Challenging the cost effectiveness of medi-cal managed care.

INTRODUCTION

Researchers and consultants have promoted expansion of Medi-Cal managed-care (MCMC) to additional Medi-Cal beneficiaries currently covered under the Medi-Cal Fee-for-Service (FFS) program to achieve greater cost efficiency and quality of care. Proponents have also promoted MCMC as a cost-effective way to expand state-subsidized health insurance for many of the State’s 6.5 million uninsured,1 even though claims of this cost effectiveness have been disputed.2 This paper presents data that challenge the cost effectiveness of MCMC, and suggests that cost savings may actually represent cost shifting to the Medi-Cal FFS system. This in turn places an unfair burden on emergency physicians and other fee-for-service Medi-Cal providers. This cost shifting appears to have been facilitated by the unique manner in which MCMC has been implemented, allowing health plans to not enroll or dis-enroll the most costly beneficiaries without a concomitant adjustment in the state’s per-member-per-month capitation payments. Because of the highly skewed distribution of the cost of care, shifting even a small number of high-cost patients from a MCMC health plan into the Fee-for-Service program allows these plans to falsely promote the reduced monthly cost of care per enrollee as the result of cost-effective care management.

Of the 6.62 million Medi-Cal enrollees in the fiscal year 2007–08, 3.33 million were enrolled in MCMC, which receives $6.06 billion of the $33.98 billion total Medi-Cal budget. Even a small overestimate in the monthly cost per member used to calculate capitation rates has a large aggregate financial impact, easily reaching hundreds of millions of dollars annually. Ever since the Medi-Cal program began shifting its beneficiaries into MCMC in 1994, numerous claims have been made regarding the beneficial fiscal impact of the managed-care model on program expenditures.3 Organizations such as the California Legislative Analyst’s Office4 and the Little Hoover Commission have touted the ability of MCMC to contain costs. Based on these assertions, MCMC proponents have recently introduced state legislation (SB 1332) to expand MCMC enrollment for aged, blind and disabled (ABD) populations, many of whom are prime candidates for dis-enrollment and carve-out cost shifting once per member per month (PMPM) capitation payments are set. However, there is little evidence that substantiates anticipated savings in the Medi-Cal program. In (all appendixes are available online as a related file at http://repositories.cdlib.org/uciem/westjem/vol10/iss2/art17/), the author assesses the validity of the claims in support of, and challenging, managed-care cost effectiveness in California and elsewhere in the country.

The Skewing of Cost Distribution in Medi-Cal

Selection bias in managed-care is a systematic assignment of beneficiaries to a health plan based on health needs. The Medi-Cal program employs a methodology for setting capitation rates based on the average per capita cost of the eligibility cohort or population to which each beneficiary belongs. If, however, the population actually enrolled is healthier and therefore less expensive than the estimates for the eligible population as a whole (risk-averse selection), the plan benefits. Economists have studied the distribution of healthcare costs within populations and consistently found, both over time and across patient groups, that medical spending tends to be highly concentrated among a small percentage of patients, rather than spread evenly through the population.5,6 This skewed cost distribution means that the departure of even a small number of high-cost patients can have a large impact on the average cost of care for the remaining population. This perverse disincentive also potentially discourages health plans from including the most qualified specialists in their provider networks, lest this attract potential enrollees with more complicated health problems. reviews the relevant literature on the distribution of costs in healthcare populations.

Focusing on the Medi-Cal population, Thomas MaCurdy et al.7 found that, “Medi-Cal spending is extremely concentrated among a small segment of the enrollee population. An enormous share of all expenditures goes to a small number of cases; 60 percent of all Medi-Cal expenditures went to benefits for only five percent of the enrollees.” The same study noted a near absence of expenditures for services associated with the least expensive 25% of the Medi-Cal population.

The per capita cost method of calculating capitation rates used by Medi-Cal is similar to the methodology employed, and then abandoned, by the Medicare program during the 1990s. Then, actuaries discovered that the introduction of managed-care into the Medicare program led to the “favorable selection” of healthier and less expensive beneficiaries into health plans, and subsequent overpayments to these plans.8 Since the same conditions appear to exist in MCMC as did in Medicare, this should cause concern over the possibility that MCMC plans might be receiving an inappropriately higher allocation of limited Medi-Cal program funds. reviews the Medicare experience with risk adjustment of Medicare managed-care capitation rates.

