• 沒有找到結果。

管理式醫療的網絡效應

N/A
N/A
Protected

Academic year: 2021

Share "管理式醫療的網絡效應"

Copied!
9
0
0

加載中.... (立即查看全文)

全文

(1)

行政院國家科學委員會專題研究計畫 成果報告

管理式醫療的網絡效應

計畫類別: 個別型計畫 計畫編號: NSC92-2415-H-004-012- 執行期間: 92 年 08 月 01 日至 93 年 07 月 31 日 執行單位: 國立政治大學財政系 計畫主持人: 連賢明 報告類型: 精簡報告 處理方式: 本計畫可公開查詢

中 華 民 國 93 年 11 月 4 日

(2)

中文摘要 管理式醫療的一大特徵是建立合作網絡---健康保險機構選擇性和醫療供給者合作來 提供醫療服務。然而,文獻上對合作網絡為何建立以及合作網絡如何運作並沒有一個清楚 的解答。本文假設健康保險機構透過兩項機制管理合作網絡:排除(僅允許被保人使用進 入網路內醫療供給者),導引(將被保人轉移至較優良醫療供給者),這兩項機制來提高品 質和降低成本。由於無法清楚觀察到各供給者的治療品質和成本,保險機構利用『相對』 成本和品質來決定機制的使用。因此,本文推論過去的治療成本和品質和現在醫院是否納 入合作網絡和現有治療病患數存在相當關係。 透過 1995 至 1999 年間的麻省各主要保險機構受保婦女的懷孕住院資料,我們利用資 料驗證這兩項機制在網絡中扮演的角色。我們以孕婦的剖腹產比例當作品質的指標,以醫 院生產服務(常規性)定價和生產中非常規性和常規性費用比例作為成本指標。我們發現 排除這項機制在實際上很少被使用。其次,我們發現當過去醫療品質較差,或是成本較高 時,醫院從該保險機構獲得的現有病患數目下降;這些證據說明導引機制被較廣泛使用。 關鍵詞:合作網絡,醫療成本,醫療品質,管理式醫療

(3)

Abstract

One of the most prevailing practices in managed care is the creation of network---a managed plan selectively contracts with health providers to supply medical services. Yet, it is still unclear why and how the network is operated. This paper hypothesizes that the network is operated through two mechanisms: “exclusion”, enrollees are restricted to providers within the network, and “redirection”, enrollees are directed to favored providers in the network, to contain the health cost and improve the quality of care. Because a plan imperfectly observes the cost and quality of care, a plan utilizes these two mechanisms by comparing the cost and quality of a specific provider with that of the average. As a result, a link exists between the previous periods’ treatment performances and the exclusion and the patient flow of a provider at this period.

Using discharge records of baby deliveries in acute hospitals between 1995 and 1999, we empirical test the presence of the two mechanisms in the hospital networks of several large managed plans in Massachusetts. We measure the quality of care by the cesarean rate, and the hospital cost by the hospital charge on non-routine delivery as well as the ratio of routine over non-routine charges. We found that “exclusion” mechanism is rarely used in practice: very few hospitals are excluded from a plan’s network over the sample years. We also found that patients are directed away from hospitals with higher costs or lower quality, the evidence in support of “redirecting” mechanism.

(4)

1. Introduction

The landscape of the U.S health care industry is changing dramatically. Traditionally, the provision of health services is provided through the indemnity plan---an insurance where patients and providers decided on appropriate treatment, and insurers cover the expenses. Increasingly, however, the medical services are provided by managed typed plans---an insurance where providers and insurers are integrated, and jointly manage patients' health services. By 1995, over 75% of privately insured Americans receive care under some form of managed care; a percentage more than tripled since 1987 (Jensen et. al, 1997).1 And two public insurances: Medicare, the insurance for the old, and Medicaid, the insurance for the poor, both largely increase their shares in managed care in the recent years.

The growth of managed care results in significant impacts on the cost and quality of health care. On the macro level, the high penetration of managed care is reported to be the key reason for slowing the medical expenditure over the late eighties and early nineties (Glied, 2000). On the micro level, Culter et al. (2000) found that the treatment cost for heart diseases in HMO

enrollment is about 30 to 40 percent lower than those enrolled in traditional plans, in spite of a moderate outcome difference between managed and non-managed plans. While the evidence on the effect of managed care on the quality of care is still not conclusive (Luft and Miller, 1994, 1997), managed care unambiguously has changed the delivery of care in the medical sector.

