Certificate of need laws: a systematic review and cost-effectiveness analysis.

BACKGROUND: Certificate of Need (CON) laws, currently in place in 35 US states, require certain health care providers to obtain a certification of their economic necessity from a state board before opening or undertaking a major expansion. We conduct the first systematic review and cost-effectiveness analysis of these laws.
METHODS: We review 90 articles to summarize the evidence on how certificate of need laws affect regulatory costs, health expenditures, health outcomes, and access to care. We use the findings from the systematic review to conduct a cost-effectiveness analysis of CON.
RESULTS: The literature provides mixed results, on average finding that CON increases health expenditures and overall elderly mortality while reducing heart surgery mortality. Our cost-effectiveness analysis estimates that the costs of CON laws somewhat exceed their benefits, although our estimates are quite uncertain.
CONCLUSIONS: The literature has not yet reached a definitive conclusion on how CON laws affect health expenditures, outcomes, or access to care. While more and higher quality research is needed to reach confident conclusions, our cost-effectiveness analysis based on the existing literature shows that the expected costs of CON exceed its benefits.

Background

Certificate of need (CON) statutes were designed to hold down health costs by limiting unnecessary proliferation and duplication of health facilities, to improve quality by regionalization of selected types of surgical or other procedures where a volume-quality relationship exists and to improve access to care by preventing competitors from “cream-skimming” paying patients, leaving selected facilities with disproportionately high uncompensated care loads. The first CON statute was enacted in New York in 1964 and due to strong federal encouragement in the early 1970s, such programs were adopted by virtually all states by the early 1980s. Following the abandonment of federal support in the mid-1980s, 14 states fully repealed their CON laws, leaving 37 states and District of Columbia in which such programs were operational in 2008. Debate over CON has revived in recent years, with New Hampshire repealing its CON program in 2016 and Florida repealing most CON requirements in 2019.

New contribution

Research has examined the impact of various dimensions of CON programs, but we are unaware of any published systematic review of this literature or any net assessment of the benefits and costs of these statutes. To evaluate the existing research, we performed a systematic review to address two key questions with several related parts. The questions are as follows:

Key Question 1: What are the regulatory costs of CON programs, including a) federal and state government regulatory costs and b) industry compliance costs?

Key Question 2: What are the key impacts of CON programs on a) health expenditures; b) health outcomes; and c) access to care?

We performed a database literature search that identified 1035 articles (939 unduplicated). Of these, we excluded 885 that clearly did not meet our inclusion criteria after abstract review. Of the 54 remaining articles subjected to full review, 13 were rejected and 41 retained. In addition to the database search, we solicited articles from systematic web searches and experts in the field. These sources provided 54 additional items, of which 49 were used in the preparation of the net assessment, yielding a full total of 90 articles.

We summarize the findings of this literature and use its findings to estimate the costs and benefits of CON laws. Cost-benefit analysis is generally used to convert disparate societal costs (such as increased spending or travel time) and benefits (such as reduced mortality) of a treatment or policy into dollar terms. The benefits are added together and the costs subtracted in order to provide a single estimate (or range of estimates) for how the policy affects overall societal welfare (wellbeing). See Viscusi [1] for a more detailed explanation and a defense of pricing health outcomes.

We estimate that the average cost-benefit ratio of CON is 1.08, meaning costs exceed benefits by 8%, with the average costs exceeding benefits by an estimated $302 million per year. However, these estimates are quite uncertain given that the literature on how CON affects key outcomes such as mortality and spending is quite mixed.

History

The history of CON is well-documented elsewhere, but as a general proposition, state policy currently has been moving in the general direction of eliminating CON or softening its stringency since the 1980s [2]. Simpson [3] provides an exhaustive early history of CON; Conover and Sloan [2] provide an update of CON trends since the early 1980s.

CON regulation first began in New York in 1964. The National Health Planning and Resource Development Act of 1974 (P.L. 93–641) provided federal funds for state efforts to implement CON and proposed severe financial penalties for failure to do so. Due to repeated congressional postponement of effective dates, the latter provisions never were put into effect, but the prospect of these rules being implemented induce virtually every state to make concerted efforts to comply [3]. P.L. 93–641 was repealed in 1986. Over the subsequent decade, 20 states have elected to drop their CON programs for acute care services (6 of these retained CON for nursing homes or other long term care services).

There are no longer any federal regulations governing CON. The American Health Planning Association [4] puts out an annual report that codifies by state the key features of all current CON programs (e.g., the dollar value of review thresholds, scope of services subject to review), as well as Web links to all operational programs where further details on pertinent regulations may be found.

Key elements

CON programs generally establish dollar thresholds for review of proposed projects related to new building or expansion of health services. Some states set different dollar thresholds (generally lower) for long term care facilities than for other acute care facilities such as hospitals. In addition, the thresholds for equipment generally are lower for equipment such as lithotripters than for capital projects; in the most stringent states, all projects involving equipment of a particular type are reviewed regardless of the size of the project. Likewise, states sometimes draw a distinction between new services as opposed to expansion of existing services. All in all, there is wide variation in the scope and mechanics of CON review across states (see AHPA 2009 [5] for a detailed breakdown).

As of 2008, there were 31 states (including District of Columbia) with CON for both hospitals (and other acute care services) and nursing homes, while another 6 states have retained CON for nursing homes only [5].

Historically, CON was enforced by the threat of disallowing Medicare or Medicaid payment for facilities or services that had not undergone CON approval. Since repeal of these provisions, states have adopted a variety of mechanisms for enforcing CON regulations. States generally can enforce CON requirements by denying, suspending or revoking the license or certification of a facility not in compliance, but in addition, some states impose sizable administrative penalties (e.g., $5 million) for specific violations of CON statutes.

Methods

We investigated two broad research areas related to the impact of CON regulation in the U.S. The questions are listed below, along with a brief description of our analytical approach.