To understand the impact of the skewing of costs in the Medi-Cal population, it may help to consider a hypothetical TANF (Temporary Aid to Needy Families) population of 200,000 Medi-Cal patients in a medium-sized California county. Assume an average cost of $125 PMPM, with eight months of average enrollment during the year, and total annual expenditures of $200 million. Next, assume that the patients are randomly assigned to two groups of 100,000 each, one enrolled in MCMC, and the other in the FFS program. Applying the distribution found by Berk and Monheit in 1996, and assuming a $125 average PMPM cost, the distribution of expenses for this hypothetical TANF patient population would appear as in the table in Exhibit 1. (Additional supporting tables for this illustration are in ).

Figure 1.

Changes to average per member per month costs resulting from movement of most expensive beneficiaries from managed care (MC) to fee-for-service (FFS)

The shift of the most expensive 1% of patients (for example, those with chronic illnesses) out of MCMC and into FFS causes the average PMPM cost of the remaining population to decline 26% or $32, from $125 to $92 (Figure 1). Shifting the most expensive 10% reduces the PMPM cost of the remaining population to $43. The shifting of the most expensive 1% of patients to FFS causes the average PMPM cost of the expanded FFS population to increase by $32, from $125 to $157. Shifting the most expensive 10% increases the PMPM cost under FFS to $192. As this illustration demonstrates, the shifting of a very small number of high-cost beneficiaries from managed-care to FFS can drastically alter the average cost of both populations. To an observer unfamiliar with this phenomenon, it would be easy to mistakenly conclude that managed-care does hold down costs.

Aggregate payments to the managed-care program reflect the loss of member months as the most expensive patients are dis-enrolled. The PMPM rates, however, are not recalculated to reflect the reduced average cost of the population, and this significantly improves plan profitability, even though capitation rates paid to MCMC plans are reduced by 5% to build in some of the anticipated cost savings. For example, using a capitation rate set at $118.75, which is predicated on 95% of the original average PMPM cost of the population, the shifting of even the most expensive 1% of 100,000 beneficiaries originally enrolled in managed-care to FFS would result in an overpayment of $2.6 million a month, or $31.6 million a year. Unfortunately, these plans bank these unearned profits while the burden for the care of these patients is shifted to FFS providers at rates that are among the lowest in the country.9

Causes of Non-Enrollment in Mandatory Managed-care Enrollment Populations

During calendar year 2005, 1,120,964 Medi-Cal beneficiaries assigned to aid codes for which enrollment into a MCMC plan is mandatory were initially enrolled under the FFS model (47%), or subsequently moved from MCMC into FFS (53%). These are disturbing figures that may not have been appropriately considered in the determination of PMPM costs or capitation rates. Certainly, the previous illustration suggests that the cost-shifting impact of selection bias that eliminates one-fourth of all potential MCMC enrollees is likely to be huge. The reasons a mandatory beneficiary may not be enrolled, or may have dis-enrolled from a health plan, are described in , and include logistical problems such as the patient changing counties or being homeless, pregnancy, chronic diseases such as HIV/AIDS or cardiomyopathy, current cancer therapy, consideration for organ transplant, or pending major surgical procedures. All of these would be expected to drive up the cost of care for these prospective MCMC beneficiaries.

There appear to be policy loopholes and incentives allowing or encouraging pregnant women to never enroll into MCMC plans, or to dis-enroll when they become pregnant or near delivery. (In 2005, 71% of deliveries were under FFS.) As a consequence, some MCMC plans had higher rates of deliveries relative to the rates of initial enrollment of pregnant women in MCMC than other plans, even though the administrative and demographic characteristics of the enrollees were equivalent. Having identified this discrepancy, Mercer and Associates,10 the actuarial consultant for the Department of Health Care Services, recommended that the Medi-Cal program implement a maternity supplemental payment to the plans to cover the cost of all deliveries to level the playing field. The department recently followed this recommendation, providing approximately $7000 in supplemental payments to MCMC plans for each delivery (which, perversely, may reverse the incentive and encourage MCMC providers to provide prenatal care under FFS and enroll these pregnant women into MCMC as they near term).

