusted means, $35 753 vs $32 122) higher at quaternary hospitals. Adjusted prices for diabetes and pneumonia admissions were, respectively, 10.0% (95% CI, 3.2%-16.7%; P = .004; adjusted difference, $924; unadjusted means, $10 739 vs $9299) and 8.3% (95% CI, 0.5%-16.1%; P = .04; adjusted difference, $767; unadjusted means, $11 011 vs $9300) higher in quaternary hospitals.

Table 1. Characteristics of the Study Sample.

Characteristic	Nonquaternary sample	Quaternary sample	Standardized mean differencea
No. of included care episodes	4 166 586	727 321
No. of hospitals	4278	152
Hospital teaching statusb
Major teaching	276	112
Minor teaching	882	41
Nonteaching	3293	3
No. of hospital-yearsc	12 200	454
No. of commuting zones	688	81
No. of hospitals per commuting zone
1	77	41
2	114	24
3	103	6
4-6	206	9
≥7	186
Hospital discharge market share, mean (SD), %d	15.8 (23.5)	17.3 (17.3)	0.065
Demographic and clinical characteristics, %
Sexe
Male	8.8	10.9	0.074
Female	91.2	89.1	−0.074
Age group, y
0-17	1.3	2.4	0.097
18-24	1.1	1.8	0.074
25-34	1.5	2.6	0.091
35-44	14.3	15.5	0.034
45-54	37.3	36.0	−0.026
55-64	44.7	41.6	−0.062
Episode year
2017	14.1	14.7	0.016
2018	43.6	42.8	−0.016
2019	42.3	42.5	0.005
Charlson Comorbidity Index score, mean (SD)f	0.17 (0.49)	0.20 (0.59)	0.05
0	86.2	85.5	−0.019
1	11.5	11.3	−0.007
2	1.9	2.3	0.03
≥3	0.4	0.9	0.066
Insurance plan
PPO	66.3	62.6	−0.079
POS	28.4	31.1	0.059
HMO	3.8	4.7	0.048
Other	1.5	1.6	0.012

Abbreviations: HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization.

^a Standardized mean difference is calculated as the difference in group means divided by the pooled SD for each variable and allows comparison of group differences on a common scale. Values less than 0.25 signify good balance.

^b Some hospitals changed teaching status during the study period, so total hospitals from the hospital teaching status rows may not add up to the total number of unique hospitals.

^c Not all hospitals had qualifying care episodes for all years.

^d Market shares are constructed based on all hospital discharges from all payers in a study year. Markets are defined as commuting zones, which are groupings of counties developed by the US government to reflect local economies and commuting patterns.

^e Excluding mammography, men account for 44.1% and 44.6% of included care episodes in nonquaternary and quaternary hospitals, respectively (standardized mean difference, 0.009).

^f Charlson Comorbidity Index score is an index of specific disease burden that predicts mortality risk. The scale reflects the sum of each relevant diagnostic code that is present in the patient multiplied by the weight of that particular diagnostic code.

Table 2. Prices for Common, Relatively Uncomplicated Services at Quaternary and Nonquaternary Hospitals.