Regulatory costs of CON regulation

Key Question 1a. What is the amount of government regulatory costs related to the CON regulation? This includes state costs to monitor and enforce rules related to certificate of need for hospitals, nursing homes or other facilities to which CON is applicable.

Key Question 1b. What is the amount of industry compliance costs related to CON regulation? This includes all administrative costs and enforcement penalties borne by private, state or locally owned health facilities subject to state CON rules. Monetary penalties may be viewed as a transfer, but the remaining costs represent real resource losses to society.

Major impacts of CON regulation

Key Question 2a. What is the net impact of CON regulation on health expenditures? Historically, CON policy was justified on market-perfecting grounds to overcome the weak incentives for economic discipline resulting from a combination of cost-based reimbursement and pervasive third-party payment for health care. According to this theory, CON could enhance efficiency by regionalizing expensive tertiary facilities and preventing the costly duplication of technology or facilities. Skeptics argue that CON is a form of industry protection from competition, pointing out that states that first adopted CON were more likely to have more hospital beds and lower occupancy rates. Reduced competition could have adverse effects on health expenditures (by allowing facilities to charge higher prices). Therefore, our search allowed for the possibility that CON could decrease, increase or have no impact on health expenditures.

Key Question 2b. What is the impact of CON regulation on health outcomes? To the extent that facilities with higher volumes of selected procedures have better outcomes, [6] regionalization resulting from CON could have a corollary benefit in the form of improved patient outcomes. Likewise, to the extent that CON efforts to prevent “cream-skimming” were successful, this might allow the survival of certain facilities such as large urban public hospitals that might otherwise be forced to shut down for lack of sufficient paying patients. In theory, this too could result in health benefits and/or reductions in avoidable hospitalizations if indigent patients were able to receive essential care on a timely basis. But limitations on competition also have the potential to result in lower quality care, so we sought literature that related CON to outcomes in either direction. Changes in either morbidity or mortality could be monetized using conventional methods.

Key Question 2c. What is the impact of CON regulation on access to care? Another rationale for CON is to ensure access to disadvantaged populations or to maintain provider financial margins to allow them to cross-subsidize indigent care, i.e., help offset uncompensated care costs [7]. Whether informally or formally through explicit commitments required for approval, CON regulators have the power to restrict approval to facilities willing to supply services perceived to be in the public interest, such as charity care or care in medically underserved areas [8]. Even if they had no measurable impact on health outcomes, such improvements in access to care would be of value, so we sought to ensure to include literature focused on this dimension of CON performance. However, some have raised concerns that CON may restrict access to care through output restriction and market division. That is, the CON process may allow regulated hospitals or facilities to “carve out” the distribution of patients, beds or suites (both geographically or by specialty niche). “Output restriction and market division are two classical tactics that are used by economic oligopolies to manage supply and create market power”. ([8] , p.1085) Such practices may be tacitly encouraged through CON programs. Therefore, our search allowed for the possibility that CON could decrease, increase or have no impact on access to health services.

Literature search and review

We searched the MEDLINE, CINAHL, Lexis-Nexis, and Public Affairs Information Service (PAIS), ISI Web of Knowledge, ProQuest Dissertations & Theses, and Health Affairs databases through 2010. A professional librarian conducted each search, customizing the searches for each research question. We present our full search strategy for MEDLINE in Appendix B as an example of our process. In general, our search strategy is similar to that of Conover and Wiechers [9].

In addition to our searches of formal databases, we searched the websites of health policy firms, research organizations, and foundations for relevant articles; the only relevant articles from this search came from the American Health Planning Association and the Employee Benefits Research Institute.

Study selection

We excluded studies with no original data, no outcomes of interest, studies superseded by an updated version, studies published in abstract form only, and studies using only pre-1975 data.

Further detail

A detailed evidence table documenting all the methods and sources used in the net assessment appears in Appendix A. Appendix B codifies all the search terms used for each database examined. We did not use a registered protocol.

Results

We report the results in two main sections here and then turn to a net assessment assembled using all the available evidence. A summary table and graph are contained at the end of this section.

Results of literature search

The literature search identified 1035 articles (939 unduplicated). Of these, we excluded 885 that clearly did not meet our inclusion criteria after abstract review. Of the 54 remaining articles subjected to full review, 13 were rejected and 41 retained. In addition to the database search, we solicited articles from systematic web searches and experts in the field. These sources provided 54 additional items, of which 49 were used in the preparation of the net assessment. One investigator extracted information from each article into evidence tables. Another investigator independently assessed the accuracy and completeness of the literature review and the net assessment developed using evidence from the synthesis.

Literature quality

In contrast to the dearth of studies available on other aspects of health regulation, there is a comparative embarrassment of riches in the area of CON, as it was the first major state-level initiative to curtail health spending. Kessler and McClellan [10] show that in the 1980s, the welfare effects of hospital competition were ambiguous, but in the 1990s hospital competition unambiguously improved social welfare, i.e., lowering average expenditures per patient and mortality among Medicare patients being treated for heart attacks in the 1990s. Thus, in cases where evidence is conflicting, more recent evidence about CON’s effects has been given greater weight than earlier studies.

Another important consideration in evaluating the quality of a study’s results is whether it addresses the issue of endogeneity or omitted variables bias. Researchers generally try to control for known differences between the states that might have an influence on an outcome variable of interest. But they can never be certain they have taken into account every unobserved difference that might matter. We put greater weight on studies that use methods to address these problems, such as state fixed effects, difference-in-difference analysis, or two-stage least squares.

Key question 1: regulatory costs of CON regulation

Public administration

KQ 1a concerned the effects of CON regulation on the costs of public administration. We found no nationwide estimate of such costs; however, total staffing for CON agencies by state are reported in a 1986 DHHS report [11].

