Can We Measure Hospital Quality from Physicians' Choices? with Ricardo Mora and Antonio Romero-Medina, CEPR DP6850, 2008.

Publications

Can We Measure Hospital Quality from Physicians' Choices?

with Ricardo Mora and Antonio Romero-Medina, CEPR DP6850, 2008.

Abstract: In this paper, we propose an alternative methodology for ranking hospitals based on the choices of Medical School graduates over hospital training vacancies. Our methodology is therefore a revealed preference approach. Our methodology for measuring relative hospital quality has the following desirable properties: a) robust to manipulation from hospital administrators; b) conditional on having enough observations, it allows for differences in quality across specialties within a hospital; c) inexpensive in terms of data requirements, d) not subject to selection bias from patients nor hospital screening of patients; and e) unlike other rankings based on experts' evaluations, it does not require physicians to provide a complete ranking of all hospitals. We apply our methodology to the Spanish case and find, among other results, the following: First, the probability of choosing the best hospital relative to the worst hospital is statistically significantly different from zero. Second, physicians value proximity and nearby hospitals are seen as more substitutable. Third, observable time-invariant city characteristics are unrelated to results. Finally, our estimates for physicians' hospital valuations are significantly correlated to more traditional hospital quality measures.

Other references to our paper:

http://www.voxeu.org/index.php?q=node/1231

Moral hazard and the demand for health services: a matching estimator approach

With Anna Sanz de Galdeano and Pedro Pita Barros

Journal of Health Economics, Vol 27, July 2008, pp 1006-1025 

Abstract: We estimate the impact of extra health insurance coverage beyond a National Health System on the demand for several health services. Traditionally, the literature has tried to deal with the endogeneity of the private (extra) insurance decision by finding instrumental variables. Since a priori instrumental variables are hard to find we take a different approach. We focus on the most common health insurance plan in Portugal, ADSE, which is given to all civil servants and their dependents. We argue that this insurance is exogenous, i.e., not correlated with beneficiaries' health status. This identifying assumption allows us to estimate the impact of having ADSE coverage on the demand for three different health services using a matching estimator technique. The health services used are number of visits, number of blood and urine tests, and the probability of visiting a dentist. Results show large positive effects of ADSE coverage for number of visits and tests among the young (18 to 30 years old) but only the latter is statistically significantly different from zero. The effects represent 21.8 and 30 percent of the average number of visits and tests for the young. On the contrary, we find no evidence of moral hazard on the probability of visiting a dentist.

Keywords: demand for health services, matching estimator, moral hazard, Portuguese health system

CEPR Discussion Paper 5488
Latest version (February 2006):[ PDF ]

Benchmarking for Productivity Improvement: a health care application

With Daniel A. Ackerberg and Michael H. Riordan

International Economic Review, vol47 nº1, February 2006

Abstract: A methodology is developed and applied to compare the performance of publicly funded agencies providing treatment for alcohol abuse in Maine. The methodology estimates a Wiener process that determines the duration of completed treatments, while allowing for agency differences in the effectiveness of treatment, costs of treatment, standards for completion of treatment, patient attrition, and the characteristics of patient populations. Notably, the Wiener process model separately identifies agency fixed effects that describe differences in the effectiveness of treatment ("treatment effects"), and effects that describe differences in the unobservable characteristics of patients ("population effects"). The estimated model enables hypothetical comparisons of how different agencies would treat the same populations. The policy experiment of transferring the treatment practices of more cost-effective agencies suggests that Maine could have significantly reduced treatment costs without compromising health outcomes by identifying and transferring best practices.

Keywords: benchmarking, relative performance, Wiener process, structural estimation, alcohol abuse, productivity

WP version:NBER Working paper 8385

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=877102

Executive summary

Bilateral Market Power and Vertical Integration in the Spanish Electricity Spot Market

With Kai-Uwe Kühn

Previously presented and distributed as: "Market Power and Vertical Integration in the Spanish Electricity Market"

Abstract:

The Spanish electricity spot market is highly concentrated both on the seller and the buyer side. Furthermore, unlike electricity spot markets in other deregulated electricity systems, large buyers and sellers are typically vertically integrated. This allows both large net sellers and large net buyers to strategically influence the spot market price. We develop a supply function model of this market to analyze the impact of market power on prices and productive efficiency and use it empiricially to detect such bilateral market power. Our estimates suggest that market power has had little impact on spot market prices but that substantial productive inefficiencies arise from the excercise of bilateral market power.

Keywords: bilateral market power, electricity markets, market power test, supply function equilibrium, vertical integration

Recently listed on SSRN's Top Ten download list for Studies: Market Structure (Topic) Recent Hits.

CEPR Discussion Paper N. 4590
Latest version (February 2006):[ PDF ]

Substance Abuse Treatment: what do we know? An Economist's perspective

European Journal of Health Economics, Vol. 6 (1), pp 53-64, March 2005

Abstract: The literature on treating substance abuse has dealt basically with four important ques-tions: (a) Is treatment effective? (b) Are all programs equally effective? (c) Why do pro-grams differ in their effectiveness? (d) Which treatments are more cost-effective? This pa-per reviews the substance abuse treatment literature around these four questions and discusses methodological issues that hin-der the interpretation and generalization of results to date. The answer to the first ques-tion is a sounding “yes,” treatment is effec-tive but not all programs are equally effec-tive. Researchers have moved beyond the “black box” literature that concentrated on patient and program characteristics as expla-nations for differences in effectiveness and search for the “active” ingredients of treat-ment. These include, for example, the treat-ment philosophy of the program’s director and staff attitudes towards patients. Cost-ef-fectiveness studies are less common, and their conclusions are mixed. In general, it is probably safe to say that for the majority of patients, outpatient or shorter programs are more cost-effective.

Keywords: substance abuse; alcohol misuse; survey; treatment evaluation; effectiveness

A Consistent Estimator for the Binomial Distribution in the Presence of Incidental Parameters, an application to Patent Data

Journal of Econometrics, vol 119 (1) pp73-98, March 2004.

Abstract: In this paper a consistent estimator for the Binomial distribution in the presence of incidental parameters, or fixed effects, when the underlying probability is a logistic function is derived. The consistent estimator is obtained from the maximization of a conditional likelihood function in light of Andersen’s work. Monte Carlo simulations show its superiority relative to the traditional maximum likelihood estimator with fixed effects also in small samples, particularly when the number of observations in each cross-section, T, is small. Finally, this new estimator is applied to an original dataset that allows the estimation of the probability of obtaining a patent.

Keywords: Count data, Fixed effects, Conditional likelihood, Binomial, Patents

Dollars and Performance: Treating Alcohol Misuse in Maine

Journal of Health Economics, vol 20 (4) pp 645-672, 2001.

Abstract:  

If public funds are allocated efficiently, then an increase in funding should improve the performance of substance abuse treatment programs. In the data used in this paper, performance (measured as abstinence rates) and expenditures per patient are not positively correlated. One explanation is that funding is endogeneous, i.e. programs treating more difficult patients receive more funding. The data comes from all Maine’s outpatient drug-free programs that received public funding between 1991 and 1994. After controlling for endogeneity, this paper concludes that the marginal impact of expenditures per patient on abstinence rates is small and statistically insignificantly different from zero.

Keywords: Alcohol, Substance abuse, Non-profit, Performance, Cost-effectiveness.