Evaluation of the Arizona health care cost-containment system

Abstract

This article evaluates Arizona's alternative to the acute portion of Medicaid, the Arizona Health Care Cost-Containment System (AHCCCS), during its first 18 months of operation from October 1982 through March 1984. It focuses on the program's implementation and describes and evaluates the program's innovative features. The features of the program outlined in the original AHCCCS legislation included: Competitive bidding, prepaid capitation of providers, capitation of the State by the Health Care Financing Administration, assignment of gatekeepers, beneficiary copayment, private administration, inclusion of private and public employees and county financed long-term care. An assessment of implementation during the second 18 months of the program reporting on more recent developments and is now being prepared by SRI International.

Introduction

The Arizona Health Care Cost-Containment System (AHCCCS), is a innovative system for providing acute medical care services to the indigent population in Arizona, the only State without a traditional Medicaid program. Arizona receives Federal funding for AHCCCS as a demonstration project of the Health Care Financing Administration (HCFA).

This system differs substantially from other States indigent health care programs. It selects its providers through a competitive bidding process; these providers are reimbursed under a prepaid capitation system. The State is reimbursed on a prepaid capitation basis by the Federal Government. In this program, beneficiaries are assigned to a particular gatekeeper who manages their medical care, and beneficiaries are required to pay small copayments for services they receive. The original legislation required that most of the program's administrative functions be contracted to a private Administrator. The State has now taken over this function, although it has retained the option to contract out specific functions. The original legislation called for the program to include private, State, and county employees in addition to the indigent. To date, these groups have not been included. The counties determine eligiblity for medically indigent and medically needy beneficiaries (MI/MN's) and continue to provide long-term care services and other services they had previously provided that are not covered under AHCCCS.

AHCCCS innovations potentially can be replicated in other geographic areas, and for programs other than Medicaid. Consequently, it is important to determine how well the program has worked. In particular, it is necessary to determine whether the program provides access to high-quality care at a lower cost than do conventional Medicaid programs.

Before AHCCCS, Arizona's indigent-care programs were administered by the individual counties. Eligibility criteria, as well as which services were covered, differed from county to county. Prior to 1980, legislation to participate in the Federal Medicaid program was routinely introduced and defeated. In 1980, when Arizona passed legislation limiting local property taxes, county revenue was no longer adequate to cover the rising cost of indigent health care and the counties needed to find a way to get Federal and State support. However, many Arizona legislators continued to be skeptical about participating in the Federal Medicaid program because they were concerned with program costs. In 1981, negotiations between HCFA and Arizona legislators concluded with an agreement on a 3-year demonstration project, which began in October 1982.

Eligibility for AHCCCS includes all categorically eligible groups under Medicaid—Aid to Families with Dependent Children (AFDC), Supplemental Security Income (SSI) recipients, medically indigent (MI), and medically needy (MN). For AFDC and SSI AHCCCS eligibles, the State receives Federal matching funds. Those eligible for these programs are considered categorically eligible for AHCCCS. MI and MN beneficiaries are poor individuals who meet Arizona requirements for eligibility for AHCCCS services but who are not categorically eligible. The State does not receive Federal matching funds for MI/MN's. As of May 1984, of the approximately 190,000 beneficiaries eligible for AHCCCS, 44 percent were AFDC beneficiaries, 19 percent SSI beneficiaries, and 38 percent MI or MN beneficiaries. Benefits covered include most acute-care services: hospital, physician, laboratory, X-ray, medical supplies, pharmacy, and emergency services. Skilled nursing facility and home health services are not included in the AHCCCS program's benefits.

The evaluation of this demonstration is being conducted under contract to HCFA by SRI International in Menlo Park, Calif. The evaluation team includes Actuarial Research Corporation in Annandale, Va., and Research Triangle Institute in Research Triangle Park, N.C. The evaluation is divided into two main parts: program outcomes, and the implementation and operation of AHCCCS. The evaluation, which began in July 1983, will last 39 months. This summary covers the first 18 months of program operation (October 1982 through March 1984).

Little information is available on program outcomes, however, case studies have been done on the implementation and operation issues. This article begins with a description of two major implementation difficulties in the first year—the administration of the program, and beneficiary eligibility and enrollment. It then describes the bidding process in the first and second years, the participating plans, the reporting of financial and utilization data by the plans, and the financial status of the AHCCCS program. Other issues discussed include quality assurance and patient satisfaction procedures, county involvement in AHCCCS, (both through the provision of long-term care services and the impact on the previous county delivery systems) and the involvement of the private State and county employees in the program. Following this, there is a discussion of the major problems and achievements of AHCCCS in the first 18 months of the program, and a discussion of policy implications.

Program administration

In designing AHCCCS, the Arizona legislature chose to contract out the administration of the program to a private Administrator. Although many Medicaid programs contract with fiscal intermediaries for claims processing, no State had contracted for the entire administration of the program. The main reason the legislature chose to do so was its belief that the private sector could operate the program in a more cost-effective manner than a government agency could.

The legislature gave the AHCCCS Division of the Arizona Department of Health Services responsibility for policy and program oversight; and the private Administrator was to handle day-to-day responsibilities. The legislature initially limited the permanent AHCCCS Division staff to 24. Their specific responsibilities were to draft policy and regulations for the program, to do planning and research, financial management, and public relations.

Through a competitive bidding process that attracted six bidders, the AHCCCS Division selected MCAUTO Systems Group Inc. (MSGI) as the program Administrator. The contract period was for 39 months. During that time, the private Administrator's responsibilities included:

• Claims and encounter data processing—Paying claims for services delivered on a capped fee-for-service basis, processing encounter data from providers, making capitation payments, and producing utilization reports to support the management of the program.

