You asked for an update of OLR report 99-R-1042, which listed the states that were using the Primary Care Case Management (PCCM) model to deliver health care to their Medicaid populations. You also asked us whether these programs have been evaluated and if so, whether the authors studying them found (1) improvements in health outcomes or (2) cost savings.

SUMMARY

Policymakers continue to look at PCCM as an alternative service delivery model for their Medicaid programs, especially as more managed care organizations (MCO) pull out of those programs. According to a June 2005 report by the National Academy for State Health Policy (NASHP), the number of states using PCCM exclusively or in combination with risk models (i. e. , MCOs) between 1990 and 2002 remained relatively constant. In 1999, we reported that 31 states used PCCM. The Kaiser Family Foundation reports that as of July 2005, 29 states were doing so.

According to the NASHP report, most of the PCCM states enroll people receiving family Medicaid coverage (e. g. , families receiving Temporary Assistance for Needy Families (TANF). Many states targeted individuals with chronic medical conditions and integrated disease management into their programs. PCCM programs are more likely to operate in rural areas, where MCO penetration is weaker but urban areas have also used them. In almost all states with PCCM, enrollment is mandatory, a phenomenon that has not changed over time. Some states that have combined PCCM and full risk programs make their MCO programs an option for people required to enroll.

Although most PCCM states have instituted quality assurance measures, some researchers suggest these are inferior to those used by states using a full-risk, MCO service delivery model.

State PCCM programs, many of which predate risk-based programs, have as their primary goals reducing Medicaid costs and improving patient outcomes. But an Internet literature search yielded few evaluations of PCCM programs, most of which are fairly dated. These generally reported on cost savings and service utilization and said little about patient outcomes. From this limited review, it appears that in some states, PCCM reduced hospitalization and service utilization. But in some, states tried PCCM and subsequently abandoned it for full-risk (MCO-only) models.

Many state Medicaid managed care programs, including states using the PCCM model, use disease management programs as a natural adjunct, since both are concerned with managing care utilization and patient health. Medicaid disease management programs have been more widely evaluated, perhaps, in part, because they target specific diseases (e. g. , asthma) and their effects are easier to measure. Evaluations of some of these can be found at www. npcnow. org

STATES USING PCCM MODEL IN THEIR MEDICAID PROGRAMS

Table 1 lists the 29 states that were using PCCM to serve some or all of their Medicaid recipients as of June 30, 2005. Illinois, as part of its new AllKids universal health program, plans to use the PCCM model and use its savings to cover more children.

Table 1: States Using PCCM in their Medicaid Programs

(as of June 30, 2005)

Table 1: –Continued-

Source: Kaiser Family Foundation, data retrieved in July 2006

Illinois

As part of its new universal health insurance for children initiative, AllKids, Illinois will use the PCCM model on a statewide basis. The estimated $ 56 million in cost savings (which includes moving kids in the state's publicly funded health care programs, Medicaid, and SCHIP into PCCM) is predicated on children getting more ambulatory, preventive care outside an emergency room or other hospital setting. The state released a request for proposals for a program administrator, who is also expected to perform outreach, education, and enrollment functions. Up until now, Illinois has served its Medicaid population using the fee-for-service model, with a small number of recipients in the Chicago area served by MCOs.

COMMON FEATURES OF PCCM

No two state PCCM programs look exactly alike, but many share common elements. For example, all offer access to care on a 24-hour a day, seven day a week basis. Some commonalities are discussed below.

Administration

The 2005 NASHP report points to numerous state strategies for PCCM administration. Texas has relied heavily on private contractors to perform program functions, from plan enrollment to claims processing. It has contracted with a plan administrator who is responsible for provider network development and contracting, provider services, member services, case management, utilization management, and health and wellness education. It even sets hospital rates for PCCM members.

North Carolina has worked hard to develop state and local partnerships. The Medicaid agency funds a position at the regional level who works with both members and providers.

Maine has probably the most complex program. Its Medicaid agency is responsible for overall program administration, and the state Unit on Aging is responsible for policy development and implementation. A private contractor handles program enrollment; the Aging Unit administers two separate contracts for services, one handling eligibility and the other services; and the Medicaid agency selects and contracts with physician practices.