Low Rates of Managed-care Enrollment in Voluntary Populations

The rules determining enrollment of Medi-Cal beneficiaries into MCMC plans are based on the “Plan Model Type” of their county of eligibility. There are three major types of MCMC plans: 1) County Organized Health System (COHS), where there is one health plan run by a public agency and governed by an independent board. Nearly all Medi–Cal enrollees residing in the COHS are required to receive care from this system; 2) the Geographic Managed Care (GMC) system allows Medi–Cal beneficiaries to choose to enroll in one of many commercial health maintenance organizations (HMOs) operating in a county; and 3) the Two–Plan Model, which consists of counties where the department contracts with only two managed-care plans, one locally developed and operated, the other a commercial HMO, selected through a competitive bidding process. In the 12 “Two-plan” model counties and the two GMC counties, certain aid codes are considered “voluntary” for managed-care enrollment. Typically, beneficiaries in aid codes for the aged, blind and disabled are considered voluntary. Additionally, newborn babies with birth complications receiving care from the California Children’s Services (CCS), are often enrolled in aged, blind and disabled aid codes; and in that way, are not enrolled into MCMC. A disabled or presumptively disabled premature newborn is assigned to a disabled aid code. The premature beneficiary is then considered enrolled in FFS, not into MCMC. These enrollment rules enable many MCMC plans to manipulate enrollment and selectively limit their exposure to high-cost patients.

According to the Department of Health Care Services, enrollment data for beneficiaries assigned to voluntary aid codes reveal that a large proportion of the population is not enrolled into managed-care. To further complicate matters, only a small proportion of the mandatory population has enrolled into a MCMC delivery system. For example, while foster-care aid codes are considered mandatory, only 10 to 12 percent of this population enrolls into managed-care plans. Foster care children are much more likely to be afflicted with psychological disorders or behavioral problems and display PMPM costs that are as much as twice that of other children of similar age, gender, and ethnicity (source: DHCS Medical Care Statistics Section).

Capitation rates based on the average cost of a given population are effective when all members of the population enroll in a plan, but become less accurate and appropriate as increasing numbers dis-enroll. In the voluntary aid code population, where 85% of the potentially eligible beneficiaries may never enroll, the difference between the cost of the potentially eligible population and the population that actually enrolls is likely very great. This is especially true if the sickest members of the population never enroll, but remain in the FFS model. The recent decision to require quarterly re-enrollment of beneficiaries in California offers MCMC plans yet another avenue to selectively manipulate enrollment.

The Impact of “Carve-Out” or Excluded Services

Another argument for the greater cost efficiency of MCMC is the slower rate of increases in Medi-Cal capitation rates compared to other indicators of health cost inflation. Data show, for example, that Medi-Cal capitation rates have risen more slowly than the medical care component of the Consumer Price Index, or employer health-insurance premiums. Alluding to the implied cost effectiveness of MCMC, Mark Smith,11 president of the California Health Care Foundation, told a 2007 conference audience, “The growth of Medi-Cal Capitation Rates has been substantially less than commercial premiums.” While the figures may be accurate, the comparison itself is misleading. Not only does this comparison ignore the underlying age and health status differences of the populations being compared; it also ignores the difference in scope of services associated with these reported cost increases.

The Medi-Cal program places a wide range of services, known as “carve-outs,” outside the scope of its capitation payments to the managed-care plans for the purposes of rate setting and contractual service obligations. Medi-Cal patients who need these services must obtain them under traditional FFS arrangements and not through their plans. Carve-outs may include: CCS (California Children’s Services) payments; Rural Health Clinics/FQHC’s (Federally Qualified Health Clinic) wrap-around payments; Long-Term Care in Skilled Facilities; Nursing, drugs and services to patients receiving treatment for mental illness; treatment for HIV; and surgeries for organ transplantation. This process limits the exposure of the MCMC program to higher cost services, more expensive specialty care, and less predictable overhead.

While capitation payments made to MCMC plans may not have increased as much as employer-based insurance premiums, the cost of the carved-out services that are excluded from capitation agreements with MCMC plans certainly have. Between 2000 and 2006, FFS payments for MCMC patients rose by 113%. Some of this increase reflects the 27% expansion of MCMC enrollment during this period. However, the PMPM cost of carved-out services also rose by 67%, indicating that on a per-person basis, carve-out costs rose by an average of 9.5% annually. In calendar year 2006, Fee-for-Service carve-out payments made by the Medi-Cal program on behalf of MCMC enrollees totaled slightly less than 1.2 billion dollars12 (Figure 2). Of these, $616 million, or 52%, were related to the CCS and the Genetically Handicapped Persons Program (GHPP). (See below).