• Provider relations—Conducting (with the State) the competitive bidding process for procurement of the prepaid plans, providing technical assistance to providers, monitoring overall performance and contract compliance, and establishing programs of quality assurance and utilization control.

• Enrollment—Providing training to counties on procedures for MI/MN eligibility, notifying individuals of their eligibility, requesting eligibles to choose a plan, conducting enrollment interviews, conducting the open-enrollment process, and assigning to a particular plan those who are eligible but not enrolled.

• Quality assurance—Assisting plans in establishing quality assurance programs, reviewing plans' quality assurance activities, and identifying potential cases of underutilization of services.

There were difficulties in each of these areas during the first 18 months of the program. Perhaps more than any other area, problems associated with eligibility and enrollment adversely affected the administration of the program. Resolution of these problems diverted major resources from long-term administrative tasks. The short time allowed for implementation, the lack of experience with programs similar to AHCCCS, rapid program changes (especially in the enrollment area), and unique systems requirements all affected the ability of the private Administrator to perform these functions. In addition, turnover of key staff and an underestimation of the resources required to do the job made it difficult for the Administrator to accomplish what was required.

During the first 18 months, priority was given to processing plans, having services provided, paying providers, and dealing with day-to-day problems such as those related to eligibility and enrollment. The Administrator delayed implementing many long-term activities, such as quality assurance, provider monitoring, collection and use of encounter data, and reporting of information.

The problems associated with program administration led to tension between the AHCCCS Division and MSGI over the latter's performance and cost. Costs by MSGI increased from the original bid of $11 million for the 39-month contract period to an estimated $30 million in February 1984. Contract disputes between the State Attorney General and MSGI led MSGI to notify the State on February 14, 1984, that, if the contract disputes were not resolved within 30 days, MSGI would terminate the contract. The State assumed the responsibility for administering the program as of March 14, 1984.

The Governor appointed a transition team composed of officials from several departments of the State government to manage the transfer of the MSGI data processing system, general functions, and personnel to the State. This system transfer was accomplished smoothly, and most of the MSGI personnel accepted positions with the State. Two areas in which there were difficulties in retaining MSGI personnel were in the management information systems and the grievance staff. A new AHCCCS director was named, and the administration was reorganized. Overall, the transition to State government administration was achieved with a minimum of disruption to the program. A new administrative organization now is in place.

Eligibility and enrollment

Eligibility determination and enrollment into the plans were among the most difficult implementation problems facing AHCCCS during its first 18 months. The decision made by the legislature to divide responsibility for eligibility and enrollment among the counties, the State, and the private Administrator created a number of operational difficulties once AHCCCS was implemented. Eligibility determination for categorically eligible groups (AFDC and SSI) was the responsibility of the Arizona Department of Economic Security and the Social Security Administration, whereas eligibility determination for MI/MN's was the responsibility of the counties. MSGI, and the AHCCCS Administrator, were responsible for enrollment—the process by which a person who has been determined eligible actually becomes a member of an AHCCCS plan. Even for the categorically eligible, for whom the eligibility and enrollment process was relatively straightforward, a number of problems arose.

The AHCCCS experience with eligibility and enrollment for MI/MN's illustrates problems inherent in trying to include these populations in prepaid, capitated programs. Because the application for eligibility will frequently be triggered by the utilization of services, care will often have been provided before capitation has begun. As a result, whenever MI/MN's are included in capitated programs, problems will arise regarding when the prepaid plans assume responsibility, and how services provided prior to plan responsibility will be covered.

Prior to AHCCCS, each county could establish its own eligibility requirements and procedures. Implementation of the AHCCCS program required that the counties develop, within a short time, a capacity to determine in a uniform manner the eligibility of MI and MN cases. The sheer volume of the cases (applications and redeterminations in the second year were estimated by county eligibility departments to be more than 85,000 per year), the complexity of procedures, and the confusion about guidelines have made eligibility determination a time-consuming and expensive county responsibility.

The responsibility for enrolling eligibles is separate from eligibility determination and, for the first 18 months of the program, was handled by the private Administrator. The Administrator's responsibilities were to conduct enrollment interviews, to assist eligibles in the enrollment process, and, if eligibles did not select a plan, to assign them to one. Delays in enrolling persons who had been determined to be eligible was a major problem during the first 18 months of the program. Prior to enrollment, providers were reimbursed on a capped fee-for-service basis for services to eligible persons. Delays in the eligibility and enrollment process during the first 18 months of AHCCCS contributed to large unanticipated fee-for-service expenditures.

Major changes were made in the eligibility and enrollment process as the program was implemented. In April 1983, legislation was passed that reduced the time that an MI/MN applicant was considered eligible for emergency AHCCCS service from 30 days prior to application to the 5 days prior to the date of determination. This legislation simplified the criteria for establishing MI/MN eligibility, thereby enabling the counties to process eligibility on a more timely basis. The old procedures were based on cumbersome tax code definitions and required extensive documentation; the legislation eliminated the dependence on the tax code. In May 1984, additional changes were made aimed primarily at reducing the fee-for-service payments—retroactive coverage of emergency services was reduced from 5 days to 1 day, and MI/MN eligibles were to be automatically assigned to a plan.

Overall, the problems in eligibility and enrollment have proven to be major source of difficulties in the implementation of AHCCCS. They have led to increased cost; diverted attention from other features of AHCCCS that needed implementation, such as encounter data and quality assurance procedures; and created tensions among counties, the private Administrator, the State, and providers. Changes in the enrollment process are now being made.

