TRENDS IN DDS EXPENDITURES: IMPACT OF COST CONTAINMENT MEASURES. April 4, 2008 FINAL

Transcription

1 TRENDS IN DDS EXPENDITURES: IMPACT OF COST CONTAINMENT MEASURES FINAL April 4, 2008 Acumen, LLC 500 Airport Blvd., Suite 365 Burlingame, CA Report Submitted to the California Department of Developmental Services (DDS)

2 Preface The Department of Developmental Services contracted with Acumen LLC in September 2006 to examine the efficacy of cost containment measures implemented beginning in This report uses expenditure data from through to make that assessment. Regulatory, policy and operational changes implemented after June 30, 2006 are not considered within the scope of this report. Such changes include provider rate increases in July 2006, the increase in the minimum wage, the establishment of permanent cost-containment measures and the elimination of the sunset date of the Family Cost Participation Program. i

3 Executive Summary Between FY93-94 and FY05-06, the California Department of Developmental Services (DDS) saw expenditures on consumers living in integrated community settings increase significantly. Adjusting for inflation, expenditures captured through the Purchase of Services data (which excludes DDS department and regional center (RC) administrative costs) rose from $774 million in FY93-94 to $2.2 billion in FY Although caseloads were growing rapidly in this period, expenditure growth matched or exceeded caseload growth in every year except FY This report uses Purchase of Service (POS) expenditure data from FY93-94 to FY05-06, coupled with consumer information from the Client Master File and Client Development Evaluation Report, to better understand the trends in DDS expenditures over this time period. In particular, we assess the effects of cost containment strategies instituted primarily in FY To do so, we examine the trends in expenditures by service and consumer characteristics over time. In addition, we use forecasts of the utilization of and per capita expenditures on services under alternative scenarios to estimate the effects of these cost containment efforts on total expenditures. Essentially, these forecasts simulate the expenditures that would have occurred in the absence of cost containment strategies, assuming that the patterns of utilization by different consumer groups followed the trends that existed before cost containment. The difference between expenditures under these counterfactuals and the actual expenditures observed with the imposition of cost containment provides an estimate of the impact of the cost containment measures. Finally, we consider consumer characteristics that will impact future costs. There are five key findings: Rapid increases in per consumer expenditures were concentrated in the period from FY96-97 to FY During this time, DDS instituted a number of rate increases and other changes in response to outside pressures, including (1) the Coffelt settlement and the Olmstead decision, providing expanded support to move former DC residents into the community; (2) quality responses required after a federal audit of the Home and Community Based Services waiver; and (3) rate increases associated with increases in the state minimum wage. Per consumer expenditures have been essentially flat in real terms since the imposition of cost containment measures in FY There is some evidence of shifting to more expensive residential service types. Residential service use declined throughout the period as consumers moved to Supported Living Service. In particular, the use of alternative residential model (ARM) Community Care Facilities (CCFs) has declined with residents moving to Supported Living Services. At the same time, there has been increased use of higher service level CCFs (both across levels and within level 4) and specialized residential settings. The simulation analysis shows significant costs avoided through the rate freezes and other cost containment measures, although in some cases this was associated with lower utilization. The rate freezes did not change the rate of decline in ii

4 residential services, but the Family Cost Participation Program appears to have reduced the use of respite care. Utilization growth also slowed for day programs. Because of the cumulative effects of fewer new cases each year, the eligibility standardization appears to have saved DDS as much as $26 million. iii

5 Table of Contents Executive Summary... i 1. Introduction Rise in Per Capita Expenditures FY96-97 FY A Overall Rise in Per Capita Expenditures... 5 B Per Consumer Expenditure by Consumer Characteristics Cost Containment and Per Consumer Expenditures on Services by Type A Services Affected by Cost Containment B Trends in Per Consumer Expenditures on Residential Services C Trends in Per Consumer Expenditures on Other Services Affected by Cost Containment Estimated Cost Avoidance from Cost Containment Measures A Overview of the Estimation Approach B Estimated Effect of Rate Freezes C Estimated Effect of Eligibility Standardization D Summary Impact Across the Cost Containment Strategies Cost Issues to Address Going Forward A Concentration of Expenditures Among High-Cost Consumers B Aging of the Caseload Appendix: Data Sources iv