Figure 2.

Fee-for-service carve-out payments for Medi-Cal managed care enrollees.

CCS is a statewide program that treats children with certain physical limitations and chronic health conditions or diseases. GHPP provides health coverage for Californians 21 and older who have specific genetic diseases, including cystic fibrosis, hemophilia, sickle cell disease, and certain neurological and metabolic diseases. GHPP also serves children under 21 with GHPP-eligible medical conditions who are not financially eligible for CCS. While it may have been assumed that services paid for by California Children’s Services (CCS) would remain static or fall in counties where managed-care has been introduced, MCMC plans have aggressively utilized CCS to provide care for premature and compromised infants, driving fee-for-service CCS payments for Managed Care beneficiaries from $230.2 million in the fiscal year 1999–2000 to $517.8 million in the fiscal year 2004–2005 (Figure 3).

Figure 3.

Change in California Children’s Services payments under fee-for-service and Medi-Cal Managed Care

A FQHC is a community-based health organization that provides comprehensive primary care, dental and mental health/substance abuse services to underserved, underinsured and non-insured populations. FQHC carve-out payments represent 41% of the total cost of MCMC carve outs not related to CCS/GHPP. Under the federal Medicaid statute, when a contract between a managed-care organization and a FQHC results in the FQHC receiving less than the amount of reimbursement due under the FQHC prospective payment system (PPS), the state must make a supplemental “wrap-around” payment to the FQHC to make up for the difference the FQHC is owed. Between 2000 and 2006 there has been a dramatic increase in the use of FQHC providers by MCMC patients. During this period MCMC wrap-around payments increased by 211%, while the number of managed-care beneficiaries utilizing FQHCs increased by 154% (Figure 4).

Figure 4.

Medi-Cal Manged Care Federally Qualified Healthcare Center wrap around payments and users - calendar year 2000–2006

The rapid increase in FQHC utilization by MCMC beneficiaries, which greatly outpaces the growth in MCMC enrollment, is worrisome for three reasons:

It suggests that there are inadequate numbers of non-clinic affiliated physicians and physician groups to enable the plans to form a primary care network for Medi-Cal beneficiaries.

Visits by managed-care patients to FQHC providers create millions of dollars in additional FFS carve-out expenditures that would not exist if the health plans were able to contract with non-FQHC affiliated physicians and physician groups.

While Medi-Cal would still be required to pay the higher FQHC rate for services incurred at clinics under a FFS scenario, it would do so without, in addition, paying a portion of a capitation rate to a health plan for these subcontracted primary care services. For the care of these patients, MCMC is ‘double dipping’ into a limited funding pool.

CONCLUSION

In light of the indicators of cost shifting identified in this review, the assertion that the MCMC program is cost effective is highly suspect. A number of questions need to be answered before program administrators can claim with confidence that billions of dollars in Medi-Cal funds are being properly allocated, and that expansion of these programs to cover other beneficiaries is indicated. These questions include but are not limited to:

What is the distribution of costs for mandatory enrollment code beneficiaries that remain in, or are excluded or dis-enrolled from, MCMC?

What is the risk-adjusted distribution of costs for voluntary enrollment codes that elect to enroll in MCMC vs. FFS?

When adjusted for risk, how do costs compare between MCMC and FFS enrollees?

What is the actual medical loss ratio for MCMC plans, after full consideration of all contracting plan and subcontracting provider group tiers in the MCMC model?

What percentage of pregnant managed-care enrollees are dis-enrolled prior to delivery?

What are the true costs of carved-out services?

Do certain mandatory aid code categories with higher relative cost risk have a lower rate of enrollment in MCMC?

How do risk-adjusted capitation payments affect Medicaid enrollment patterns, carve-outs, and provider networks?

The data presented here indicate that the MCMC program is subject to perverse incentives that adversely impact the Medi-Cal safety net. It is widely acknowledged that the average expenditure per enrollee in the Medi-Cal Program (California’s version of Medicaid) is one of the lowest in the nation. FFS providers, especially those obligated under Emergency Medical Treatment and Active Labor Act (EMTALA) to provide emergency care, believe they have had to bear the increasing burden of caring for those patients who are maneuvered out of MCMC enrollment, in return for FFS payments that inevitably fail to cover the provider’s costs. The data presented suggest that expansive carve-outs, enrollment loopholes, and dis-enrollment incentives misdirect limited Medi-Cal program funds to plans and provider groups, rather than to the FFS providers that actually provide a disproportionate share of services to high-cost Medi-Cal beneficiaries. The magnitude of this shift easily approaches several hundred million dollars per year. This undermines the potential benefits of applying the managed-care concept to those that need management the most, and under-compensates providers that actually render services to these patients under fee-for-service.