Provider Types

In general, the primary care physicians (PCPs) include physicians who practice primary care (e. g. , internist, pediatrician, OB-GYN). In some states, PCPs can be physician “extenders,” such as RNs, physician assistants (PA), and others. And in a number of states, specialists act as the PCP, provided that they also provide needed primary care services. This latter category of providers tends to serve patients who have chronic medical conditions. Still other states use safety net providers (e. g. , federally qualified health centers) as the PCP. The NASHP study showed that while physicians were the main PCP during the study period, the number of PCCM states that had physician groups and clinics as well as specialists acting as PCPs grew significantly over time.

Provider Recruitment

In addition to the monthly per patient fee paid to the PCPs, the NASHP report asserts that other factors are equally critical to recruiting and keeping PCPs, including education and ensuring that expectations are known up front. Surveyed states used a number of strategies to get providers to join and stay with PCCM, including a telephone hotline, regional advisory committees, training, allowing for PCP input into program design, and other supports. Equally important is a program's ability to provide the PCPs information (e. g. , costs and service utilization) so they can better manage their patients' care.

Provider Payments

PCCM programs are believed to be attractive to providers for two main reasons: they can better manage their patients' care and they are compensated for doing so. In addition to receiving a fee for providing primary care services (which is almost always reimbursed using the traditional fee-for-service schedule), states with PCCM programs pay their PCPs a monthly per patient case management fee (typically $ 3).

The NASHP study found additional payment methods, including (1) enhanced payments, in which the state paid the PCP an enhanced rate for any primary care services; (2) shared savings, in which the state shared with the PCP any savings attributable to the PCP's management of the patient; and (3) capitation, in which states used what is referred to as prepaid inpatient and ambulatory health plans to pay for services. The latter two methods were used by an extremely small percentage of states, though over time, the use of the capitated method rose, as did the enhanced rate.

Additional Incentives

A number of the surveyed states were using or were planning to use physician profiles as a way for quality monitoring and for physicians to get feedback. Using a complex scoring system, Maine provided monetary rewards to those physicians ranking in the top 20^th percentile in each provider group based on its attaining specific goals (e. g. , reducing unnecessary emergency room (ER) use).

Prior Authorization (PA)

NASHP reported that all PCCM states required patients to obtain PA from their PCP before accessing care from other physicians. Over half required PCP approval before obtaining inpatient hospitalization, outpatient diagnostic care, and Early and Periodic Screening Diagnosis and Treatment (EPSDT) services. (It should be noted that the 1997 federal Balanced Budget Act effectively eliminated PCPs' ability to require PA for emergency room visits. Reducing the reliance on ERs for primary care has long been a goal of managed care).

Access Strategies

NASHP examined state strategies for ensuring access to services. States most frequently offered 24-hour coverage, provider to enrollee ratios, and maximum travel times and distance. And an increasing number of states allowed specialists to act as PCPs and imposed maximum waiting times for appointments.

Quality Oversight

NASHP indicated that most PCCM states reported using a number of strategies to monitor quality of care and services. The most common ones included enrollee surveys, focus groups, and hotlines; spot checking the 24-hour availability; and random medical record review. Less common was the use of external quality review organizations, disenrollment surveys, and provider credentialing. ER use, EPSDT, and immunization were the top three subjects of performance feedback.

Recognizing that PCCM programs were becoming more common as MCOs pulled out of the Medicaid market, researchers from Harvard and Boston University undertook what appeared to be the first study of the extent to which PCCM programs were monitoring performance. Their findings were published in Health Affairs (November/December, 2004). The authors surveyed and received responses from Medicaid directors from 46 states and the District of Columbia. The universe of Medicaid managed care programs reviewed included those with PCCM exclusively (nine), those with combined PCCM and MCOs (16), and those using MCOs only (22).

The researchers found, that compared to state MCO-only Medicaid managed care programs, most PCCM programs were not using quality measurements, feedback, and improvement strategies. The authors called the contrast between the two “striking” and suggested a minimum standard for states using PCCM.

States adopting the PCCM model are working hard to control service utilization. While service utilization results were reported most often, the researchers found that less than half of the states were collecting this information, while more than 80% of MCOs collected the data. The authors also found that feedback of performance results to the public

and physicians in PCCM programs was infrequent. For example, only 12% of PCCM states provided satisfaction of care feedback, the most commonly collected indicator, compared with 66% from MCO states (Quality Oversight in Medicaid Primary Care Case Management Programs, pp. 235-242).