6 1. Introduction The California Department of Developmental Services (DDS) oversees the provision of services to nearly 220,000 individuals with developmental disabilities, including mental retardation, cerebral palsy, epilepsy, autism and other related conditions. As mandated by the Lanterman Developmental Disabilities Services Act of 1969, DDS is required to ensure that individuals with developmental disabilities receive the services they need to meet their needs and goals in the least restrictive setting possible. For the vast majority of DDS consumers, these services are provided through a network of 21 non-profit regional centers (RCs). The RCs use person-centered planning to develop individual program plans that identify the services to be provided for each individual consumer. Although DDS pays for services only when private insurance, schools or other generic service providers are not available, there is no financial eligibility or means-test for services, and no specified cap on the expenditures per individual. Among state programs, the developmental disabilities system is unique in that it is an entitlement to services for Californians. Excluding DDS and RC administrative costs, expenditures on consumers increased significantly between FY93-94 and FY Adjusting for inflation, expenditures associated with RC consumers rose from $774 million in FY93-94 to $2.2 billion in FY05-06, 1 as shown in Figure 1-1. Figure 1-1: Total POS Expenditures (Inflation Adjusted, 2005 Dollars) $2,500 $2,000 Total Expenditure in Millions $1,500 $1,000 $500 $ Fiscal Year Rising caseloads only partially account for the increase in expenditures over this period. Caseloads did grow significantly: between and , the number of consumers 1 All expenditures are inflation adjusted using the CPI to reflect real dollars in

7 served rose by 78 percent, from 166,694 to 207,624. However, expenditure growth matched or exceeded caseload growth in every year except FY As shown in Figure 1-2, expenditure growth was more than double the enrollment growth between FY97-98 and FY Figure 1-2: Growth Rates in Program Payments (Inflation Adjusted) and Enrollment 18% 16% 14% Percent Growth Rate (%) 12% 10% 8% 6% 7.1% 10.3% 4.1% 4% 2% Annual Growth in Enrollment Annual Growth in Expenditures 3.4% 0% Fiscal Year Equally striking, however, is the rapid decline in expenditure growth in recent years. By FY02-03, expenditure growth was slowing dramatically, and since FY03-04, it has tracked closely with caseload growth. A major factor in this change is a series of temporary and permanent cost containment measures, listed in Table 1-1. Rate freezes have been used since as temporary measures while DDS works on a consistent method for determining payment rates for RC vendors. The state has also stopped funding new programs, other than the Community Placement Plan program, with start-up funds unless there is an approved health and safety exemption. Additionally, Community Care Facility (CCF) vendors cannot change their service level without an approved health and safety exemption. Other cost containment measures are administrative: expanding the time allowed for assessment of new consumers, increasing the average RC case manager to consumer ratio, and standardizing the eligibility definition to conform more closely to federal standards. Finally, a program called the Family Cost Participation Program (FCPP) assesses a participation cost, based on income, to parents of children who receive specific RC services. 2

