Abstract
The use of stochastic frontier models for inference on hospital efficiency is complicated by the inability to fully control for quality differences across hospitals. Additionally, the potential existence of cross-sectional dependence due to the presence of unobserved common factors leads to endogeneity problems that can bias both cost function and efficiency estimates. Using a panel consisting of 1518 hospitals for the years 1996–2013 (T = 18), I adopt techniques for dealing with long, cross-sectionally dependent panel data in order to estimate cost parameters and hospital specific efficiency. In particular, I employ the estimation technique proposed by Bai (Econometrica 77(4):1229–1279, 2009), which assumes that the unobservable heterogenous effects have a factor structure. I find evidence of considerable scale economies and that hospital cost inefficiencies have been increasing during the period of 1996–2013, and that the growth in expenditures is, in part, driven by spending that increases patient satisfaction, but that does not significantly contribute to improved patient health outcomes.
Similar content being viewed by others
Notes
For details on hospital value based purchasing see: https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/hospital-value-based-purchasing/index.html?redirect=/Hospital-Value-Based-Purchasing/
Of particular relevance to this study are SFA studies, with a health care focus, that have used panel data; these include: Chirikos (1998), Chirikos and Sear (2000), Deily et al. (2000), Li and Rosenman (2001), Rosko (2001, 2004), Sari (2003), McKay and Deily (2005), Deily and McKay (2006), Furukawa et al. (2010), Rosko and Mutter (2010, 2014).
For excellent reviews of the SFA literature, see: Kumbhakar and Lovell (2000), Greene (2008) and Belotti (2012). For recent applications of the factor model methodology to production frontier analysis, see: Ahn et al. (2007), Mastromarco and Serlenga (2009), Kneip et al. (2012), Ahn et al. (2013), Filippini and Tosetti (2014) and Mastromarco et al. (2015).
Additional details on the potential empirical benefits of the factor model approach are provided within Appendix A.
Rather than having to imposing a set structure upon the modeled inefficiency, a factor model approach allows the researcher to use data-driven tests to determine the appropriate number of factor components to include. For details on these tests, see, e.g., Bai and Ng (2002), Bai (2004), Onatski (2009), Alessi et al. (2010), Kneip et al. (2012), Ahn and Horenstein (2013).
For additional details on the within-group interpretation of the interactive fixed effects model, please see Bai (2009, pp. 1248–1249).
In practice, this means that a interactive fixed effects approach will results in the same estimates as the fixed effects model, assuming this is indeed the true underlying model. However, in the presence of cross-sectional dependence within the data, due to latent common factors, the interactive fixed effects model is able to capture and control these latent features. As such, in motivating the use of an interactive fixed effects approach it is important to test for cross-sectional dependence using, e.g., the Pesaran (2014) cross-sectional dependence test.
An alternative example of a possible latent factor would be trends in the adoption/use of state-of-the-art medical technologies within the US health care market. For a more detailed explanation of the endogeneity problem caused due to such latent structures, please see Appendix A.
Regarding the notation throughout: the linear form is due to having taken logs. The “ln” prefix is omitted to ease the notation. That is, e.g. \(TC_{it}\) should really say ln\(TC_{it}\). Also, bold notation denotes vectors, e.g. \(x_{it}\) is not a vector, but \({\textbf{x}}_{it}\) is.
The literature on stochastic cost frontier analysis in health care has mainly consisted of working with either the Cobb-Douglas functional form or the more flexible specification of a trans-log form that nests the former as a special case (see e.g. Rosko et al., 2008 for a relevant survey of this literature). The choice to work with the Cobb-Douglas form is done for its simplicity here, however, a robustness check using the translog cost functional form is also provided within Appendix D.
The notation \({\overline{z}}_{i}\) denotes the average across the time dimension. That is, \({\overline{z}}_{i}=\frac{1}{T}\left( \sum _{t}^{T}z_{it}\right)\).
The choice of taking the min(.) operator over both i, t or just i will affect the results and a discussion of this is provided within the results section.
Note that since the given Cobb-Douglas form does not allow for second-order terms, the value of scale economies is assumed to be constant irrespective of the output level (see e.g. Farsi et al. (2005b, p. 75)).
This section provides only some basic information regarding the data sources, for additional data details please see Appendix B.