It is not clear how the managed care does that, however. Managed care encompasses a diverse array of institutional arrangements (Glied, 2000). Aside from the conventional

cost-sharing (co-payments), managed care requires health providers to share risks so as to control the health case use (capitation); they review providers' provisions of care (utilization reviews), and denied their charges if necessary (denial of care). Likewise, enrollees in managed plans have to call their insurers before seeing their caregivers (gatekeeper), and are restricted to providers willing to supply care at discounted prices (the creation of network). The overall impact of managed care is thus a combination of various effects from all mechanisms. Each mechanism is distinct and its relative importance is unknown.

In this project, we analyze the creation of network----a health plan selectively contracts with providers to supply medical services---and its effect on care received. There are a number of reasons to focus particularly on this arrangement. First, almost every plan builds its network; its prevalence illustrates the importance. Second, many other mechanisms, though they are intended to explicitly restrict the health use, are reported to be less important in practice. Finally, the usual

1 Jensen et al (1997) report these numbers based on a survey of insurance plans from private employers. The growth of managed care is continuing. Morrisey and Jensen (1997) reports that small employers, which used to cling to conventional plans, are now mostly offering managed care products. Besides, according to the Health Insurance Association of America (2002), between 1994 and 1997, the market share of fee-for-service plans for private insurance coverage fell from 32.8 percent to 21.2 percent, while the share of PPOs grew from 44.8 percent to 47.3 percent, and the share of the most recent entrant to the managed care arena, point-of-service HMOs, more than doubled, rising from 1.8 percent to 4.8 percent. During the same period, traditional HMOs rose from 22.5 percent to 26.9 percent.

(5)

classification for managed care, such as PPO or HMO, to a large extent depends on how a plan manages its network. Obviously, the use of network plays a significant role in affecting the supplied care.

2. Provider’s Network and Enrollees’ Hospital Choice

Most literature on managed care focuses on the performance of managed care arrangements under different practices (Lufts, 1981; Miller and Luft, 1994, 1997; Vistnes, 1995; Culter and Sheiner, 1997; Town and Vistnes, 2001; Lien 2002; Dor et al, 2004). Among them, Ma and McGuire (2002) provide a theoretical model of network incentives in managed health care. Town and Vistnes (2001) develop a framework for analyzing bargaining relationships between hospitals and HMOs under selective contracting.

A managed-care organization may restrict enrollees’ hospital choice through two mechanisms, the exclusion and redirection mechanism. If a managed-care organization has perfect information regarding which hospital is profitable for them, then they would know exactly which hospitals they should contract with. The literature argues that managed-care organizations directly exclude hospitals that are not profitable (Town and Vistnes, 2001). By selectively contracting with profitable hospitals, managed-care organizations restrict their enrollees’ choice of hospital and maximize their profits.

However, this argument is incomplete as least in three respects. First, most of the

managed-care organizations in Massachusetts officially contract with almost all acute hospitals; enrollees in these managed-care organizations do not have face any formal restriction on their choice of hospital.2 Second, hospitals may face binding capacity contraints, which imply a short-term increase in cost. It is difficult for a managed-care organization to immediately exclude a hospital from a managed-care network even if the hospital causes financial loss. Third,

managed-care organizations usually do not have perfect information on hospital costs and quality. Nevertheless, managed-care organizations can still affect their enrollees’ choice of hospital through a redirection mechanism. Here, enrollees are influenced by managed-care organizations when selecting hospitals. The empirical relevance of this influence is strong. Even when large managed-care organizations contract with most hospitals in Massachusetts, in practice these managed-care organizations only use a small number of hospitals in their networks. In fact, the total market shares for the “top-10” hospitals in a given managed-care organization is high. For instance, the top ten hospitals in Harvard Plan, one of the main HMOs, accounts for more than 90 percent of the business.

In summary, if managed-care organizations do direct patients, they should direct patients to their those hospitals with lower costs or higher quality. However, because managed-care

organizations may lack full information, they direct patients to hospitals that they expect to be

2

The Harvard-Pilgrim bulletin of membership and the Tufts managed care organization website, both list hospitals for enrollees participating in HMO-type plans. Almost all acute hospitals in Massachusetts are a member of both networks.