Bidding process

One of the most innovative features of AHCCCS is its reliance on bidding to secure providers. Contracts for the provision of services are awarded to health plans. Plans, in turn, have arrangements with individual physicians, hospitals, and other providers to furnish the necessary services. Plans are paid a capitation according to their winning bid rate. Plans bid an individual rate for each eligibility category in each service area (normally a county). Winning bidders are the lowest qualified bidders in each service area.

During the first year of bidding, AHCCCS permitted partial-service bids in addition to full-service bids. In the first year 50 separate organizations bid. There were 18 contracts awarded—17 of which were full-service bidders. What is surprising in the first year is the speed with which plans were able to organize themselves to bid. In the first year, 20 full-service bidders bid; 16 of them were formed specifically to bid on AHCCCS, and 13 of these 16 newly formed plans were awarded contracts.

In the first year, before winning bidders were selected in six counties, the State had to ask for voluntary price reductions. The State had initially expected to negotiate directly with the plans in counties where bids exceeded budget limits; however, the AHCCCS director was advised by the Attorney General that State law precluded negotiating the bid prices offered by providers. Following the request for voluntary price reductions, the remaining qualified bidders reduced their prices.

The bid rates for AFDC beneficiaries in the first year ranged from $43.46 by Coconino Health Care in Coconino County to $80.19 by Dynamic Health Services in Mohave County. For aged and MI/MN beneficiaries with Medicare, they ranged from $32.92 by Dynamic Health Services in Mohave County to $101.48 by Northern Arizona Family Health Plan in Yavapai County. For disabled persons, the lowest bidder was Pima Care ($105.92) in Pima County, and the highest bidder was Coconino Health Plan ($205.83) in Coconino County. The lowest bid for the blind eligibility group was $91.90 by Health Care Providers in Maricopa County, and the highest was $178.12 by Health Care Providers in Yavapai County. The overall average amount reimbursed per enrolled Federal beneficiary in the first year of the program was approximately $76.¹

During the second year of bidding, a few changes were made in the contracts. Dentures, transportation, prescription lenses, hearing aids, and dental services under the Early and Periodic Screening, Diagnosis, and Treatment Services (EPSDT) program were added as covered services; bid categories were modified; reporting requirements were clarified; a 1-year renewal clause was included; and the evaluation criteria more directly emphasized price and full-service bids. Voluntary price reductions were requested of all bidders in the second year because bid prices were too high. When that request did not result in low enough bids in 13 of the 15 counties, a second request for proposal (RFP) was released.

The lowest and highest bid rates in the second year for AFDC beneficiaries were from the same plan, ACCESS Patient's Choice: $46.97 in Maricopa County and $68.81 in Pima County. For aged and MI/MN beneficiaries with Medicare, bid rates ranged from $31.33 by CIGNA Health Plan in Maricopa County to $69.80 by Gila Medical Services in Gila County. For MI/MN non-Medicare beneficiaries, the bids ranged from $74.04 by Pinal General Hospital in Pinal County to $126.01 by Pima Health Plan in Pima County. For blind and disabled beneficiaries with Medicare, the lowest bidder was CIGNA Health Plan ($31.40) in Maricopa County, and the highest bidder was Pinal General Hospital ($105.39) in Pinal County. The lowest bid for blind and disabled beneficiaries without Medicare was $156.45 in Pima County by Pima Health Plan, and the highest was $303.80 in Pinal County by Health Care Providers. The overall average amount reimbursed per enrolled beneficiary in the second year of the program was approximately $81.

Of the 23 full-service bidders in the second year, 19 were awarded contracts. Although several plans contracting in the first year dropped out of the second-year bidding, more new plans bid, including one individual practice association (IPA) that bid statewide. The response by bidders in the first two rounds greatly exceeded expectations. However, it should be remembered that the State placed few restrictions on organizations qualified to bid and, as a result, bidding required only minimal startup costs. In addition, AHCCCS, MSGI, Arizona Hospital Association, and numerous private consultants were available to provide support and technical assistance in preparing bids.

Description of plans

The major features of the 19 main plans providing AHCCCS services in the second program year are given in Table 1, which shows that the plans vary widely along all dimensions. Two of the plans, Arizona Physicians IPA and ACCESS Patient's Choice, both of which are IPA's and provide services in almost all counties, together have almost one-half of the total enrollees. Maricopa Health Plan, a county-sponsored staff model plan, and Health Care Providers, an IPA, have the next largest numbers of enrollees. Together, these four providers have 70 percent of the enrollees.

Table 1. Characteristics of plans participating in the Arizona Health Care Cost-Containment System (AHCCCS) in the second year: October 1, 1983-September 30, 1984.

Plan	Major features				Payment mode
	Percent of enrollees1	Started for AHCCCS	Type	Sponsorship	Primary care physician	Specialist	Hospital
ACCESS Patient's Choice	9.9	Y	I	O	F	F	CH
Arizona Physicians IPA	38.2	Y	I	P	C,F	F	CH
CIGNA	1.4	N	S	O	S	S	D
Comprehensive AHCCCS	1.2	Y	I	P	C,F	F	CAP
Dynamic Health Services	1.3	Y	L	P	S	F	CH
El Rio Santa Cruz	4.0	N	S	O	S	S,F	D
Family Health Plan of Northeast Arizona	0.3	Y	I	H	C	C	CH
Gila Medical Services	1.4	Y	S	P	S	F	CAP,CH
Graham County Doctors Plan	0.7	Y	G	P	C	F	CAP
Health Care Providers	10.5	Y	I	P	C	C	CH
Maricopa Health Plan	11.9	N	S	C	S	F	CH
Northern Arizona Family Health Plan	2.2	Y	I	P	C,F	F	CAP
Phoenix Health Plan	0.6	Y	I	H	C,F	F	CAP
Pima Health Plan	7.0	N	G	C	S	S	D
Pinal General Hospital	2.2	Y	S	C	S	S	CH
Samaritan Health Services (Medical Care Systems)	1.0	Y	I	P	C,F	C,F	CH
St. Joseph's Mercy Care Plan	1.4	Y	S	H	S	2F	CH
University Famli-Care	2.9	N	S	O	S,C	C,F	D
Western Sun	1.9	N	S,I	O	S,F	S,F	CH