Compliance costs

KQ 1b concerned the costs of compliance with the Act. While we found sporadic allusions to costs incurred by CON applicants and/or delays imposed by the CON process, we found no systematic literature summarizing such costs or providing reliable parameters from which to construct a national estimate.

Key question 2: major impacts of CON regulation

Impact on health expenditures

KQ 2a concerned the Act’s effects on health spending. Because of the volume of available literature, we have divided it based on the category of facilities to which CON applies (hospitals, nursing homes and other) and discuss individual papers in detail only in Appendix C.

Hospital CON

The evidence regarding hospital CON’s effect on health expenditures is generally mixed, although one could credibly conclude that the weight of this evidence is that CON has no impact on health costs overall. Because of the sheer number of studies available, including several of high quality that examine CON’s impact on overall health expenditures, we have excluded many other narrowly focused studies that examined CON’s effects on cost per day or cost per stay since one cannot reliably extrapolate from these results to a global cost effect (see Conover and Sloan [2] for a review). Some more recent studies [2, 12]. have found that CON produces a savings in the hospital sector while having no detectable impact on health expenditures overall. In addition, most early studies did not correct for the endogeneity of CON regulation. More recent studies have corrected this oversight by using either two-stage models to estimate demand for regulation (e.g., [13]) or using state-level fixed effects estimators, (e.g. [2, 12, 14]).

Nursing home CON

We found at least a half dozen studies that examined the effects of nursing home CON on bed capacity, but most of these are at least two decades old and not directly informative of the net effect on health spending (since even demonstrated reductions in nursing home bed supply may simply spill over into expenditures on home and community-based substitutes). However, two more recent studies explicitly examined the impact of nursing home CON on health expenditures, with mixed results. We discuss these in detail in Appendix C.

Other types of CON

The literature on how other types of CON affect spending is relatively thin. Three papers, discussed in Appendix C in detail, provide suggestive evidence that hospice CON, home health CON, and dialysis CON increase spending, but these papers generally do not use direct measures of spending or costs.

Impact on health outcomes

KQ 2b concerned the effects of CON regulation on health outcomes. The Federal Trade Commission and U.S. Department of Justice, [15] after extensively reviewing the available literature and hearing from expert witnesses, concluded the following: “The Agencies believe that CON programs are generally not successful in containing health care costs and that they can pose anticompetitive risks. As noted above, CON programs risk entrenching oligopolists and eroding consumer welfare. The aim of controlling costs is laudable, but there appear to be other, more effective means of achieving this goal that do not pose anticompetitive risks A similar analysis applies to the use of CON programs to enhance health care quality and access” (emphasis added; ch. 8, p. 5). There are three bodies of CON literature that address health outcomes: the largest group examines CON’s impact on mortality, a much smaller body examines other health outcomes and the third focuses on nursing home quality.

Mortality losses

The CON literature related to mortality losses has been the most disparate, with three studies finding that CON reduces mortality, five showing no effect, and three finding that CON increases mortality. We discuss these studies in detail in Appendix C. Oddly, almost the entire literature on CON and mortality has focused on heart surgery mortality; only Shortell and Hughes [16] consider other procedures, finding that mortality is higher in states with stringent CON regulation.

Other health outcomes

A handful of studies have examined health outcomes unrelated to mortality risk, suggesting that CON worsens the quality of dialysis care [17] and has mixed effects on the quality of cardiac care [18, 19]. We discuss these studies in more detail in Appendix C.

Nursing home quality

There is a companion literature on the impact of nursing home CON on quality measured in terms of structure, process and outcome indicators.

In theory, a binding CON policy provides no incentive to nursing homes to compete for Medicaid residents on the basis of quality, the theory argues that under such a binding constraint a higher payment level actually leads to lower quality. It also can lead to access problems for Medicaid patients such as “cream-skimming” patients likely to require less intensive nursing home services. A large body of literature--excellently reviewed in Caldwell [19]--examines evidence related to this theory (and to “cream-skimming” and other access barriers faced by prospective nursing home patients on Medicaid), but recent evidence in a series of studies by David Grabowski [20-24] suggests higher Medicaid payments do result in higher nursing home quality, contradicting a raft of earlier results by Nyman [J Health Econ. 1985 ">25-28]. Nursing home occupancy rates have declined over time, suggesting less “excess demand.” This in turn suggests that either nursing home CON may impose less of a binding constraint, or that substitutes for nursing home care have arisen that dissipate some of the adverse consequences that nursing home CON might historically have imposed.

More directly though, Zinn [29] showed that the presence of a statewide nursing home construction moratorium is associated with a lower level of quality of care, measured in terms of both lower RN staffing and a higher percentage of residents who are physically restrained.

Impact on access to care

KQ 2c concerned the effects of CON regulation on access to health care. As of 1994, most CON programs required facilities to provide a “reasonable amount” of care to the poor [30]. The literature on the actual effects of CON regulation on access have been mixed. Nine studies found a positive effect of CON on access, two found no effect, and sixteen found a negative effect. We discuss these studies in more detail in Appendix C. The greatest challenge in summarizing the literature on access to care is that almost every study defines “access” in a different way (amount of care overall, amount of uncompensated care, travel time to care, racial disparities in care, et cetera).

Cost-effectiveness analysis

Overview

During this systematic review, we identified a large body of literature addressing the regulatory costs and various impacts of CON regulation on health expenditures, health outcomes and access to care. Our review systematically identified, organized, and critically analyzed the relevant studies. While we located several in-depth qualitative reviews of various components of this literature, [12, 31, 32] none has been as thorough and systematic as this one. More importantly, no prior studies have reported an aggregate “bottom line” assessment of the nationwide benefits and costs of CON regulation. Consequently, in what follows, we have synthesized the best available evidence in order to provide such a net assessment. The wide range of evidence contained in the research surrounding certificate of need policies provides mixed results, showing that CON may either save or cost both money and lives.