Alternatively, aligning incentives through appropriate risk-adjustment of capitation rates by paying much higher cap rates 1) to cover services that are currently carved out, 2) for patients that are currently exempted from managed care, and 3) for higher cost patients that are currently being selectively not enrolled or dis-enrolled, and then closing these loopholes, could provide sufficient incentive for plans and capitated provider groups to actually focus on the cost-effective and innovative case management of these patients. This approach might also encourage plans to enhance their network of qualified providers to accomplish this goal, and reduce the program’s reliance on EMTALA-obligated FFS emergency care providers and expensive emergency department visits to meet these patients’ needs.

Table.

Distribution of Health Expenditures for the U.S. Population, By Magnitude of Expenditures2

Percent of US Population Ranked By Percentile	Percent of Expenditures (1996)
Top 1%	27%
Top 2%	38%
Top 5%	55%
Top 10%	69%
Top 30%	90%
Top 50%	97%

Exhibit 2 Table.

Impact of Dis-enrollment of Most Expensive Beneficiaries on Average Cost of Remaining Managed Care Population

Percent of Population	Number of Unduplicated Beneficiaries	Amount of Expenditures	Member Months	PMPM Cost
Entire Population	100,000	$100,000,000	800,000	$125.00
Less Most Expensive 1%	99,000	$73,000,000	792,000	$92.17
Less Most Expensive 5%	95,000	$45,000,000	760,000	$59.21
Less Most Expensive 10%	90,000	$31,000,000	720,000	$43.06

Exhibit 3 Table.

Impact of Dis-enrollment of Most Expensive Beneficiaries from Managed Care and Back into Fee-For-Service

Percent of Population	Number of Unduplicated Beneficiaries	Amount of Expenditures	Member Months	PMPM Cost
Original Population (FFS)	100,000	$100,000,000	800,000	$125.00
Adding Most Expensive 1% from MC	101,000	$127,000,000	808,000	$157.18
Adding Most Expensive 5% from MC	105,000	$155,000,000	840,000	$184.52
Adding Most Expensive 10% from MC	110,000	$169,000,000	880,000	$192.05

Exhibit 4 Table.

Change in Average Per-Member, Per-month Cost of the Managed Care Population, and Resulting Over-payment, as the Most Expensive Beneficiaries are Shifted to Fee-For-Service

Expenditure	Original Population	MC’s Most Expensive 1% Shifted to FFS	MC’s Most Expensive 5% Shifted to FFS	MC’s Most Expensive 10% Shifted to FFS
Capitation Payment at 95% Original PMPM ($125)	$118.75	$118.75	$118.75	$118.75
PMPM for Remaining Population	$125.00	$92.17	$59.21	$43.06
Overpayment (PMPM)	$(6.25)	$26.58	$59.54	$75.69
Remaining Beneficiaries	100,000	99,000	95,000	90,000
Aggregate Monthly Overpayment	$(625,000)	$2,631,250	$5,656,250	$6,812,500
Aggregate Annual Overpayment	$(7,500,000)	$31,575,000	$67,875,000	$81,750,000

Table.

Beginning cost distribution of each group of beneficiaries in hypothetical California county

Total number of beneficiaries	100,000
Member months	800,000
Total annual expenditures	$100,000,000.00
Percent of unduplicated beneficiaries	Top 1%	Top 5%	Top 10%	Top 50%	Bottom 50%
Number of unduplicated beneficiaries	1,000	5,000	10,000	50,000	50,000
Amount of expenditures	$27,000,000	$55,000,000	$69,000,000	$97,000,000	$3,000,000
Percent of expenditures	27.0%	55.0%	69.0%	97.0%	3.0%
Member months	8,000	40,000	80,000	400,000	400,000
Per member per month	$3,375.00	$1,375.00	$862.50	$242.50	$7.50

Applied to Temporary Assistance for Needy Families beneficiaries in hypothetical county. Assumes $125 per beneficiary per month average cost.