The impact of a diagnosis-related group-based prospective payment experiment: the experience of Shanghai

Abstract

The purpose of this study is to examine the impact of the 2004 and 2005 Diagnosis-Related Group (DRG)-based Prospective Payment Experiments (whereby a ceiling was set on per case payments for 15 DRGs) of the health insurance system in Shanghai using Differences-In-Differences (DID) and Differences-in-Differences-in-Differences (DDD) empirical strategies. The results show that the Length Of Stay (LOS) of all inpatients suffering from the target diseases during the two test periods responded quickly to the experiment (even though only insured inpatients living in Shanghai were eligible for the DRG-based experiments), thus implying that the DRG-based prospective payment system does not induce physicians to significantly reduce inpatient expenditures but does induce them to reduce the LOS of those suffering from the target diseases. Moreover, costs were shifted to uninsured inpatients suffering from the target diseases: their overall inpatient expenditures increased by 33.6 percentage points more than that of uninsured inpatients during the nontest period in 2005. Thus, the experiments were of limited effectiveness, and cost-shifting to uninsured inpatients was observed during at least the latter test period.

Acknowledgements

I am most grateful to Charles Yuji Horioka for his helpful comments. I also thank Kooyiti Kawabuti, Midori Wakabayashi and Shizuka Sekita for their assistance.

Notes

¹The data are limited to inpatients in one of six diagnostic categories.

²‘Prepayment’ as used here is different from Diagnosis-Related Group Prospective Payment Systems (DRG-PPS) (case-based). Rather, it is more akin to a global budget or a close-and-volume contract.

³Only 16 hospitals and 409 diseases were chosen as the target of the reform in Ha'erbin.

⁴The reform in the city of Zhenjiang covers all A and B class hospitals and applied to 82 types of diseases. If inpatients suffering from a target disease are not cured completely, the hospital will be paid only 80% of the cap for that target disease.

⁵If certain conditions are met, the cost of outpatient care will also be paid from the fund.

⁶The target diseases are as follows (the last six types were added in 2005): (1) spontaneous labour; (2) caesarean section; (3) appendectomy; (4) abdominal cholecystectomy; (5) peritoneoscopic cholecystictomy; (6) thyrophymoctomy; (7) hysterectomy; (8) mastoncoctomy; (9) artificial lens implantation; (10) discogenictomy; (11) varicosity; (12) hernioplasty; (13) prostatic operation; (14) abdominal ovariotomy; and (15) peritoneoscopic ovariotomy.

⁷There are nine classes of hospitals in China ranging from C − to A + with A + being the highest, and the deductible amount is 10% of the average income of workers in the previous year. The average income of workers in Shanghai was 22 160 yuan in 2003 and 24 398 yuan in 2004 (The Statistical Yearbook of Shangai, 2005).

⁸The test periods were October to December in 2004 and 2005 (a total of 6 months).

⁹The amount of the deposit depends on the nature and severity of the diagnosis, whether the patient is insured, and an additional deposit is sometimes required during hospitalization.

¹⁰In addition, the duration of the experiments was only 3 months, so the change of the composition of patients was negligible.

¹¹Official government data on hospitals in the same class are used here (The Hygiene Yearbook of China, 2004).

¹²The available information on patients includes age, gender, disease, whether insured and whether a citizen of Shanghai. The ratio of insured inpatients, uninsured patients from Shanghai and uninsured patients from other cities is 5:3:2.