The first of these comes from the Dartmouth atlas data, while the rest (including all variables in Table 3) are from the CMS’s hospital compare data.
This section draws on two computational resources for the purpose of performing estimation and various specification tests. These are: (i) the plm package by Croissant and Millo (2008), and (ii) the phtt package by Bada and Liebl (2015).
The tests for establishing the optimal number of factors vary in their recommendations from 1 to 4 factors (for details see Appendix C). Within the main analysis I employ a specification that controls for individual fixed effects and that allows for two factors.
The global CD test considers dependence across all hospitals within the whole of the US. The local CD test, on the other hand, considers dependence on a state level.
Running additional specification tests recommended by Bai (2009) and Kneip et al. (2012) I find that both of these tests reject the null hypothesis of there only being additive two-way-effects–again, lending support to a richer specification with interactive fixed effects. For additional details on these tests and the results from Augmented Dickey Fuller tests that reject the null of a unit root, see Appendix C.
This result appears robust when using even broader (25 or 50 mile) radius measures.
The finding of increasing scales within health care seems supported by the high volume of hospital mergers and system formation within the industry (again, for a nice review of this literature please see Gaynor and Town (2012)).
References
Ahn, Seung C., & Horenstein, Alex R. (2013). Eigenvalue ratio test for the number of factors. Econometrica, 81(3), 1203–1227.
Aigner, Dennis, Knox Lovell, C. A., & Schmidt, Peter. (1977). Formulation and estimation of stochastic frontier production function models. Journal of Econometrics, 6(1), 21–37.
Alessi, Lucia, Barigozzi, Matteo, & Capasso, Marco. (2010). Improved penalization for determining the number of factors in approximate factor models. Statistics & Probability Letters, 80(23), 1806–1813.
Bada, Oualid, & Liebl, Dominik. (2014). phtt: Panel Data Analysis with Heterogeneous Time. Journal of Statistical Software, 59(6).
Bai, Jushan. (2004). Estimating cross-section common stochastic trends in nonstationary panel data. Journal of Econometrics, 122(1), 137–183.
Bai, Jushan. (2009). Panel data models with interactive fixed effects. Econometrica, 77(4), 1229–1279.
Bai, Jushan, & Ng, Serena. (2002). Determining the number of factors in approximate factor models. Econometrica, 70(1), 191–221.
Battese, G. E., & Coelli, T. J. (1988). Prediction of firm-level technical efficiencies with a generalized frontier production function and panel data. Journal of Econometrics, 38(3), 387–399.
Battese, G. E., & Coelli, T. J. (1992). Frontier production functions, technical efficiency and panel data: With application to paddy farmers in India. Journal of Productivity Analysis, 3(1–2), 153–169.
Battese, G. E., & Coelli, T. J. (1995). A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empirical Economics, 20(2), 325–332.
Belotti, F., et al. (2012). Stochastic frontier analysis using Stata. Stata Journal, 3(ii), 1–39.
BLS. (2017). Local Area Unemployment Statistics. http://www.bls.gov/lau/%7B%5C#%7Dtables, visited on 10/01/2017
Census. (2017). SAIPE Program. https://www.census.gov/programs-surveys/saipe/data.html, visited on 10/01/2017
Chirikos, Thomas N. (1998). Further evidence that hospital production is inefficient. Inquiry, 35(4), 408–416.
Chirikos, Thomas N., & Sear, Alan M. (2000). Measuring hospital efficiency: a comparison of two approaches. Health Services Research, 34(6), 1389–1408.
Chudik, Alexander, & Pesaran, M. Hashem. (2013). Large panel data models with crosssectional dependence: a survey. Mimeo.
CMS. (2017a). Cost Reports. url: https://www.cms.gov/Research-Statistics-Dataand-Systems/Downloadable-Public-Use-Files/Cost-Reports/?redirect=/costreports/02_hospitalcostreport.asp, visited on 10/01/2017
CMS. (2017b). Historical Impact Files for FY 1994 through Present. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Historical-Impact-Files-for-FY-1994-through-Present.html, visited on 10/01/2017
CMS. (n.d). Hospital compare data archive. https://data.medicare.gov/data/archives/hospital-compare.