How benefits and costs were calculated

We have calculated the regulatory costs in the following fashion (lower- and upper-bound estimates are shown in parentheses: full details of methods and sources are in Table A1).

Government regulatory costs

We multiply average CON staffing per state in 1986 [11] times the number of states with CON in 2008 and multiply this by the average total compensation of state employees in 2008. ([33], p. Table 843) Since we do not know whether CON employees have total compensation that is lower or higher than the average for all state employees, we use +/− 25% for upper and lower bounds.

Industry compliance cost

Lacking a firm national estimate, we calculated CON compliance spending per $10,000 personal health care expenditures (PHCE) in 2013 using data from the State of Washington [33, 34] and adjusting this figure to 2008 dollars using the change in U.S. PHCE between 2008 and 2013 [35]. The Lewin Group ([36], p.12) characterizes the Washington program as representing a “middle of the road” approach to CON, so no further adjustment to this figure appeared necessary. We applied this estimate to the total amount of personal health care spending in states with CON [5, 37, 38].

Key Impacts: Health Expenditures. We estimated three separate health spending effects.

Reduction in Coronary Artery Bypass Grafting (CABG) Facilities. We calculated the number of new CABG programs averted per 1000 CABG patients using Kolstad’s [39] estimates from Pennsylvania. We applied this to the estimated number of CABG surgeries in states with CON (allocating total CABG surgeries in 2008 [35] based on the distribution of total adults age 18 and older [40]. We multiplied the result times the amortized annual capital cost per CABG cost; this was derived from total capital costs reported in the literature, [41, 42] adjusted for inflation using the Turner Building Cost Index, [43] using the same 15-year amortization period as Cutler et al. (2010), and the prevailing bank prime loan rate (5%) as of July 2008 [44].

Medicare Spending (Stringent CON States). We estimated the reduction in Medicare spending in stringent CON states using Sloan and Conover’s [45] estimate of 1.8%, applying this to estimated Medicare spending in states with stringent CON in 2008 [37].

All Other Health Spending. In light of all the varied findings regarding CON on total expenditures, no reduction in overall spending seems most consistent with the available evidence. The 13.6% reduction in spending reported in Lanning, Morrisey and Ohsfeldt [13] is both very dated and inconsistent with subsequent studies by Sloan and Conover [45]--using an updated version of the same data series and similar methods for controlling for endogeneity--showing no statistically significant effect of CON (or even stringent CON) on overall health spending. But the same study shows there are savings to Medicare in stringent CON states. Moreover, the Sloan and Conover analysis was performed using data through 1998; hence, it hypothetically should have picked up any beneficial effects of CON on CABG spending. The only way to reconcile these results is to assume that Medicare and CABG savings are offset by corresponding increases in health spending elsewhere, resulting in zero overall net measured impact. Based on estimating aggregate personal health spending in CON states using CMS estimates of per capita spending by location of provider [37]. This implies an expected increase in all other health spending in CON states of 0.19%.

Home Health Cost Increases. We elected not to use the Anderson and Kass [46] finding of a slight increase in Medicare-certified home health agency costs due to CON both because the finding is so dated and has so many methodological limitations.

Key Impacts: Access to Care, Patient Time Losses. We used Kolstad’s estimate of $7.50 per CABG patient in time losses related to restricted supply of CABG facilities under CON (no further adjustment was made since it is based on Pennsylvania’s hourly wage, which is nearly identical to the national average [47]. This was multiplied by total CABGs in states with CON in 2008.

Key Impacts: Health Losses. We calculated four separate effects on mortality.

General Elderly Hospital Mortality Rates, Stringent CON States. Because it measured effects using data now nearly 2 decades old and also has its own methodological limitations, the Shortell and Hughes [16] finding that mortality was 6% higher than expected in states with stringent CON was assigned a weight of 25% (consistent with how other mortality estimates of that vintage were treated) and used as the 99th percentile cost estimate. This increase is applied to all elderly hospitalization deaths for the elderly in states with stringent CON (estimated by applying a national hospital death rate for the elderly [48] to total reported elderly hospitalizations by state [49]. This estimated mortality increase is monetized using an adjusted value of statistical life for the elderly in 2008 [50] that takes into account their lower life expectancy relative to the general population (i.e., this calculation assumes the elderly place the same value on an added year of life as the general population, but this means the aggregate value they place on mortality risk reductions on a per life saved basis is lower since they have fewer years of life remaining).

CABG Mortality, All Patients. The expected number of deaths averted due to CON is calculated using the estimated total CABGs in states with CON, 2008 times the weighted average reduction of 1.1 deaths per 1000 patients due to CON (detailed in Table A2 using results from Ho, [51] DiSesa et al., [52] and Robinson et al. [42] This estimated mortality reduction is monetized by first multiplying the estimated added life expectancy per CABG patient (8 years [53]) times the average quality of life of a typical CABG survivor (0.9 on a scale of death [0] to perfect health [1 [54];]) to obtain the net increase in quality-adjusted life years (QALYs) attributable to CON. The monetary value of these mortality gains is calculated by multiplying the number of QALYs times a willingness-to-pay estimate of the value of a QALY ($239,000 [50]).

CABG Mortality, Medicare Patients (Stringent CON States). The expected number of deaths averted due to stringent CON is calculated using the estimated total CABGs in states with stringent CON, 2008 times the weighted average reduction of 7.3 deaths per 1000 elderly patients due to stringent CON (detailed in Table A2 using Popescu et al. [55]). Each death averted is monetized as described immediately above except that the added life expectancy per CABG patient is adjusted to reflect the shorter life expectancy for elderly CABG patients [56].