NOTES: Started for AHCCCS: Y = Started just for AHCCCS; N = Was in existence before AHCCCS. Type: G = Group model; I = Individual practice association; S = Staff model; L - A loosely organized configuration of physicians and hospitals. Sponsorship: P = Physicians; H = Hospitals; C = County; O = Others. Primary care physician payment mode: S - Salary; F = Fee for service; C = Capitation payment. Specialist payment mode: S = Salary; F = Fee for service; C = Capitation. Hospital payment mode: CH = Charges; CAP = Capitation; D = Discounted charges.

¹These enrollees represent those Federal categorical (AFDC and SSI) or MI/MN eligibles served through a prepaid capitated contract with AHCCCS plans as of May 1984.

(See the Technical note for a detailed description of AHCCCS contracts.)

²Negotiated discount of fee for service.

Table 1 indicates that 71 percent of the enrollees are in plans that were specifically started for AHCCCS. Of the enrollees, 64 percent are in the eight AHCCCS IPA-type plans, 26 percent in the seven staff model plans, 8 percent in the two group staff model plans, and 3 percent in two other types of plans. More than one-half of the enrollees (56 percent) are in the eight plans sponsored by physicians; 21 percent are in the three county-sponsored plans; and 22 percent are in eight plans sponsored by hospitals or having other sponsorship.

With respect to primary care physicians' mode of payment, the majority of the enrollees (55 percent) are in the eight plans that pay their primary care physicians on capitation or some combination of capitation and fee for service. Eight other plans (having 31 percent of the enrollees) pay their primary care physicians a salary, and one plan (with 10 percent of the enrollees) pays by fee for service. The remaining two plans (serving 5 percent of the enrollees) pay physicians through a combination of salary and capitation or salary and fee for service.

Specialists are normally paid on a fee-for-service basis. Fee-for-service specialist payment is used by 10 plans (having 69 percent of the enrollees). Three plans (with 11 percent of the enrollees) pay their specialists a salary, and two plans (with 11 percent of the enrollees) on capitation. Four of the plans (serving 10 percent of the enrollees) pay specialists a combination of capitation and fee for service of salary and fee for service.

Ten plans pay hospitals the billed charges; these plans have almost 80 percent of the enrollees. Four of the plans (having 15 percent of the enrollees) pay hospitals discounted charges, four plans (serving 5 percent of the enrollees) pay capitation, and one plan (with 1 percent of the enrollees) pays a combination of capitation and charges.

Gatekeeping and copayment

Visits to the plans in the first year revealed that the gatekeeping concept had been implemented and was operational, but the collection of beneficiary copayments had been actively implemented by only 12 plans.

The original AHCCCS legislation required that all eligibles choose or be assigned to a primary care physician within a plan who would serve as a gatekeeper for all care received. The 19 plans participating in the second program year have implemented these gatekeeper functions in a variety of ways. IPA-type plans generally have patients assigned to an individual gatekeeper; staff and group model plans and plans sponsored by hospitals frequently do not require a beneficiary to pick an individual gatekeeper. Generally, for all plans, internists and obstetricians/gynecologists can serve as primary care gatekeepers if they wish. For some of the plans in the rural areas, general surgeons and other specialists who have traditionally acted as primary care physicians in those areas can also serve as gatekeepers.

Of the seven IPA-type plans, all but two require preauthorization for referrals within the plan. Seven of the 12 plans that are not IPA's require preauthorization for such referrals. All of the plans indicated that they had a prior-approval process for hospital admissions.

Four of the seven IPA-type plans indicated that their doctors actively seek copayments; however, verification of copayment collection by the plan is more difficult in IPA-type plans (where each physician's office would have to be visited) than in plans that are more centrally organized. The one plan that is both an IPA and a staff model plan has a mechanism to collect copayments. On the non-IPA-type plans, four indicated that they do not attempt to collect copayments, and seven indicated that they do attempt to collect copayments. Many plan administrators feel that the collection of copayments is not worth the effort required. They generally did not seem to feel that copayments serve an important function in promoting a sense of personal dignity or in curtailing unnecessary utilization of services. There were, however, a few rural plans that voiced support for copayments because they believe that their AHCCCS beneficiaries take pride in paying for part of their care. None of the plans see the collection of copayments as an important revenue-generating tool.

Financial status

Lack of data reporting by the plans makes an objective evaluation of the financial status of the plans impossible at the time of this report. Peat, Marwick, Mitchell and Co. signed a contract in April 1984 to conduct a financial review of the plans from May to August 1984.

Financial and utilization data

As part of their first-year contract, AHCCCS plans were required to submit financial reports and utilization data of defined types to the AHCCCS program. Because of the initial problems with the program, in the first 18 months little emphasis was given to monitoring the collection of either of these kinds of data. Based on estimates from the AHCCCS Division on the number of encounters expected, and on our analysis of the approved encounters in the Administrator's data file as of February 1984, we estimate that less than one-fourth of expected encounter data had been collected. Many plans had submitted no data at all, and others had submitted only a small fraction of that expected.