An article in Health Affairs (January/February, 2003) suggests that in some states, PCCM performance monitoring is strong: Vermont, North Carolina, and Arkansas, all of which have used PCCM exclusively for some time, have instituted sophisticated mechanisms for monitoring their programs, include performance measurement and provider profiling. The authors suggest that these programs appear as “well-crafted, state- run MCOs” (Medicaid and Managed Care: A Lasting Relationship, pp. 77-88).

GAO Report. Between June 2001 and December 2002, the General Accountability Office (GAO) studied children's access to services under Medicaid and State Children's Health Insurance (SCHIP) programs in 16 states. The researchers compared risk-based models with FFS (with which, interestingly, they chose to include PCCM). Seven of the states were using PCCM.

None of the seven states that had PCCM programs set limits on the number of beneficiaries a specialist could serve, and few set specific standards for appointment waiting times. In general, PCCM programs had fewer standards than MCO models imposed. And while PCCM states had current information on PCCM providers, they did not indicate how easily beneficiaries could see their PCP. The study further found states with both risk-based and PCCM programs set PCCM appointment waiting time limits less frequently than they did under the MCO model—three of the states set no standards. Massachusetts was the exception, requiring PCPs to see new patients within a time frame while imposing no such requirement for MCOs. And of the four states that set these standards, only Texas reported monitoring compliance with them.

The report also looked at state monitoring of service utilization. Most often this came through an analysis of FFS claims data. The authors reported that the claims profiles often focused on high-cost services or those at risk of over-utilization. Three states also included primary and preventive care utilization. Specialist care was typically not monitored.

EVALUATIONS FOR COST SAVINGS AND PATIENT OUTCOMES

Alabama and Georgia

Alabama and Georgia have used a PCCM model for their Medicaid population since the mid-1990s. (Like a number of other states, both states also use a Prepaid Inpatient Hospital Plan (PIHP) for most of their Medicaid recipients' inpatient services, which is a risk-based, less than comprehensive care, model. Georgia also uses a similar model, prepaid ambulatory health plan (PAHP) for outpatient services. ) Researchers at the University of Alabama evaluated PCCM in Alabama and Georgia for both states (1996-99 in Alabama and 1994-97 in Georgia). They analyzed care use by a 25% sample of children between birth and age 18. Their findings were presented in Health Care Financing Review, Summer 2005.

The researchers tracked changes in Medicaid primary, preventive, and emergency care visits. They found a slight decline in primary and preventive care visits during the study periods, but they had hoped to see increases in these areas. They found a larger decline in emergency room utilization. Table 2 shows the changes in utilization for the two states.

Table 2: Effect of PCCM on Children Receiving Medicaid

in Alabama and Georgia

Arkansas

Arkansas established its Medicaid Primary Care Physician Management waiver program in 1993. The waiver required the state to evaluate the program by September 30, 1995.

The state began recruiting PCPs in October 1993. By September 1995, nearly 1,300 physicians had agreed to participate. Each provider was allowed to enroll up to 1,000 patients, but some could take more if they had a high Medicaid patient load before the program began or no other providers were located within a reasonable distance of the patients.

The program was implemented statewide. About 30% of the entire Medicaid population was exempted (e. g. , dually eligible, nursing home residents). Enrollment was staggered by Medicaid eligibility group with families receiving cash assistance the first to re-enroll. Others did not enroll due to delays so the actual number of participating clients was only 51. 8% of the state's Medicaid population.

The monthly management fee was $ 3 per month per patient. The PCP was expected to refer patients to primary care services instead of more costly specialists, unless referral was medically necessary.

A University of Arkansas evaluation was positive. The PCP enrollment rate was negatively related to Medicaid expenditures, indicating a cost savings estimated to be 6. 7% when adjusted for enrollment levels. And families (adults and children) receiving cash welfare benefits accounted for the greatest savings (16. 3% for adults and 13. 2% for children enrolled in June 1995).

Some cost areas showed statistically significant downward changes while others did not. For example, Medicaid inpatient hospital expenditures decreased by an adjusted rate of 39. 9% and outpatient hospitalization expenditures fell by 7. 9%, but physician expenditures declined by a statistically insignificant amount. As a whole, prescription drug expenditure reductions were not statistically significant. However, using an eligibility specific model, statistically significant drug expenditure reductions were found for adults receiving cash welfare. Overall, the researchers estimated the program saved $ 57 million in its first 17 months of operation (Evaluation of the Arkansas Medicaid Primary Care Case Management Program, Health Care Financing Review, Summer, 1996).