CABG Mortality, Medicare Patients (Non-Stringent CON States). The expected number of deaths attributable to non-stringent CON is calculated in the identical fashion by authors using the weighted average increase of 4.9 deaths per 1000 elderly patients (detailed in Tables A1 and A2 using Vaughn-Sarrazin et al. [57] Popescu et al. [58], DiSesa et al. [52] Popescu et al. [55] and Ho et al. [59].

Social Welfare Losses: Efficiency Losses from Tax Collections. We accounted for the marginal overhead costs—which account for government collections costs, taxpayer compliance costs, and marginal excess burden (deadweight losses)—of state taxes by multiplying state administration costs times 40.5% (Duke University Center for Health Policy and Inequalities Research (CHPIR, 2015a). By the same logic, every dollar of Medicare savings is associated with companion “hidden” savings to society from the marginal overhead costs of federal taxes. Hence, all Medicare savings have been multiplied by 48.2% to account for such gains.

Social Welfare Losses: Efficiency Losses from Regulatory Costs. All industry compliance costs are presumed to be roughly equivalent to a sales tax, i.e., raising prices and reducing demand/output correspondingly. We therefore multiply the sum of industry compliance costs and health expenditure changes times the marginal excess burden associated with sales taxes (23.3%) [60].

Summary findings

The results are summarized in Table 1. The expected benefits of CON regulation in 2008 ($3.6 billion) are exceeded by its costs ($3.9 billion) by $302 million. Put another way, its benefits are 8% less than its costs (conversely, its expected costs exceed expected benefits by 8%). There is considerable uncertainty surrounding all these figures.

Table 1

Benefits and Costs of Certificate of Need (millions of 2008 dollars)

		DISTRIBUTION
	EXPECTED	Lower bound(5%)	25%	75%	Upper bound(95%)
Benefits	$ 3647	$ 1658	$ 2791	$ 4471	$ 5821
Costs	3949	932	1965	5185	9380
Net Costs	302	(3556)	(1754)	1768	5987
Benefit: Cost Ratio	0.92	0.30	0.65	1.86	4.08
Cost-Benefit Ratio	1.08	0.24	0.54	1.57	3.34

Probability benefits exceed costs: 54%

Consequently, as shown in Fig. 1, the weight of the evidence suggests that CON creates more costs than benefits. We estimate that the probability that benefits exceed costs is 54%. However, because net costs are skewed in the direction of reducing rather than increasing social welfare our best estimate is that social welfare would increase by several hundred million dollars a year if CON were repealed in the 37 states that retain it, or if it were modified in some fashion to either increase benefits, reduce the costs of achieving them or some combination. It is beyond the scope of this synthesis to make recommendations about how the latter might be achieved.

Fig. 1

Net Costs of Certificate of need, 2008

Limitations

The information compiled in this report may permit policymakers to identify areas in which regulatory costs appears excessive relative to benefits. However, this report is not designed to provide specific guidance on ways in which the objectives of CON restrictions might be pursued more cost-effectively.

Our cost-effectiveness analysis is limited by the quality and scope of the literature upon which it depends. In particular, the literature on CON and mortality focused almost entirely on heart surgery mortality, and so this provides most of the evidence weight for our mortality estimates even though it represents only a small fraction of all mortality. Our estimates of the effect of CON would likely be larger if more broad mortality measures were available (either larger costs or larger benefits depending on whether this literature found CON to increase or reduce mortality). Going forward, we hope researchers will study the effect of CON on types of mortality other than just post-heart-surgery. The literature on CON and access to care, by contrast, is too varied, with almost every study measuring “access to care” in a different way. We used only patient travel time in our cost-effectiveness analysis, as the measure that was easiest to translate into dollar terms, but it is certainly not the only measure of access and arguably not the best. The effect of CON on travel time was a small cost, but other access measures may have yielded either larger costs or benefits. Finally, the literature on the regulatory and compliance costs of CON was sparse and out-of-date.

In addition to the limitations of this overall body of literature and the particular challenges it poses, our review process also had some limitations. Because of time and resource constraints, we did not conduct dual, independent, blinded review of articles for inclusion or abstraction of information into evidence tables. We also did not rely on a formal scoring process for grading the quality of individual articles. Our review also stopped in 2010, leaving room for future work to focus on the most recent literature. Newer literature may find systematically different results, not only due to new econometric techniques or areas of focus, but because the true underlying effect of CON may change over time (in particular given insurance-market reforms; see Bailey [61]).

However, we have labored to be as transparent as possible in demonstrating to readers the methods, sources and analytic assumptions used to develop the net assessment. This approach provides an exceptionally high degree of granularity, allowing skeptical readers to substitute their own assumptions to calculate an alternate result. As well, we have made a point of codifying the uncertainty inherent in all of our estimates rather than relying simply on point estimates. While there may be components of benefits or costs that we have missed or inaccurately measured, we have no reason to believe that the process we used would have systematically biased our estimate of net costs in any obvious direction.

Conclusion

Based on the available evidence, CON programs appear to have achieved some benefits. However, the costs imposed such programs, including regulatory costs as well as adverse effects on health spending, exceed those benefits by an estimated $302 million a year. On average, the cost-benefit ratio is 1.08, meaning costs exceed benefits by 8%. Consequently, the weight of the evidence suggests that CON creates more costs than benefits. However, our estimates are quite uncertain; we have 90% confidence the true value of the cost-benefit ratio lies between 0.30 and 4.08.

We estimate that the probability that CON’s benefits exceed its costs is 54%. However, because net costs are skewed in the direction of reducing rather than increasing social welfare, our best estimate is that CON decreases social welfare by several hundred million dollars per year. These effects are largely driven by health expenditures and health status; the literature yields central estimates that CON has no effect on expenditures and slightly improves health, but suggests that CON is more likely to lead to large negative effects (more spending and worse health) than large positive ones (less spending and better health). Therefore, expected social welfare would increase if the 35 states that continue to maintain CON programs repealed them or modified them in some fashion to either increase benefits, reduce the costs of achieving them, or some combination.