Most of the financial data required from the plans for the first year also were not submitted, and those data that were submitted were not audited by AHCCCS. Based on a review of 3 of the 10 major reports required of most plans during the first year, we estimate that, by April 1984, fewer than 20 percent of these first-year reports had been submitted.

With the takeover of Administrator functions by the State, the appointment of a new AHCCCS director, and increased pressure from HCFA to collect such data, considerable attention is now being given to obtaining accurate and complete data from the plans. Such data is necessary to administer the program effectively and evaluate its outcome.

Financial support

Financial support is received from three major sources: the counties, the Federal Government, and the State. Counties contribute a fixed amount per year based on each county's expenditures for indigent health care before AHCCCS. The Federal contribution is based on the number of AFDC and SSI enrollees in AHCCCS. The State of Arizona provides the balance of the funding. In the first 9 months of operation (October 1982 to June 1983), county contributions made up 47 percent, Federal funds 32 percent and State appropriations 20 percent of the total $108.5 million. Adding projected revenues for the period July 1983 to June 1984, it was estimated that the Federal share would be 29 percent, the counties' 37 percent and the State's 34 percent. It should be remembered that the Federal Government provides matching funds only for the categorical eligibles, for whom it pays approximately 60 percent of the cost.

Actual program expenditures in the first 9 months of the program were $91.7 million. Adding projected expenditures as of February 1984, for the period July 1983 to June 1984, to the actual program expenditures for the first 9 months brought estimated program costs to $307 million for the first 21 months of operation. During the second year, projected expenditures exceeded budget revenues by $40 million because of significantly greater amounts paid for fee-for-service claims and higher-than-budgeted administrative expenses. Because the State of Arizona was at risk, supplemental appropriations had to be provided by the State legislature.

An important element in the evaluation relates to whether the program actually saves money. To conduct such an evaluation it is necessary to know both paid and incurred program costs. Because of the lack of data from the plans and the administrative problems with the program discussed above, estimates of incurred program costs during the first year must be made with caution. Possible incurred but not paid program costs include:

• Fee-for-service claims payments for claims incurred but not paid.

• Capitation payments to the plans for enrolled beneficiaries who were not known to be enrolled at the time of the capitation payment to the plan.

• Claims by the plans for services received by beneficiaries that exceed $20,000 during the year (AHCCCS provided reinsurance for 100 percent in the first year and 90 percent in the second year of the costs in excess of $20,000 per person per year).

• Payments by AHCCCS for Part B Medicare premiums and for crippled children's services.

Of special importance is an estimate of fee-for-service claims payments for claims incurred but not paid. Because of the administrative difficulties of the program in its first months and the fact that there is no program limitation on the time for submitting or resubmitting claims, this amount may be substantial. At the time of this report, sufficiently detailed information has not been available about these liabilities to make an accurate estimate.

Other issues

A number of other issues were important in the first year of implementation. One important issue was the county involvement including the impact of AHCCCS on long-term care and on county delivery systems. Another was the quality assurance and patient satisfaction procedures at the plan level. Another issue was the impact of AHCCCS on the private sector in Arizona.

County involvement

Although AHCCCS was intended to relieve the counties of responsibility for providing health care to indigents, the counties continue to have certain obligations. The decision to exclude long-term care from AHCCCS produced an ongoing obligation for counties to pay for long-term care (exclusive of acute medical services). The AHCCCS legislation also locked counties into the eligibility standards and coverage of services that were in effect on January 1, 1981, so that no one would be disenfranchised as a result of the AHCCCS program. This provision, called the maintenance-of-effort requirement, created county residual-care responsibilities for populations served and for services provided by prior county systems.

Although long-term care services are excluded from AHCCCS, the cost of acute medical services provided to AHCCCS-eligible residents of long-term care facilities is paid for by the program. Persons residing in long-term care facilities are treated as a special population in the AHCCCS bidding process. In counties for which no contracts are issued, services are provided on a capped fee-for-service basis. In the first and second years of AHCCCS, five plans were awarded contracts to provide AHCCCS-covered services to long-term care patients.

Long-term care was specifically excluded from AHCCCS services. The dual county and AHCCCS system that has evolved as a result has fragmented responsibilities and has created the potential for patient dumping between the counties and the AHCCCS providers. Policies and procedures in providing acute-care services to those in long-term care have been pieced together in an attempt to accommodate the existing county long-term care systems; however, fragmentation is a serious problem. Several efforts are under way in Arizona to find ways to address the current problems in the long-term care system, including the establishment of the Pritzlaff Commission by the Governor and the legislature (but funded by the Flinn Foundation), which is developing recommendations for a comprehensive approach to long-term care.

Prior to AHCCCS, the counties varied greatly in eligibility criteria, covered services, and delivery systems. Although most county responsibilities for indigent care (other than long-term care) were absorbed by AHCCCS, legislated responsibilities for residual care required them to continue serving people eligible under the old county systems but ineligible under AHCCCS, and to continue providing previously covered county services not included in AHCCCS. Some counties serve poor people who are ineligible for AHCCCS, not covered under county residual-care responsibilities, not covered by private insurance, and are unable to pay for the cost of medical care.

Of the 15 counties, 4 have no residual-care responsibilities other than for long-term care, 5 have residual-care responsibilities to population groups not served by AHCCCS, and 10 have residual-care responsibilities for services (in addition to long-term care) not covered by AHCCCS.

Before AHCCCS, five of the counties operated their own hospitals. These five counties had more difficulty in adjusting their delivery systems to AHCCCS than did the nine counties (Arizona's 15th county, La Paz, was not incorporated until after the implementation of AHCCCS) that did not operate hospitals, because the five with hospitals had to be concerned about keeping their hospitals viable, while the other countries could simply reduce their purchase of services.

