HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response View responses Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Tsai, Y.-W. Articles by Hu, T.-w. Search for Related Content PubMed PubMed Citation Articles by Tsai, Y.-W. Articles by Hu, T.-w. Related Collections Ethics Insurance Health Professionals Other Maternal and Infant Health

Yi-Wen Tsai, PhD and Teh-wei Hu, PhD

Yi-wen Tsai is with the National Health Research Institutes, Taipei, Taiwan. Teh-Wei Hu is with the School of Public Health, University of California, Berkeley.

Correspondence: Requests for reprints should be sent to Yi-Wen Tsai, PhD, Division of Health Policy Research, National Health Research Institutes, 2F, 109, Min-Chuan East Road, Section 6, Taipei 114, Taiwan, ROC (e-mail: ivytsai{at}nhri.org.tw).

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 

Objectives. Taiwan’s National Health Insurance Program (NHI) was implemented on March 1, 1995. This study analyzed the influences of the Case Payment method of reimbursement for inpatient care and of physician financial incentives on a woman’s choice for primary cesarean delivery.

Methods. Logistic regressions were used to analyze 11 788 first-time deliveries in a nonprofit hospital system between March 1, 1994, and February 29, 1996.

Results. After implementation of the NHI’s Case Payment scheme, the likelihood that a woman would choose primary cesarean delivery increased by four to five times compared with the choice behavior of uninsured individuals prior to NHI (P < .0001).

Conclusion. Out-of-pocket payment discourages the selection of primary cesarean delivery. No robust statistics were found relating physician financial incentives to delivery choice.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 
On March 1, 1995, the Taiwan government instituted a national health insurance (NHI) program providing universal health insurance coverage. Before the introduction of NHI, nearly 60% of Taiwan’s population was enrolled in 1 of 3 major social insurance systems: the Labor Insurance scheme, the Government Employees’ Insurance scheme, and the Farmers’ Health Insurance scheme. To control the rapid growth of medical expenses, the Bureau of National Health Insurance developed several payment methods. For instance, the Case Payment method reimburses inpatient care, including vaginal delivery (since May 1995) and cesarean delivery (since October 1996), on the basis of a constant fee schedule per case. Cheng and Chiang found that NHI significantly increased the use of inpatient and outpatient medical care and that this effect was particularly prominent with respect to previously uninsured patients.¹

In this study, we investigated the effects of Case Payment on primary cesarean delivery, which is defined as a woman’s choice of cesarean delivery for her first childbirth. In addition, we analyzed the relationship between physician financial incentives and the likelihood of primary cesarean delivery.

Before the introduction of NHI, all 3 social insurance programs covered the fees for cesarean deliveries, whereas Government Employees’ Insurance was the only program that also covered the fees for vaginal deliveries. The 3 insurance programs required no coinsurance from enrollees. NHI is similar to Government Employees’ Insurance in that it covers both vaginal and cesarean deliveries and requires no coinsurance for child deliveries. Compared with self-paying schemes, Government Employees’ Insurance and NHI thus present patients with significantly smaller price differences between the 2 delivery options. Both Labor Insurance and Farmers’ Health Insurance covered only cesarean deliveries in full and provided a fixed-rate subsidy for vaginal deliveries, often forcing patients to pay out of pocket to meet the full charge.

The 3 social insurance programs preceding NHI reimbursed health care providers on a fee-for-service basis. The difference between the fees paid to physicians for cesarean deliveries and those paid for vaginal deliveries ranged from NT $64 000 to NT $71 000 (US $1 = NT $32). In contrast, the Case Payment method of NHI reimburses health care providers for medical services on a per-case basis. Reimbursement for complicated cases is based on a fee-for-service scheme, but the total percentage of fee-for-service cases per provider cannot exceed a ceiling set by the Bureau. This payment method is designed to increase health care providers’ cost consciousness. In addition, NHI reduced the reimbursement differences between cesarean and vaginal deliveries by around NT $15 000 to limit physicians’ financial incentive to perform cesarean deliveries.

Our hypothesis was that women covered by NHI would be more likely than uninsured women to undergo cesarean deliveries but less likely than those covered by Labor Insurance or Farmers’ Health Insurance before the introduction of NHI. The insurance effect of NHI on primary cesarean deliveries was similar to, but weaker than, that of Government Employees’ Insurance, largely owing to the differences in the reimbursement policies: NHI implemented Case Payment, and Government Employees’ Insurance applied general fee for service. We examined the influences of physician financial incentives on the use of primary cesarean deliveries. Because the reimbursement fees for cesarean deliveries were higher than those for vaginal deliveries, there were financial motives for physicians to perform more cesarean deliveries. This was particularly true when a physician’s actual earnings fell short of his or her desired income (the actual earnings and desired income factors are defined in the following section).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 
Data
The data used in our study were drawn from the largest nonprofit hospital system in Taiwan, which encompasses 3 hospital branches and a total of more than 6000 inpatient beds. Childbirth observations are those coded with the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM)² codes of all-patient diagnosis-related groups (AP-DRGs) 370, 371, 372, 374, 375, 650, 651, and 652. From March 1, 1994, through February 29, 1996, 11 788 primary deliveries took place within the hospital system; they were attended by 26 obstetricians and gynecologists.

Variables and Analysis
Two logistic regressions—one an ordinary logistic regression (model 1) and the other a random-effects logistic regression (model 2)—were used to compare the likelihood of selection of primary cesarean delivery by NHI and pre-NHI insurance enrollees with that of its selection by uninsured patients. Other covariates in the model included physician’s unmet desired income, clinical conditions, maternal age and education level, convenience, and hospital branches.