Iowa

An abstract published in Health Services Research (August 2006), reports on a University of Iowa study of Iowa's Medicaid PCCM program and its effect on Medicaid costs from 1989 to 1997. The study authors compared actual health care costs per member with “expected” costs in the absence of a PCCM model. They estimated that PCCM saved the state $ 66 million over the study period, with Medicaid expenses declining 3. 8%. This effect grew stronger over time. The authors reported increases in outpatient care and pharmacy expenses and decreases in physician and hospital expenses.

In a 1996 study some of the same researchers looked at the effect of PCCM on birth outcomes. Specifically, they compared prenatal care and birth outcomes in a PCCM environment with a strictly FFS one.

Iowa had been running its PCCM program as a demonstration program in seven counties. One hypothesis was that PCCM would improve the adequacy of prenatal care utilization and result in more positive birth outcomes. But, using a well-established index for measuring prenatal care utilization, the researchers found an apparent trend of FFS mothers having higher percentages of prenatal care. The effects of prenatal care on birth outcomes was inconclusive, though the researchers expected to see better outcomes in the FFS group given the higher levels of prenatal care (Primary Care Case Management and Birth Outcomes in the Iowa Medicaid Program, American Journal of Public Health, January 1997).

A 2001 NASHP paper reported on a later birth outcome study in Iowa that compared risk-adjusted quality of care provided to Medicaid mothers and babies in FFS, PCCM, and MCOs. The authors compared outcomes of PCCM and MCO in 1999 and MCO and FFS in 2000. The paper reported only one conclusion, which is a bit difficult to interpret: PCCM neonates had longer lengths of stay for very low birth weights, but MCO neonates had longer stays for low birth weights.

Rural Florida, North Carolina, and Oklahoma

University of North Carolina researchers looked at the effect of PCCM in rural Florida, North Carolina, and Oklahoma. Rural areas have typically been harder to serve under risk-based managed care because the smaller numbers of Medicaid enrollees do not provide the financial incentives health plans find in more urban areas where Medicaid populations are significantly higher. They found that Medicaid recipients living in rural areas benefited from PCCM's additional clinicians and individualized care associated with disease management. But care management was more difficult to maintain because of the enrollee dispersion. And when states implemented phone management of care to address this, inconsistent access to phones often made management elusive.

North Carolina uses the PCCM model almost exclusively. The state began a pediatric asthma disease management program in 1998, which was found to lower hospital admissions by 34% and emergency room visits by 8%. And although medication costs rose (an expected result), the average episode cost for children in PCCM was 24% less than for other children. A separate study found a 37% decrease in the average monthly rates of non-urgent ER visits.

Virginia

Researchers from Virginia Commonwealth University and the University of Virginia evaluated immunization rates across Virginia's three Medicaid managed care models: PCCM and mandatory and voluntary MCO enrollment programs. They looked at children who turned two during the period between July 1, 1996 and June 30, 1997 by surveying their physicians in 1998. The overall response rate was almost 80%.

They found that immunization rates in the PCCM programs were consistently higher than the rates in the MCO groups. The voluntary MCO group's rates were higher than that of the mandatory group. They concluded that the type of managed care program in which the child participated was a key predictor of immunization status.

The researchers noted a few limitations of the study. First, mandatory MCO enrollment was the most recently introduced model compared to PCCM, which had been operational for a while. And one MCO pulled out of Medicaid, which could have affected immunization compliance.

Virginia's PCCM program contains an adjunct disease management program. The Virginia Health Outcomes Partnership was implemented in 1995 as a pilot asthma management program for Medicaid recipients. (It was the nation's first disease management program in FFS Medicaid. ) A widely touted 2000 study of the program found that the number of emergency room visits dropped, while dispensing of asthma relief drugs (the two outcomes measures used) rose when Medicaid PCPs in the pilot who received training were compared with other community physicians. The authors projected a direct savings of Medicaid of $ 3 to $ 4 for every incremental dollar spent on providing disease management support to the PCPs. (But not all disease management program evaluations have been positive. For example, a 2001 study of Florida's program, which projected significant savings, actually cost the state money and was plagued by problems (that were subsequently addressed)).