Service	No.	Mean (SD) price, $	Difference, % (95% CI)		P valueb
			Unadjusted	Adjusteda
Inpatient pooled c , d
Nonquaternary		NA	[Reference]	[Reference]	[Reference]
Quaternary		NA	10.7 (4.9 to 16.6)	8.2 (2.8 to 13.6)	.003
Hip replacement
Nonquaternary	55 849	33 924 (14 599)	[Reference]	[Reference]	[Reference]
Quaternary	9701	37 683 (16 879)	11.5 (4.7 to 18.3)	10.2 (4.2 to 16.1)	.001
Knee replacement
Nonquaternary	85 073	32 122 (13 791)	[Reference]	[Reference]	[Reference]
Quaternary	11 199	35 753 (15 262)	10.1 (4.0 to 16.2)	7.6 (1.1 to 14.1)	.02
Pneumonia
Nonquaternary	13 815	9300 (6341)	[Reference]	[Reference]	[Reference]
Quaternary	1872	11 011 (11 103)	10.3 (1.4 to 19.3)	8.3 (0.5 to 16.1)	.04
Diabetes
Nonquaternary	11 882	9299 (6113)	[Reference]	[Reference]	[Reference]
Quaternary	2459	10 739 (8552)	10.9 (3.1 to 18.7)	10.0 (3.2 to 16.7)	.004
Outpatient pooled d
Nonquaternary		NA	[Reference]	[Reference]	[Reference]
Quaternary		NA	3.4 (−3.9 to 10.6)	3.7 (−2.0 to 9.4)	.20
Lower joint MRI
Nonquaternary	336 724	1251 (827)	[Reference]	[Reference]	[Reference]
Quaternary	52 558	1379 (854)	11.7 (0.0 to 23.4)	11.8 (1.9 to 21.7)	.02
Spirometry
Nonquaternary	59 989	217 (183)	[Reference]	[Reference]	[Reference]
Quaternary	61 894	254 (163)	21.4 (6.5 to 36.3)	14.1 (2.4 to 25.8)	.02
Colonoscopy
Nonquaternary	257 602	1874 (1133)	[Reference]	[Reference]	[Reference]
Quaternary	36 833	2197 (1435)	13.0 (−0.5 to 26.5)	14.0 (7.0 to 20.9)	<.001
Mammography
Nonquaternary	3 345 660	282 (131)	[Reference]	[Reference]	[Reference]
Quaternary	550 805	285 (135)	0.7 (−7.4 to 8.9)	0.9 (−5.3 to 7.1)	.77

Abbreviations: MRI, magnetic resonance imaging; NA, not applicable.

^a Adjusted difference controls for patient characteristics (sex, age, and Charlson Comorbidity Index score), insurance plan type, hospital market share, hospital teaching status, market (commuting zone), and year.

^b P values are in reference to the adjusted differences.

^c Inpatient admissions included are those from the Diagnosis-Related Group category without a complication, comorbidity, or major complication or comorbidity.

^d The pooled percentage difference reflects the regression-estimated percentage difference in the mean price of the underlying inpatient or outpatient services between the 2 hospital groups, weighted by the volume of each service in the aggregate.

For common outpatient services, the adjusted pooled price was 3.7% (95% CI, −2.0% to 9.4%; P = .20) higher in quaternary vs nonquaternary hospitals although not statistically significant. Adjusted prices for lower joint magnetic resonance images, spirometry, and colonoscopies were 11.8% (95% CI, 1.9%-21.7%; P = .02; adjusted difference, $148; unadjusted means, $1379 vs $1251), 14.1% (95% CI, 2.4%-25.8%; P = .02; adjusted difference, $31; unadjusted means, $254 vs $217), and 14.0% (95% CI, 7.0%-20.9%; P < .001; adjusted difference, $262; unadjusted means, $2197 vs $1874) higher at quaternary vs nonquaternary hospitals, while the difference in adjusted prices for mammography, the most common of the outpatient services considered, was not statistically significant.

Discussion

Within the same geographic market, quaternary hospitals negotiated higher commercial prices than nonquaternary hospitals for several common inpatient admissions and outpatient studies. Prices did not differ for screening mammography, which may reflect competition from nonhospital imaging facilities. Limitations of this study included potential unmeasured variation in patient characteristics, a narrow definition of quaternary hospitals, data from a subset of insurers, a narrow set of services selected for their relative homogeneity but with possible residual quality variation, and commuting zones proxying for geographic markets. Whether price premiums for common services performed by quaternary hospitals cross-subsidize specialized services, enable other valuable activities, or deliver meaningful quality benefits remain questions for future study. Yet to the extent that price premiums reflect market power and bargaining leverage against insurers derived from offering quaternary services, additional regulation and/or policies that promote competition may be appropriate.

Section Editors: Jody W. Zylke, MD, Deputy Editor; Karen Lasser, MD, and Kristin Walter, MD, Senior Editors.

Supplement 1.

eMethods

eTable 1. Definitions of Inpatient and Outpatient Services Used in Analysis

eReferences

Supplement 2.

Data Sharing Statement