CMS. (2017c). Hospital Value-Based Purchasing. https://www.cms.gov/edicare/Quality-Initiatives-Patient-Assessment-Instruments/hospital-value-based-purchasing/index.html?redirect=/Hospital-Value-Based-Purchasing/, visited on 10/01/2017
Croissant, Y., & Millo, G. (2008). Panel data econometrics in R: The plm package. Journal of Statistical Software, 27(2), 1–43. issn: 1548-7660.
Deily, Mary E., & McKay, Niccie L. (2006). Cost inefficiency and mortality rates in Florida hospitals. Health Economics, 15(4), 419–431.
Deily, Mary E., McKay, Niccie L., & Dorner, F. (2000). Exit and inefficiency: The effects of ownership type. Journal of Human Resources, 35(4), 734–747.
Devers, Kelly J., Brewster, Linda R., & Casalino, Lawrence P. (2003). Changes in hospital competitive strategy: A new medical arms race? Health Services Research, 38(1p2), 447–469.
Dzau, Victor J., et al. (2017). Vital directions for health and health care: priorities from a National Academy of Medicine Initiative. Jama, 317(14), 1461.
Eberhardt, Markus, Helmers, Christian, & Strauss, Hubert. (2013). Do Spillovers matter when estimating private returns To R &D ? The Review of Economics and Statistics, 95(May), 436–448.
Farsi, Mehdi, Filippini, Massimo, & Greene, William. (2005). Efficiency measurement in network industries: Application to the Swiss Railway companies. Journal of Regulatory Economics, 28(1), 69–90.
Farsi, Mehdi, Filippini, Massimo, & Kuenzle, Michael. (2005). Unobserved heterogeneity in stochastic cost frontier models: An application to Swiss nursing homes. Applied Economics, 37(December), 2127–2141.
Farsi, Mehdi, Filippini, Massimo, & Kuenzle, Michael. (2006). Cost efficiency in regional bus companies: An application of alternative stochastic frontier models. Journal of Transport Economics and Policy, 40(1), 95–118.
Ferrier, G. D. (2014). Improving the “usefulness’’ of stochastic frontier analysis for health care. Economics and Business Letters, 3(1), 27–34.
Filippini, Massimo, & Tosetti, Elisa. (2014). Stocastic Frontier Models for Long Panel Data Sets: Measurement of the Underlying Energy Eciency for the OECD Countries. FRED, 2017, Economic Data. https://research.stlouisfed.org/fred2, visited on 10/01/2017
Furukawa, Michael F., Raghu, T. S., & Shao, Benjamin B M. (2010). Electronic medical records and cost efficiency in hospital medical-surgical units. Inquiry, 47(2), 110–123.
Goldman, Dana P., Vaiana, Mary, & Romley, John A. (2010). The emerging importance of patient amenities in hospital care. The New England Journal of Medicine, 363(23), 2185.
Green, William. (2011). Econometric Analysis (7th ed.). Pearson.
Greene, William. (2005). Fixed and random effects in stochastic frontier models. Journal of Productivity Analysis, 23(1), 7–32.
Green, William. (2005). Reconsidering heterogeneity in panel data estimators of the stochastic frontier model. Journal of Econometrics, 126(2), 269–303.
Hollingsworth, Bruce. (2003). Non-parametric and parametric applications measuring efficiency in health care. Health Care Management Science, 6(4), 203–218.
Hollingsworth, Bruce. (2008). The measurement of efficiency and productivity of health care delivery. Health Economics, 17(10), 1107–1128.
Hussey, Peter S., et al. (2009). A systematic review of health care efficiency measures: Health care efficiency. Health Services Research, 44(3), 784–805.
IOM. (2010). The healthcare imperative: lowering costs and improving outcomes: workshop series summary. National Academies Press.
IOM. (2012). Best care at lower cost: The pathway to continuously learning health care in America. September. National Academies Press.
Jondrow, James, et al. (1982). On the estimation of technical inefficiency in the stochastic frontier production function model. Journal of Econometrics, 19(2–3), 233–238.
Kneip, Alois, Sickles, Robin C., & Song, Wonho. (2012). A new panel data treatment for heterogeneity in time trends. 28, 03 590–628.
Kumbhakar, Subal C., & Lovell, C. A. (2012). Stochastic Frontier Analysis (2nd ed., pp. 1–204). Cambridge University Press.