(6)

profitable for them. The market share of hospital h in managed-care organization k is assumed to be a function of this expectation and other market variables. That is,

) , , ( , , , , k t h t h t h t t h f X s = Κ Ε (1)

where shk,t is hospital h’s market share in managed-care organization k in year t. 3

t h

X , is a vector of

market variables (demand and cost shifters). Κ is a variable indicating the status of capacity constraint h,t

of hospital h in year t.4 Ε is managed-care organization k’s expectation on profitability from hospital h. kh,t A managed-care organization will direct higher patient flow to a hospital that is expected to be profitable (high Ε ) for it. kh,t

Managed-care organizations’ expectations on profitability from hospitals are characterized in the following way. Assume that demand elasticity is constant and hospitals within a managed-care network compete in a Cournot fashion. Let p denote the unit service price paid by managed-care organizations to providers, and c the provider’s marginal cost. Under Cournot competition among network providers, the equilibrium-pricing rule is:

ε = − p c p .

where ε denotes demand elasticity. This implies that the equilibrium reimbursement from managed care organizations to providers (p) equals to αc, whereα =1/(1−ε). Therefore, managed-care organizations’ costs of caring for the patients are proportional to hospital costs.

When dealing with a hospital with cost higher than the market average, managed-care organizations expect that this hospital is less profitable for them. However, if unobservable factors affect hospital costs and quality, managed-care organizations must form their expectation of profitability according to all the information they possess. For simplicity, I assume that the current expectation is also based on the expectation in the previous period, and the adjustment is assumed to be partial:

3 Denote k t h

N , as the total number of patients admitted to hospital h and enrolled in managed-care organization k in year t. Ntk is the total number of patients enrolled in managed-care organization k. The market share of hospital h in managed care organization k in year t is calculated as follows:

= = h k t h k t h k t k t h k t h N N N N s , , , ,

4 A managed care organization is less possible to direct patients away from a less capacity-constrained hospital in the short run

(7)

k t h k t h k t h k t h, = 0 + 1⋅dc ,−1 + 2dq ,−1 +Ε ,−1 Ε λ λ λ (2) where k t h

dc ,1 is the deviation of hospital costs from the market average ( k t h k t h k t h c c dc ,1 = ,1,1), and likewise k t h

dq ,1 is the deviation of hospital quality from the market average

( k t h k t h k t h q q dq ,1= ,1,1).5

3. Finding and Conclusion

This paper examines whether managed-care organizations can affect their enrollees’ choice of hospitals through a redirection mechanism. Our estimations have focused on Massachusetts’s hospitals that have treated women aged 15 to 45 for childbirth in fiscal years 1995 to 1999. We defined a hospital’s market by geographical distance to the hospital. Due to data limitation, only the HMO-type managed-care organizations have been included for our analysis, and only the six largest managed-care organizations have been included for estimations.

We have found that a managed-care organization direct its enrollees away from a hospital when the hospital’s costs are high and quality is low. This result is evidence for a managed-care organization’s ability to direct patients. Furthermore, we have also found that this redirection mechanism is stronger for larger hospitals in a provider network. A larger hospital in a network usually is a hospital favored by a managed-care organization. Therefore, this result has

strengthened the evidence for redirection. In addition, we have also found a stronger effect when we use a narrower market definition. This has provided additional evidence for redirection.

Furthermore, our robustness tests contain three variations. First, the distance to define the market has been relaxed to narrower and wider one. Second, managed-care organizations have been divided into two types, tightly and loosely controlled managed-care organizations. Lastly, we have estimated models based on a subsample of hospitals that (1) treat more than a certain percentage of enrollees of a managed-care organization; (2) are among top-five to top-ten hospital in a managed-care organization. These robustness tests have confirmed the findings in this study, and shown that the findings in the baseline estimation are robust.

5 This is a strong assumption. This assumes that a hospital’s performance will affect a managed-care organization’s expectation forever.

(8)

4. References

American Health Line (1998), “Health Care Spending Reaches $1 Trillion Mark for First Time,” January 13.

Burns, L., and D. Wholey (1992), “The Impact of Physician Characteristics in Conditional Choice Models for Hospital Care,” Journal of Health Economics, 11(1), pp. 43-62.

Culter, D. M. and Sheiner, L. (1997), “Managed Care and the Growth of Medical Expenditure,” NBER working paper: w6140.

Dor, A., S. M. Koroukian, and M. Grossman (2004), “Managed Care Discounting: Evidence from the Marketscan Database,” NBER working paper: w10437.