Supplementary information

Additional file 1.

Table 2

Costs and Benefits of Certificate of Need Regulation (millions of 2008 dollars)

COST CATEGORY	COSTS			BENEFITS			NOTES
	Expected	Lower bound	Upper bound	Expected	Lower bound	Upper bound
Government Regulatory Costs	20	9	31	–	–	–
Federal	–	–	–	–	–	–
State	20	9	31	–	–	–	[A]
Compliance Costs	146	109	182	–	–	–
Administrative costs	146	109	182	–	–	–	[B]
Key Impacts	3272	677	7836	2786	1291	4418
Health expenditures	2023	104	6044	2023	848	3229
Reduction in CABG facilities	–	–	–	456	233	711	[C]
Medicare spending	–	–	–	1566	416	2748	[D]
All other health spending	2023	104	6044	–			[E]
Patient time losses	2	1	3				[F]
Health status	1247	314	3746	764	(1)	1974
Morbidity losses	–	–	–	–	–	–
Mortality losses	1247	132	3541	764	(1)	1974
General elderly hospital mortality	1027	44	3303	–			[G]
CABG mortality, all patients				463	(199)	1455	[H]
CABG mortality, Medicare patients, stringent CON states				300	(4)	807	[I]
CABG mortality rates, Medicare patients non-stringent CON states	220	(97)	701				[J]
Social Welfare Losses	512	66	1453	861	295	1511
Efficiency losses from tax collection	8	4	13	755	192	1400	[K]
Efficiency losses from regulatory costs	504	58	1446	106	54	168	[L]
GRAND TOTAL	3949	932	9380	3647	1658	5821
NET BURDEN (COSTS MINUS BENEFITS)	302	(3556)	5987	Probability that benefits exceed costs:			54%

Update: 5/16/2015

Note: All lower and upper bounds are calculated using Latin Hypercube simulations in @RISK. Therefore, sub-totals generally are smaller than the sum of their components. All reported simulation results are based on 100,000 iterations. For simplicity, both costs and benefits are shown as positive numbers. However, costs are equivalent to negative benefits and vice-versa. Cases in which negative numbers appear under either costs or benefits illustrate instances in which an impact is not statistically significant using conventional standards: that is we are not 95% certain that the estimated impact does not fall above or below zero.

Parameters:		DISTRIBUTION	@RISK FORMULA	EXPECTED	LOWER BOUND (5%)	UPPER BOUND (5%)	NOTES
Government Regulatory Costs
[P1]	Total states with CON, 2008	Constant	37	37	37	37	[a]
[P2]	Average FTE CON employees in states with CON	Truncated Normal	10.2	10.2	5.1	15.3	[b]
[P3]	Annual total compensation, state employees, 2008	Normal	52,058	52,058	39,044	65,073	[c]
Compliance Costs
[P4]	Average spending per CON application, WA, 2013	Constant	124,936	124,936	124,936	124,936	[d]
[P5]	Average annual number of CON applications, WA	Constant	52	52	52	52	[e]
[P6]	Personal health care expenditures (PHCE), WA, 2008 (millions)	Constant	$ 42,411	$ 42,411	$ 42,411	$ 42,411	[f]
[P7]	Percentage change in PHCE, U.S., 2008–2013	Constant	17.9%	17.9%	17.9%	17.9%	[g]
[P8]	CON compliance spending per $10,000 PHCE	Normal	$ 1.29	$ 1.29	$ 0.97	$ 1.61	[h]
[P9]	Personal health care expenditures (PHCE), 2008 (000’s)	Normal	$ 1,997,199	$ 1,997,199	$ 1,964,445	$ 2,029,953	[i]
[P10]	Share of PHCE in states with CON, 2008	Constant	56.6%	56.6%	56.6%	56.6%	[j]
[P11]	PHCE spending in CON states, 2008 (millions)	Calculated	$ 1,130,663	$ 1,130,663	$ 1,112,100	$ 1,149,200	[k]
Key Impacts
Health expenditures
[P12]	Medicare expenditures, 2008 (millions)	Normal	$ 440,758	440,758	$ 433,530	$ 447,987	[l]
[P13]	Share of Medicare spending in states with stringent CON, 2008	Constant	19.2%	19.2%	19.2%	19.2%	[m]
[P14]	Reduction in Medicare/eligible spending due to CON	Normal	1.8%	1.8%	0.5%	3.2%	[n]
[P15]	All other PHCE spending in CON states, 2008 (millions)	Calculated	$ 1,045,988	$ 1,045,988	$ 1027,402	$ 1,064,578	[o]
[P16]	Increase in per capita health spending due to CON	Exponential	0.19%	0.14%	0.01%	0.58%	[p]
[P17]	New CABG programs averted per 1000 CABG patients	Normal	0.97	0.97	0.75	1.19	[q]
[P18]	Capital cost per CABG program (millions)	Normal	$ 20.5	$ 20.5	$ 17.8	$ 23.2	[r]
[P19]	Amortized annual capital cost per CABG program (millions)	Normal	$ 2.0	$ 2.0	$ 1.7	$ 2.2	[s]
Patient time losses
[P20]	Time losses per CABG patient due to restricted supply	Normal	$ 7.50	$ 7.50	$ 3.75	$ 11.25	[t]
[P21]	Total CABGs in states with CON, 2008	Normal	238,586	238,586	131,404	267,020	[u]
General elderly hospital mortality
[P22]	Elderly hospital patients, 2008 (thousands)	Normal	13,902	13,902	11,504	16,300	[v]
[P23]	Elderly in-hospital mortality risk, 2008	Normal	4.1%	4.1%	4.0%	4.2%	[w]
[P24]	Increase in Medicare mortality rates due to stringent CON	Exponential	0.3%	7.9%	6.1%	9.7%	[x]
[P25]	VOSL, elderly, 2008 (millions of dollars)	Triangle	$ 2.9	$ 2.9	$ 0.6	$ 5.9	[y]
CABG mortality, all patients
[P26]	Change in CABG mortality rate/1000 patients due to CON	Calculated	−1.1	−1.1	−2.8	0.6	[z]
[P27]	Added life expectancy per CABG patient	Normal	8.0	8.0	6.0	10.0	[aa]
[P28]	Quality of life, CABG patients after 10 years	Normal	0.9	0.9	0.7	1.1	[ab]
[P29]	Value of a quality-adjusted life year (QALY), 2008 (thousands)	Triangle	$ 239.1	$ 239.1	$ 120.7	$ 407.0	[ac]
CABG mortality, Medicare patients, stringent CON states
[P30]	Elderly CABGs in states with stringent CON, 2008	Normal	43,840	43,840	24,145	49,065	[ad]
[P31]	Change in elderly CABG mortality rate/1000 due to stringent CON	Calculated	−7.3	−7.3	−14.76	0.13	[ae]
[P32]	Ratio of elderly CABG life expectancy to average CABG LE	Constant	0.54	0.54	0.54	0.54	[af]
CABG mortality rates, Medicare patients non-stringent CON states
[P33]	Elderly CABGs in states with CON, 2008	Normal	91,540	91,540	50,417	102,450	[ag]
[P34]	Change in elderly CABG mortality rate/1000 in all CON states	Calculated	−0.9	−0.9	−2.0	0.1	[ah]
[P35]	Change in elderly CABG mortality rate/1000, non-stringent CON	Calculated	4.9	4.9	−5.8	32.9	[ai]
Social Welfare Losses
[P36]	Marginal tax overhead costs, federal taxes	Normal	48.2%	48.2%	38.1%	64.3%	[aj]
[P37]	Marginal tax overhead costs, state taxes	Normal	40.5%	48.2%	32.4%	53.9%	[ak]
[P38]	Marginal excess burden, sales taxes	Normal	23.3%	48.2%	20.4%	26.1%	[al]