Lee, Y. H., & Schmidt, Peter. (1993). A production frontier model with flexible temporal variation in technical inefficiency. In H. Fried, C. A. K. Lovell, & S. Schmidt (Eds.), The measurement of productive efficiency: techniques and applications. Oxford Newey: Oxford University Press.
Li, T., & Rosenman, R. (2001). Cost inefficiency in Washington hospitals: a stochastic frontier approach using panel data. Health Care Management Science, 4(2), 73–81.
Mas-Colell, A., Whinston, M. D., & Green, J. R. (1995). Microeconomic Theory. Oxford University Press.
Mastromarco, Camilla, & Serlenga, Laura. (2009). Dynamic Efficiency Modelling for Technological Convergence (pp. 1–35). Mimeo.
Mastromarco, Camilla, Serlenga, Laura, & Shin, Yongcheol. (2013). Globalisation and technological convergence in the EU. Journal of Productivity Analysis, 40(1), 15–29.
Mastromarco, Camilla, Serlenga, Laura, & Shin, Yongcheol. (2015). Modelling technical efficiency in cross sectionally dependent stochastic frontier panels. Journal of Applied Econometrics, 31(1), 281–297.
McKay, Niccie L., & Deily, Mary E. (2005). Comparing high- and low-performing hospitals using risk-adjusted excess mortality and cost inefficiency. Health Care Management Review, 30(4), 347–360.
Mutter, Ryan L., Rosko, Michael D., & Wong, Herbert S. (2008). Measuring hospital inefficiency: The effects of controlling for quality and patient burden of illness. Health Services Research, 43(6), 1992–2013.
Newhouse, Joseph P. (1994). Frontier estimation: How useful a tool for health economics? Journal of Health Economics, 13(3), 317–322.
Onatski, Alexei. (2010). Determining the number of factors from empirical distribution of eigenvalues. The Review of Economics and Statistics, 92(4), 1004–1016.
Pitt, Mark M., & Lee, Lung-fei. (1981). The measurement and sources of technical inefficiency in the Indonesian weaving industry. Journal of Development Economics, 9, 43–64.
Rosko, Michael D. (2001). Cost efficiency of US hospitals: A stochastic frontier approach. Health Economics, 10(6), 539–551.
Rosko, Michael D. (2004). Performance of U.S. teaching hospitals: A panel analysis of cost inefficiency. Health Care Management Science, 7(1), 7–16.
Rosko, Michael D., & Mutter, Ryan L. (2008). Stochastic frontier analysis of hospital inefficiency. Medical Care Research and Review, 65, 131–166.
Rosko, Michael D., & Mutter, Ryan L. (2010). Inefficiency differences between critical access hospitals and prospectively paid rural hospitals. Journal of Health Politics, Policy and Law, 35(290), 95–126.
Rosko, Michael D., & Mutter, Ryan L. (2014). The association of hospital cost-inefficiency with certificate-of-need regulation. Medical Care Research and Review, 71(3), 280–298.
Roth, Jean. (2017). NBER ZIP Code Distance Database. http://www.nber.org/data/zip-code-distance-database.html, visited on 10/01/2017,
Sari, Nazmi. (2003). Efficiency outcomes of market concentration and managed care. International Journal of Industrial Organization, 21(10), 1571–1589.
Schmidt, Peter, & Sickles, Robin C. (1984). Production frontiers and panel data. Journal of Business and Economic Statistics, 2(4), 367–374.
The Dartmouth Atlas of Health Care. (2017). Hospital Discharges and Post-Acute Care. url: http://www.dartmouthatlas.org/tools/downloads.aspx?tab=41, visited on 10/01/2017,
Truven Health Analytics. (2016). 100 Top Hospitals. http://100tophospitals.com/studies-winners/100-top-hospitals/year, visited on 08/01/2016
Wang, Hung-jen. (2002). One-step and two-step estimation of the effects of exogenous variables on technical efficiency levels. Journal of Productivity Analysis, 18, 129–144.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Linde, S. Hospital cost efficiency: an examination of US acute care inpatient hospitals. Int J Health Econ Manag. 23, 325–344 (2023). https://doi.org/10.1007/s10754-023-09356-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10754-023-09356-x
Keywords
- Cost inefficiency
- Economies of scale
- Principal components
- Stochastic frontier analysis
- Value-based purchasing