Quality assurance and patient satisfaction

The AHCCCS plans are required to establish quality assurance and patient satisfaction procedures, and the Administrator and AHCCCS Division are responsible for monitoring those procedures. These two issues did not receive priority attention during the initial months of the implementation of AHCCCS at either the plan or Administrator level.

During 1983, MSGI contracted with the Accreditation Association for Ambulatory Health Care (AAAHC) to conduct quality assurance reviews of all the AHCCCS plans. The AAAHC review found that the quality of medical care provided to AHCCCS patients “… appears to be at least equivalent to the care rendered by AHCCCS providers to their private non-AHCCCS patients” (Moen, 1983). In addition, AAAHC found that most plans lacked quality assurance programs. Only five plans had quality assurance systems in operation, and only two plans, with less than 6 percent of the enrollees, had fully organized systems in place. For most plans, quality assurance activities were not a high priority during the first year.

These are several ways to assess patient satisfaction. These include formal grievance procedures, informal complaint mechanisms, and patient surveys. Formal procedures were established by AHCCCS for addressing patient complaints and grievances. A specific set of rules specify how plans must address these complaints and grievances. At the time of enrollment, each member must be given material explaining grievance procedures available through the plan, the Administrator, and the Arizona Department of Health Services. According to the HCFA regional office State assessment report dated January 31, 1983, in some plans, handbooks describing grievance procedures had not yet been distributed to members. In any event, there were few grievances filed. Between October 1982 and September 1983, the Administrator had received 78 formal grievances, 10 of which were filed by plan members and 68 by providers. Between October 1983 and August 1984, the Administrator received 101 formal grievances, 31 filed by plan members and 70 filed by providers.

In addition to the formal grievance procedures AHCCCS established an Ombudsman Office to handle informal complaints, and MSGI handled informal complaints in its member relations group. Finally, AHCCCS commissioned a study of patient satisfaction in Maricopa County in November 1983, which found that 76 percent of AHCCCS enrollees were satisfied with the health services they received from AHCCCS. However, 71 percent reported that they did not know how to report a complaint about medical care provided them.

Private sector impact

The original design of AHCCCS made State and county government employees and employees of private businesses eligible for AHCCCS. The legislature viewed enrollment of these groups as a way for AHCCCS to become a program of mainstream medical care delivery rather than simply an indigent health care program. AHCCCS has not enrolled public and private employer groups. On December 24, 1982, the Arizona Department of Health Services issued an RFP requesting providers to submit bids to provide services to public and private groups, but the bid rates, when combined with the 18 percent administrative cost of MSGI, were considered too high. The limited term of the demonstration and the problems associated with implementation have also made employers reluctant to join AHCCCS.

An important impact of the AHCCCS program has been to stimulate the formation of prepaid plans that will market to private employers. A number of the plans in AHCCCS (under different corporate entities) have applied for status as State-licensed health care service organizations, which will permit them to market to employers. The four largest corporations in Arizona have formed an employer's coalition (the Coalition for Cost-Effective Quality Health Care) that is actively promoting cost-containment strategies. The coalition's long-term goal is to increase choice among health providers and encourage the development of prepaid health options. In this regard, business leaders of the coalition view AHCCCS as a positive step toward a more competitive health care system in Arizona because it has encouraged prepaid health care providers to be formed.

Major findings

Given the short timeframe for implementation, the establishment of the AHCCCS program itself was an accomplishment. More people are eligible for AHCCCS than were eligible under the county systems for indigent care. The program has been judged by the AAAHC to be giving the same level of medical services to its clients that is given to those receiving services on a fee-for-service basis. Providers have participated, and it has provided a stimulus for private sector interest in health care cost-containment.

However, the program's first 18 months were fraught with controversy. AHCCCS was implemented quickly, and the Adminstrator functions were contracted out without sufficient thought as to what was being contracted for or how to monitor performance. Eligibility and enrollment procedures created difficulties for the counties, the Administrator, and the State, because a larger number of fee-for-service claims had to be paid for those beneficiaries served between the time of their eligibility determination and their enrollment. Plans did not know who their members were; MSGI did not know who was enrolled; and the counties had difficulty making eligibility determinations in the allotted time. Plans bid, were awarded contracts, and are providing services to beneficiaries but the lack of outcome data makes it impossible to know whether the volume and nature of services being provided are reasonable and adequate. Emphasis on day-to-day problems caused little attention to be given to monitoring the collection of financial and utilization data and to the development of systems for quality assurance or patient grievance procedures. Private-sector involvement has been postponed in part because of the high bid and administrative costs of private participation in AHCCCS and in part because of the poor public perceptions of the program.

Following are the highlights of the first 18 months of the program.

The list of achievements includes:

• The program was able to attract a large number of full-service bidders—Seventeen of the 18 first-year winning bidders and all of the 19 second-year winning bidders were full-service bidders. Of the 19 current plans, 13 were formed just for AHCCCS. Thus, a significant number of new prepaid health plans now exist in the State, creating a more competive environment.

• Provider acceptance has not been a problem— Providers in all counties are participating in the AHCCCS program. Although the Arizona Medical Association took a neutral position on the program, it appears that enough primary care physicians and specialists are participating.

• More people under statewide uniform eligibility criteria are enrolled in the program than under the previous county system—An estimated 136,000 to 142,000 beneficiaries were enrolled in the county systems before AHCCCS. As of May 1984, 175,196 beneficiaries were enrolled in AHCCCS.