The selection of primary cesarean delivery was coded with the ICD-9-CM procedure codes of 74.0, 74.1, 74.2, 74.4, and 74.9. Insurance programs were classified into 4 groups (INS1–INS4), with the uninsured group (INS4) used as reference group. The Labor Insurance and Farmers’ Health Insurance schemes (INS1) and the Government Employees’ Insurance scheme (INS2) were the major insurance plans before the advent of NHI. NHI was coded as INS3. The Farmers’ Health Insurance group was combined with the Labor Insurance group into a single variable (INS1), partly because this insurance plan accounted for only 3% of the study population. In addition, the payment method of Farmers’ Health Insurance was similar to that of Labor Insurance. The total population of 11 788 first-time deliveries represented childbirth observations across the 2 study periods: the pre-NHI period (51.1% of the study’s cases) and the post-NHI period (48.9%). During the first period, Labor Insurance covered nearly 57.3% of the study population pre-NHI, which had a cesarean delivery rate of 50.7%. Among those who were uninsured pre-NHI, the cesarean delivery rate was 8.3%. During the second period, women covered under NHI had an average cesarean delivery rate of 33% (Table 1).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 
The estimated results from the 2 models (Table 1) showed that clinical indications were the strongest correlates of cesarean deliveries compared with other factors. With regard to insurance effects, when the effects of other covariates were held constant, insured women were more likely than uninsured women to undergo primary cesarean deliveries. Women covered by NHI had primary cesarean delivery rates that were higher than those of uninsured women, similar to those of women covered by Government Employees’ Insurance, and much lower than those of women covered by Labor Insurance. The insurance effects continued to hold in the random-effects model. After the effects of other variables were controlled, primary cesarean deliveries were also found to be less likely to occur on Saturdays (odds ratio [OR] = 0.58, P < .01). This result was quite robust, given that the odds ratios for Saturday were consistent in the 2 model specifications.

Results from model 1 showed that a woman was more likely to have a cesarean delivery when attended by a physician who had failed to meet the previous month’s desired income. The DIFF variable showed that reaching desired income in excess of NT $100 000 decreased the likelihood of a primary cesarean delivery at a rate of 4% per NT $100 000 (P < .01), although the effects of this financial incentive were not as robust as one might expect (model 2).

Physicians at the lower professional levels (levels 2 and 3, respectively) were 27% (P < .01) and 33% (P < .001) less likely to perform primary cesarean deliveries than were physicians at level 1, the effects of other variables being held constant. These estimates, however, were not significant in model 2. Women at the 2 medical centers (HOSP1 and HOSP2) were 36% (P < .001) and 72% (P < .001) less likely, respectively, than those at HOSP3 to have primary cesarean deliveries.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 
This study can be viewed as a case study of the situation in Taiwan’s nonprofit health care system. Because this nonprofit hospital system serves a diverse population of patients, our research findings provide useful information about physician and patient behavior.

The results showed that clinical indications were the strongest correlates of cesarean deliveries. As to the economic factors, health insurance was significantly associated with primary cesarean delivery rates. NHI increased the likelihood of primary cesarean deliveries by a factor of 4 to 5 compared with the uninsured category (P < .0001), but NHI’s effects were significantly lower than those of most pre-NHI insurance programs. This finding suggests that out-of-pocket payment influenced patients’ decisions regarding type of delivery and therefore supports the notion that cost sharing is a viable strategy for cost containment. Physicians may perform cesarean deliveries on weekends for patients’ convenience—justifying their demand for leisure time. In our study, there was no robust evidence that physician financial incentive affected the decision to perform cesarean deliveries. Future studies should use either samples taken from more sites or samples taken over a longer period.

	Acknowledgments

 
The authors are grateful for data support from Yi-Chou Chuang, chief executive officer of Chang-Gung Memorial Hospital in Taiwan, and for the suggestions provided by reviewers of this journal.

Human Participant Protection
No protocol approval was needed for this study.

	Footnotes

 
Peer Reviewed

Y.-W. Tsai identified the study issue, designed methods, analyzed the data and drafted the manuscript. T.-wei Hu participated in study design, data analysis, and manuscript editing.

Accepted for publication October 5, 2001.

	References

TOP ABSTRACT INTRODUCTION METHODS RESULTS CONCLUSIONS References

 
1. Cheng SH, Chiang TL. The effect of universal health insurance on health care utilization in Taiwan: results from a natural experiment. JAMA. 1997;278:89–93.[Abstract]

2. National Center for Health Statistics. International Classification of Diseases, Ninth Revision, Clinical Modification. Hyattsville, Md: National Center for Health Statistics; 1980. DHHS publication PHS 80-1260.

This article has been cited by other articles:

				R. P. Kauffman One Cheer for Choice Obstet. Gynecol., February 1, 2007; 109(2): 457 - 457. [Full Text] [PDF]

				H.-C. Lin and S. Xirasagar Institutional Factors in Cesarean Delivery Rates: Policy and Research Implications Obstet. Gynecol., January 1, 2004; 103(1): 128 - 136. [Abstract] [Full Text] [PDF]

				C.-N. Liu and M.-C. Yang RETHINKING "WOMAN'S CHOICE" OF CESAREAN DELIVERY Am J Public Health, July 1, 2003; 93(7): 1036 - 1037. [Full Text] [PDF]

eLetters:

This Article Abstract Full Text (PDF) Submit a response View responses Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Tsai, Y.-W. Articles by Hu, T.-w. Search for Related Content PubMed PubMed Citation Articles by Tsai, Y.-W. Articles by Hu, T.-w. Related Collections Ethics Insurance Health Professionals Other Maternal and Infant Health