Table 3

Summary of Studies on the Effect of CON on Mortality for Coronary Artery Bypass Graft (CABG) Procedures

STUDY	DESIGN	STATES	YEARS	POP-ULATION	CONTROLS	ESTIMATED EFFECT OF CON ON MORTALITY						NOTES
						Mortality measure	Mean change1	Stand-ard error	Net change, deaths per 1000 patients2	Pop-ulation weight3	Recency weight4
Total Population Studies
Ho [65]	Retro-spective cohort	18 CON vs. 8 no CON	1988–2000	All	State fixed effects, patient characteristics	Inpatient mortality	−2.6%	0.8%	−1.00	77.1%	50.0%	Only one state in sample dropped CON during the study period. Finds no mortality effect of CON on PTCA. SE imputed from reported t statistic.
DiSesa et al. [52]	Retro-spective cohort	27 CON vs. 24 no CON	2000–2003	All	State and hospital fixed effects, patient controls	Operative mortality	−4.9%	5.7%	−1.25	100.0%	100.0%
Robinson et al. [42]	Pre-post	PA	1994–1999	All	Patient characteristics	Inpatient mortality	0.0%	0.9%	0.00	0.0%	50.0%	After CON lifted, actual mortality matched expected mortality for both old and new cardiac programs; uses same PHC4 data as Kolstad
Kolstad [39]	Pre-post	PA	1994–2003	All	Compares incumbent hospitals to new entrants		2.9%	0.9%	0.62	4.3%	100.0%	Kolstad calculates that 11 deaths are averted annually by CON repeal. His Table 1.1 shows that in 2000–2003, the 40 incumbent hospitals performed an average of 349 CABGs, for a total of 13,960 (RAMR = 2.17%) while 24 new entrants performed an average of 160 (RAMR = 2.04%). Weighted average mortality = 2.14% vs. 2.20% if 11 additional deaths had occurred.
Cutler et al. [53]	Pre-post	PA	1994–2003	All	Compares incumbent hospitals to new entrants		2.9%	0.9%	0.62	0.0%	100.0%	Essentially the same paper as Kolstad [39]
Weighted average:									−1.13
Medicare Patient Studies
Vaughan-Sarrazin et al. [57]	Retro-spective cohort	27 continuous CON vs. 18 no CON	1994–1999	Medicare (excludes managed care)	Patient characteristics	In-hospital mortality	−17.3%	2.6%	− 8.70	77.1%	25.0%	States without CON exhibited CABG higher mortality (OR = 1.22) than states with continuous CON; this implies CON is associated with an 17.3% decrease in mortality rates, derived algebraically. No effect in intermittent CON states
Popescu et al. [58]	Retro-spective cohort	27 CON vs. 23 no CON	1998–2000	Medicare	Patient characteristics	30-day all-cause mortality	−5.0%	1.0%	−8.90	100.0%	50.0%
DiSesa et al. [52]	Retro-spective cohort	27 CON vs. 24 no CON	2001	Medicare patients age 65 and older (excludes managed care)	State and hospital fixed effects, patient controls	Operative mortality	−0.3%	4.9%	−0.10	100.0%	75.0%
Popescu, Vaughan-Sarazin and Rosenthal [55]	Retro-spective cohort	27 CON vs. 24 no CON	2000–2003	Medicare (age 68+)	Patient characteristics	30-day all-cause mortality	0.0%	1.5%	0.00	100.0%	100.0%	Vaughan-Sarazin is co-author on this paper; her most recent work, using the most recent data she uses, finds zero effect (a true 0.00 estimate; not just statistically insignificant)
Ho et al. [59]	Retro-spective cohort	27 continuous CON vs. 7 dropped CON	1989–2002	Medicare patients age 65 and older (excludes managed care)	State fixed effects, extensive hospital and patient controls	Procedural mortality	10.8%	3.3%	5.20	63.9%	100.0%	Dropping CON reduces mortality at first, but the effect dissipates 5 years after CON is removed
Weighted average:									−0.93
Popescu, Vaughan-Sarazin and Rosenthal [55]	Retro-spective cohort	27 CON vs. 24 no CON	2000–2003	Medicare (age 68+)	Patient characteristics	30-day all-cause mortality	−4.2%	2.6%	−7.31	100.0%	100.0%	States with stringent CON lower mortality but effect is of borderline statistical significance