• Gatekeeping appears to be implemented and functioning—Plans either have a specific individual physician assigned as a patient's gatekeeper or they have implemented the concept more generally with a group of physicians or the entire plan serving as the gatekeeper. IPA's generally have a specifically assigned gatekeeper. Those without IPA structures are usually not assigned a particular physician as a gatekeeper.

• Plans have been stimulated to develop internal cost-containment mechanisms—Many of the plans have developed internal cost-effective mechanisms, such as prior approval for referrals and admissions and emergency clinics to provide checks on the indiscriminate use of medical services.

• AHCCCS has contributed to the development of prepaid plans that are now being marketed to the private sector—Of the 13 plans that were formed just for AHCCCS, 5 plans are planning to market (under other corporate entities) to the private sector in the future.

• The entire health care delivery system in Arizona has been affected by a more cost-sensitive health care environment—Private employer groups have been more involved in the issue of private-sector health care cost-containment, and publicity given to the program has increased consumers' awareness of health care costs.

The problems encountered were:

• Failure of the private Administrator implementation—The concept of the private Administrator could not be adequately tested in Arizona because of significant implementation problems. These implementation problems include: the selection of a contractor without the requisite expertise, the inadequate monitoring of the contractor, the newness of the AHCCCS program and its unique features, the shortness of time for program startup, and the underestimation of required resources. Whether as a result of some or all of these factors, the private Administrator was not a successfully AHCCCS innovation. A true test of this concept should be conducted in a manner that takes into consideration the difficulties encountered in Arizona.

• Eligibility/enrollment interaction—Major problems were encountered in the areas of eligibility determination and enrollment because of the extended length of time between the application for eligibility and the final processing of enrollment information, the number of parties required to complete the transactions, and the inadequate communication links between the parties. Because of these gaps in the process, there was a significant chance that the parties had different information regarding an individual's status, creating confusion as to who was responsible for providing and for reimbursing services.

• Collection of financial and utilization data—During the first year, AHCCCS paid little attention to the collection of financial and utilization data from the plans. Thus, little information on cost and utilization is currently available for program mangement, monitoring, and analysis. More specifically, lack of cost data has resulted in the inability to determine whether the AHCCCS program has cost less than a traditional Medicaid program. Lack of utilization data makes it difficult to determine service use by the poor to medical-care services in Arizona.

• Copayment has not been implemented by many plans—Even though mandatory nominal copayment is part of the AHCCCS legislation, it has not been implemented by many of the plans. Many feel that, because the amounts of money are so small, it does not make good business sense for the plan administrators or providers to actively enforce collection.

• There have been few quality assurance procedures or patient satisfaction activities—Few of the plans had quality assurance and patient satisfaction procedures in place during the first 18 months. The State undertook only limited activity in this area as well. However, the quality assurance review done by AAAHC for the State found that the quality of medical care provided to AHCCCS patients appeared to be at least equivalent to that provided to their private nonAHCCCS patients. In terms of patient satisfaction few grievances were filed.

• County health care systems may be adversely affected—Since the implementation of AHCCCS, operating a county hospital has become an increasingly large financial liability, and counties feel that their expenditures on their hospitals have grown as a result of the loss of patient base to private AHCCCS providers. This problem is exacerbated because the remaining patients are often indigents without public or private health coverage, whose bills are likely to become bad debts for the county.

Policy implications

Major policy implications can be identified from the experience of AHCCCS in its first 18 months. Some of them are specific to particular features of the program, and some are more general and relate to overall program design and managment. The specific ones listed first are most important for those interested in taking one or several particular features of the program and implementing them in other areas. The final section is relevant for those considering the design of an entire program.

Specific implications

Program administration

The major lessons from AHCCCS involve determining what functions are appropriate to contract out, and designing and administering the contract.

• A State needs to think carefully about what functions are suitable for contracting out to private firms and those that are more suitable to be performed by the State itself.

• The State should exercise caution in basing an award decision primarily on bid price and should explicity assess the likelihood of cost overruns given a low bid.

• Responsibilities delegated to a private Administrator should be clearly specified and be capable of being monitored and supported by the contracting agency.

• Contract modification procedures are required that can deal effectively with changing responsibilities.

• The contracting agency needs to devote sufficient resources to monitoring the private Administrator's performance.

Eligibility and enrollment

Major policy choices involve deciding who should have responsibility for eligibility and enrollment functions and determining how these function can be related in a manner that covers eligibles under capitation as quickly as possible.

• The eligibility and enrollment system should be integrated in a manner that reduces delays. This may require either that the State or the contractor operate the entire system or that an on-line communications system among all responsible parties be established to reduce delays due to inefficient transfer of data.

• Eligibility forms and procedures should be streamlined to reduce unnecessary delays in eligibility determination.

• In a capitated system of delivery of medical care, it is especially important to cover eligibles under the capitation arrangements as quickly as possible. This may be by streamlining the eligibility and enrollment system and by designing the system so that providers accept retroactive capitation.

• Special attention in the design of a eligibility and enrollment system must be given to MI/MN groups, who tend to be high users of medical services at the time of their initial program eligibility.

Competitive bidding

The design of the bidding process is critical to determine whether competition is effectively promoted. This process involves choices concerning both the time schedule and rules for bidding.

• Sufficient time should be allowed for the State to develop the provider RFP and for the plans to respond to the RFP.

• To promote a more competitive process, the State should: not allow its budget amount to be known prior to bidding; have the ability to negotiate with bidders; not allow other bidders' bid prices to be known while providers are submitting revised bids; and create an environment where bidders perceive that there is a real probability of losing.