Glied, S. (2000), “Managed Care,” Chapter 13 of Handbook of Health Economics, edited by A. J. Culyer and J. P. Newhouse, 1st edition, Amsterdam ; New York : Elsevier

Health Insurance Association of America. (2002), Source Book of Health Insurance Data 2001-2001, HIAA press, Washington DC.

Jensen, G.A., M. A. Morrisey, S. Gaffney, and D. K. Liston. (1997), “The Dominance of Managed Care: Insurance Trends in the 1990s,” Health Affairs, 16(1), pp. 125-138.

Jensen, G. A., M. A. Morrisey, and J. Marcus (1987), “Cost Sharing and the Changing Pattern of Employer-sponsored Health Insurance,” Milbank Quarterly, 65(4), pp. 521-550.

Keeler E. and Brodie M. (1993) “Economic Incentives in the Choice between Vaginal Delivery and Cesarean Section,” The Milbank Quarterly, 71(3), pp. 365-404.

Kessler D. P. and M. B. McClellan (2000), “Is Hospital Competition Socially Wasteful?” Quarterly

Journal of Economics, 115(2), pp. 577-615.

Lien, H. M. (2002), “Does the Use of Network Reduce Quality?” working paper. Boston University. Luft, H. S. (1981), Health Maintenance Organizations: Dimensions of Performance, New York: John

Wiley and Sons.

Ma, C. A. and T. G. McGuire (2002), “Network Incentives in Managed Health Care,” Journal of

Economics and Management Strategy, 11(1), pp. 1-35.

McGuirk, M., and F. Porell (1984), “Spatial Patterns of Hospital Utilization: The Impact of Distance and Time,” Inquiry, 21(1), pp. 84-95.

Miller, R. H. and Luft, H. S. (1994), “Managed Care Plan Performance since 1980: A Literature Analysis,”

Journal of American Medical Association, 271, pp. 1512-1515.

Miller, R. H. and Luft, H. S. (1997), “Does Managed Care Lead to Better or Worse Quality of Care?”

Health Affairs, 16(5), pp. 7-25.

Morrisey, M. A. and G. A. Jensen (1997), “Switching to Managed Care in the Small Employer Market,”

Inquiry, 34, pp. 237-248.

Reed, M. C. (2000), “Why People Change Their Health Care Providers?” HSC Data Bulletin, 16.

Town, R. and Vistnes, G. (2001), “Hospital Competition in HMO markets,” Journal of Health Economics, 20-5, pp. 733-753.

Vistnes, G. (1995), “Hospital Mergers and Antitrust Enforcement,” Journal of Health Politics, Policy and

Law, 20(1), pp. 175-190.

White, W., and M. Morrisey (1998), “Are Patients Traveling Further?” International Journal of the

Economics of Business, 5(2), pp. 203-p222.

Zwanziger, J., and G. Melnick (1993), “Effect of Competition on the Hospital Industry: Evidence from California.” In Competitive Approaches to Health Care Reform, R. Arnould, R. Rich, and W. White, eds, Washington, D.C: Urban Institute Press.

(9)

5. Personal Evaluation

This paper aims to submit the health economic journals, like Journal of health economics, or Health economics.

參考文獻

相關文件

Probably the greatest challenge to those managing patients with oral diseases is the dilemma of attempting to predict which oral erythroplakias, leukoplakias, liche- noid and

In clinical trials involving patients with DDwoR, however, most authors usually define their samples based on the duration of symptoms (i.e. locking duration or time since

SUMMARY The purpose of this systematic review was to describe the prevalence of whiplash trauma in patients with temporomandibular disorders (TMDs) and to describe clinical signs

The percentage of positive LSGBs was significantly higher in patients in whom the biopsy was performed by or on the request of either the department of Rheumatology and Internal

patients with stage I/II disease but not in those with stage III disease.43 A high serum level of VEGF is associated with poor survival among patients with small cell lung

According to his protocol, patients are categorised as Stage I if they exhibit exposed bone in a field of radiation that has failed to heal for at least 6 months and do not have

The mean salivary level of alpha-amylase and the mean scores of state and trait anxiety in geographic tongue patients were higher than those of healthy controls.. But these

Seals, if any, essential for sealing the pressure sensing element, and in direct contact with the process medium, made of or protected by aluminum, aluminum alloy, aluminum