States Abandoning PCCM for MCO-Only Model—Maryland and Mississippi

Maryland. Researchers at the Project Hope Center for Health Affairs in Bethesda evaluated Maryland's PCCM program in 1997. It used five years of Medicaid claims and eligibility before and after the program began to determine its impact on expenditures and services utilization. The study's abstract states that PCCM increased the probability that a Medicaid enrollee would use primary and preventive services, but PCCM had little effect on use of specialty or ER services. And although the program successfully controlled program expenses, largely through a reduction in the use of “ancillary” services, the increase in the use of services resulted in an estimated 3. 4% increase in Medicaid expenditures.

Maryland moved from a mixed model to MCO-only model in 1998. The state extensively evaluated the new program and reported its findings in 2002. The evaluation was generally favorable and suggested that the program's success was due largely to a built-in “medical home” for each enrollee and MCO care management, two aspects of PCCM that carried over. But perhaps an even stronger reason for its success was that the MCOs paid providers more than they received under the FFS system.

Some of the successes included increases in well-child visits and the number of children receiving them, decreases in ER use, and progress in access to dental care. But results for access to care for children with special health care needs were mixed: fewer children in foster care received services and the amount of services declined, while children eligible for Supplemental Security Income (SSI) experienced improved access. (Interestingly, a 2000 health journal article referenced a 1998 study of the state's PCCM program that found a 120% increase in the probability of an enrolled child having a well-child visit, with a 10% decrease in his probability of having an avoidable hospitalization).

A 1999 Urban Institute report suggests that a major impetus for the change was a strong push by Governor-elect Gledenning in 1994, who felt that mandating Medicaid enrollees participation in MCOs or other capitated plans was the state's best way to achieve spending cuts. The “medical home” aspect of managed care was carried over from PCCM. The report also suggests that the reforms produced smaller savings than expected, in part due to advocates successfully lobbying for required services and fixed payments to providers (e. g. , hospitals). An official from the state's Medicaid agency adds that patient access to care has improved significantly as the number of providers MCO networks can offer far exceeds the number available under PCCM.

Mississippi. Mississippi began experimenting with Medicaid managed care in the 1990s. It used both PCCM and traditional MCOs. The PCCM program, HealthMACS, ran from 1993 until April 2002. Researchers from the Mississippi Health Policy Research Center at Mississippi State University evaluated HealthMACS and reported their findings in a February 2006 brief. The researchers conducted interviews with all stakeholders, including providers, consumers, legislators, and state agencies. They also analyzed state and national health care organization documents.

Begun as a pilot, HealthMACS operated in seven counties and went statewide in 1997. The authors suggested that its biggest problem was administrative difficulties created by its structure, with frequent communication failures. Clients found it difficult to communicate with their PCPs after hours and often ended up in ERs. And the ERs needed a special code to be reimbursed for this care. The code changed monthly, making reimbursement more difficult.

The authors also cited communication problems between state agencies. Before HealthMACS, the state's health department acted as a PCP for many Medicaid recipients; after it was implemented, the Medicaid agency would not allow this. This was particularly problematic for children's care as the health department had historically been the major provider for both immunizations and EPSDT screening. The authors asserted that HealthMAC's PCPs lacked an effective tracking mechanism to ensure that children received the screenings.

Utah. The Congressional Budget Office evaluated Utah's PCCM program in 1988. Utah had one of the country's first Medicaid case management programs. Savings from reductions in unnecessary use of services were expected to more than offset the increase in costs resulting from an increase in access to care. According to the study's abstract, findings revealed that the use of PCP services increased significantly, but so did the use of specialty services and prescription drugs (both of which the program administrators had hoped to lower). But hospital use for outpatient services declined. While the program achieved its goal of increased access, it was less successful in containing costs, which rose by 25% in the program's early years.

According to the Kaiser list of PCCM states, Utah currently uses PCCM for relatively few Medicaid recipients (those living in rural parts of the state), with the vast majority being served with the PIHP and PAHP model. (Since 2002, the state has provided Medicaid to low-income adults under an 1115 waiver in which the state expanded access to Medicaid but reduced the services to certain individuals and increased cost sharing. )

RC: ts