1Mean change in probability of death, calculated by authors using data reported at original source

2All figures calculated by authors: 1000 x (Mean Mortality Rate in CON States) x (1–1/(1 + Mean Change)) using data on the mean mortality rate for the relevant population and mortality measure shown as reported at original source

3Population weights represent the fraction of the theoretical population of interest included in a study. All figures are calculated by authors based on the total number of CABG surgeries in 2008, allocated to states based on 2008 Census figures on total adult population age 18 and older (for total population studies) and total population age 65 and older (for Medicare patient studies). A weight of zero has been assigned to studies that either duplicate other reported findings or have been entirely superseded by analyses using the same (overlapping) data source but with additional newer years of data

4Recency weights are calculated to provide greater weight to results that rely on more recent data and/or improved methods

Table 4

Summary of Studies on the Effect of CON on Quality for Procedures Other than Cardiac Surgery

STUDY	DESIGN	STATES	YEARS	CONTROLS	KEY FINDINGS	NOTES
Zinn [29]	Cross-sectional	50	1987	Facility characteristics	Moratorium on building makes nursing home patients 7% more likely to be physically restrained, reduced RNs per patient 32%	Physical restraint increases 2.5 pp. from a 36% base; RNs per patient falls 1.3 pp. from a .04 base
Ford and Kasserman [17]	Time series of cross sections	50	1982–1989	State characteristics	Both presence of CON and CON stringency have a significant negative impact on both entry of new firms and expansion of capacity in dialysis industry	No direct evidence on quality; authors show CON increases market concentration, then cite other work saying that concentration reduces quality

Table 5

Summary of Studies on the Effect of CON on Access

STUDY	DESIGN	STATES	YEARS	KEY FINDINGS
National Studies
Fric-Shamji and Shamji [66]	Retrospective cohort	26	2004–2005	CON has 0% effect on procedure rates, but may shift care to non-profit hospitals
Fric-Shamji and Shamji [67]	Retrospective cohort	26	2004–2005	CON has 0% effect on procedure rates
Fric-Shamji and Shamji [68]	Retrospective cohort	26	2004–2006	CON has 0% effect on procedure rates, but may shift care to teaching hospitals
Popescu [55]	Retrospective cohort	50	2000–2003	CON reduces that chance that a patient with AMI is admitted for revascularization by 18%
Ho [65]	Retrospective cohort	50	1989–2002	CON results in 19.2% fewer PCIs being performed
Ho et al. [59]	Retrospective cohort	50	1989–2002	Removing CON increases PCIs and CABGs by 0%
Short et al. [69]	Retrospective cohort	50	1989–2002	CON has 0% effect on cancer resection procedures
Ho, Ross et al. [65]	Retrospective cohort	50	1989–2002	CON increases CABGs by 0%
Case Studies
DeLia et al. [70]	Retrospective cohort	NJ	1995–2004	Removing CON decreases racial disparity in cardiac angiography by 3%
Robinson et al. [42]	Retrospective cohort	PA	1994–1999	Removing CON increases CABGs by 0%
Kolstad [39]	Retrospective cohort	PA	1994–2003	Removing CON decreases travel distance for CABG by 2.3 miles (9%)

Table 6

Summary of studies on effect of CON on health spending

STUDY	YEARS	CONTROLS	KEY FINDINGS
Conover and Sloan [2]	HCFA; 1980–1998	State characteristics, State fixed effects	Dropping CON has a 0% effect on all expenditures
Lanning, Morrisey and Ohsfeldt [13]	HCFA; 1969, 1972, 1976–1982	2SLS accounts for endogeneity of CON	CON increases hospital spending 20.6%, overall spending 13.6%
Hellinger [63]	No source reported; 1985, 1990, 1995, 2000	State characteristics	CON decreases hospital beds by 10%, which in turn decreases spending by 1.8%
Rivers et al. [71]	AHA; 1999–2003	Hospital and state characteristics, state fixed effects	CON has a 0% effect on hospital spending; strict CON increases hospital spending 4.9%
Grabowski [72]	CMS; 1981–1998	State fixed effects	CON repeal increases Medicaid nursing home expenditures 0%

Database: Ovid MEDLINE(R) < 1975–2010> Search Strategy #1: Costs.
#	Searches	Results
1	exp certificate of need/	801
2	“certificate of need”.mp.	886
3	1 or 2	886
4	(cost$ or burden$ or impact$).mp.	736,742
5	exp “costs and cost analysis”/	157,748
6	ec.fs.	287,061
7	4 or 5 or 6	885,552
8	3 and 7	368
9	..l/ 8 lg = en	366
10	..l/ 9 yr = 1975–2004	325
11	..l/ 9 yr = 2005–2010	40

Database: Ovid MEDLINE(R) < 1975–2010> Search Strategy #1: Benefits.
#	Searches	Results
1	exp certificate of need/	801
2	“certificate of need”.mp.	886
3	1 or 2	886
4	(mortalit$ or morbid$).mp.	489,833
5	exp Mortality/	235,682
6	exp morbidity/	298,906
7	4 or 5 or 6	855,189
8	3 and 7	18
9	access$.mp.	243,710
10	3 and 9	58
11	8 or 10	73
12	..l/ 11 lg = en	73
13	..l/ 12 yr = 1975–2004	49
14	..l/ 12 yr = 2005–2010	24