Formation of prepaid plans

Formation of prepaid plans is influenced by the specific requirements established for bidders and the degree of technical assistance provided to help plans get established.

• Providers do seem willing to organize themselves into prepaid plans to provide service to indigents.

• State government should provide technical assistance to newly formed prepaid plans. This should include technical assistance in the development of plan management information systems, effective utilization control mechanisms, quality assurance programs, and sound financial systems; and it should promote the sharing of best practice information among providers.

• Specific requirements for bidders regarding working capital, adequate management capability, quality assurance, and reporting systems should be developed and enforced.

Enrollee cost sharing

Policy choices must be made concerning the importance of copayments as a means of maintaining personal dignity and as a cost control measure versus the difficulties in implementing copayments.

• If the State does not wish to enforce mandatory beneficiary copayments, such copayments should be made optional at the discretion of the plans.

• If the State believes that copayments should be mandatory, there should be strong efforts by the State to convince plans and providers of their importance (as personal dignity and cost-containment mechanisms), assistance in designing accounting systems that would aid in copayment collection, and strong penalties for noncompliance.

Gatekeeping

Policy choices concerning the design of gatekeeping need to take into account the differences among types of prepaid plans.

• IPA's and non-IPA-type plans may, because of their structures, tend to implement the concept of gatekeeping in different ways. If the State wishes to permit these different concepts of gatekeeping, requirements should be written to allow flexibility.

• Incentives may be required to reward primary care gatekeepers for their added responsibilities for patients.

• Plans should be encouraged to implement data systems that could help them monitor gatekeepers' performance.

Information systems

Policymakers must recognize that the design of information systems for a prepaid program is different from that of systems necessary for traditional fee-for-service programs.

• Careful consideration should be given to the design of an information system specifically tailored for a prepaid environment.

• States and prepaid plans should commit resources early in the planning process to ensure that information systems are in place.

• Incentives and sanctions need to be established that encourage prepaid plans to submit encounter and financial data on a regular basis.

• Useful comparative management reports should be provided to the plans on a timely basis to encourage submission of complete and accurate financial and utilization data.

Quality assurance and patient satisfaction

Policy choices concerning quality assurance and patient satisfaction involve the design of procedures at both the plan and State levels and the establishment of mechanisms for implementing those procedures.

• States and prepaid plans should commit resources in the early stages of program operations to ensure that quality assurance and patient grievance procedures are in place at the program and plan levels.

• Bidders should be required to have in place functioning quality assurance and patient grievance procedures before starting to provide services.

• States need to play a regulatory and enforcement role through regular medical quality audits and monitoring of plans to ensure that quality assurance systems and patient grievance systems are in place.

Relationship to the counties

Policy consideration should be given to impacts of a statewide prepaid program on existing county health systems.

• The potential impacts on county hospitals of an AHCCCS-type program shoud be anticipated and planned for in advance. Other means of providing health care for those not eligible for the prepaid program but otherwise uninsured should be considered.

• Long-term care services need to be integrated into an AHCCCS-type program, either by completely defining roles and responsibilities with a separate program that has responsibility for long-term care or by including these services directly in the program. Because of the relatively small number of plan enrollees who need long-term care, it would be difficult to include such services in the plan capitation rates. However, these services could be integrated into an AHCCCS-type system by having the long-term care facilities themselves bid directly for the provisions of long-term care services.

Private sector impact

Although such an arrangement is often difficult to implement, programs serving indigents ideally would include beneficiaries of all payment sources. This not only provides a cross section of beneficiaries and risks to a plan but also encourages the provision of mainstream medicine to the indigents.

If private-pay patients are to be brought into the program, priority must be given to their inclusion early; premiums developed for them must be competitive with other insurance available; and the program itself must project an image of viability to the private sector.

General implications

AHCCCS has attracted a sufficient number of providers to serve a large number of indigents in all parts of the State. For those who wish to implement an AHCCCS-type program, four lessons can be learned from the program's early implementation.

• It is important to allow adequate time to plan for all features of the program before going operational. Without such planning, a program of this type begins with serious administrative problems.

• There needs to be a clear definition of responsibilities within the program. Administrative responsibilities have to be clearly defined and their functioning assigned to responsible individuals who have sufficient time to handle their implementation.

• Processes need to be completely and carefully designed so that the process itself does not contribute to the problems.

• Early attention must be given to monitoring cost, utilization, quality, and patient satisfaction. Early attention to these functions is imperative to their effective implementation because their operation can be easily postponed when programs are faced, as they will be, with unanticipated, short-term problems. In the final analysis, however, cost, utilization, quality, and satisfaction are the considerations on which the worth of a system of this type will be based. Attention to documentation and monitoring of these systems requires early and constant commitment of human and financial resources. Without such a commitment, their implementation will drag into subsequent years of the program operation.

Technical note

Information on the plans and bidding process is discussed only for the regular AHCCCS plans. However, as of May 1984, there were 22,483 additional AHCCCS eligibles served through other kinds of AHCCCS contracts. AHCCCS-covered-services are paid through prepaid capitated contracts or on a capped fee-for-service basis for 3,649 eligibles for acute medical care services in long-term care facilities. There are five long-term care contractors— Maricopa Health Plan, Pima Health Plan, Pinal General Hospital, Western Sun Associates, and Dynamic Health Services. The Department of Economic Security receives a captitated fee, negotiated with AHCCCS, for 547 foster children. Plans are reimbursed on a capped fee-for-service basis for 1,039 Native Americans living on reservations who choose to receive their services from an AHCCCS provider. Finally, AHCCCS administers Federal payments to Indian health service facilities for Native American AHCCCS eligibles who choose to receive their health care in those facilities.