8 Table 1-1: DDS Purchase of Service Savings Measures Effective FY02-03 through FY05-06 Cost Containment Measures Effective Date Part of Analysis (Y/N) 1. Non-CCP Start-up Funding Freeze 7/1/02 N 2. Day Program, Work Activity Program, and 7/1/03 Y In Home Respite Rate Freeze 3. CCF Service Level Freeze 7/1/03 Y 4. Contracted Services Rate Freeze 7/1/03 Y 5. Habilitation Services Rate Freeze 7/1/03 N 6. SSI/SSP Pass Through Elimination N & Unallocated Reductions N 8. Change in Eligibility Definition 7/1/03 Y 9. Family Cost Participation Program 1/1/05 Y Despite the cut in expenditure growth, the California Legislative Analyst s Office and individual legislators have questioned the efficacy of these cost containment measures, given the continuing increases in expenditures and the fact that the temporary rate freezes sunset in 2008, and FCPP sunsets in In light of these concerns, this report considers four major questions regarding the trends in expenditures between FY93-94 and FY05-06: 1. What factors are associated with greater expenditure growth in the RC system? 2. How have costs shifted across expenditure categories? 3. Have cost containment efforts been effective? 4. Where are the risks for future expenditures? Among the cost containment measures, we focus on those items in Table 1-1 that we can directly link to specific services. In particular, we examine the role of the day program and inhome respite care rate freezes (#2); the community care facility (CCF) ARM service level freeze (#3); the contracted services rate freeze (#4); the change in eligibility definition (#8); and the Family Cost Participation Program (#9). To address the questions listed above, we analyze expenditures as captured in DDS s Purchase of Services (POS) data, which comes from the Uniform Fiscal System (UFS). 3 DDS department and regional centers operational costs are not included in these data. The POS data identify payments by month for each vendored service for each consumer, as well as payments for services to groups of consumers that are paid for under larger contracts (such as transportation services). All expenditures are adjusted for inflation using the Consumer Price Index (CPI) to reflect real dollars in FY Although service codes identify the type of service, the POS data does not provide information on the number of units of service or the rates. For this reason, for a given consumer, we cannot determine the exact amount of service provided at what rate. 2 These cost containment measures have now become permanent, however this report treats only operational policies put into place before June 30, 2006, and thus treats these measures as temporary. 3 For FY93-94 to FY02-03, the dataset was downloaded in 10/2006, while for FY03-04 to FY05-06, the dataset was downloaded in 1/

9 To link the POS data to information on consumers, we obtained consumer enrollment, demographics, and disability information from the Client Master File (CMF) and Client Development Evaluation Report (CDER). 4 Unless otherwise noted, our analysis includes all consumers who were enrolled in a Fiscal Year (FY) under status 1 (high risk 0-3 year olds) or status 2 (active consumers). We break the population into different consumer groups in order to account for variations in growth patterns within the overall population. Our consumer groups are defined by unique age range-diagnosis combinations, resulting in 13 groups. In addition, we separate out consumers who were ever residents of Developmental Centers in the period covered in this analysis. 5 This grouping allows us to track the distinctive characteristics of each group with respect to patterns in utilization, expenditures and population growth over time. Although consumers may have multiple developmental disabilities, our diagnosis groups are mutually exclusive and exhaustive. That is, consumers placed in the cerebral palsy group are only those who are not diagnosed with autism; consumers with diagnosed epilepsy are in the epilepsy group only if they do not have autism or cerebral palsy; a consumer with mental retardation will be placed in the mental retardation group if s/he does not have autism, cerebral palsy or epilepsy; Other Diagnosis is for consumers that do not have autism, cerebral palsy, epilepsy or mental retardation; 6 finally, consumers who do not have Client Development Evaluation Reports our source of diagnosis information are placed in the not in CDER group. The remainder of the report is organized into four chapters. Chapter 2 examines trends in per capita expenditure growth. Chapter 3 examines the role of cost containment efforts, while Chapter 4 presents findings from a simulation analysis that explores the likely impact of the cost containment. Finally, Chapter 5 identifies ongoing issues that need to be considered to control costs in the future. 4 The CDER data include only consumers aged 3 and over. 5 Determined according to whether the consumer was status 8 in the Client Master File between FY93-94 and FY Because we aim to determine the consumer s primary diagnosis (the diagnosis most debilitating), we categorize consumers with particular disabilities only when that disability causes, moderate or substantial impact on the consumer s functioning ( mild impact cases do not qualify.) This means that if a consumer has autism with a mild impact and epilepsy with a severe impact, s/he will be placed in the epilepsy group. 4

10 2. Rise in Per Capita Expenditures FY96-97 FY00-01 A Overall Rise in Per Capita Expenditures Although total expenditures for RC consumers rose from FY93-94 through FY05-06, average expenditures per consumer calculated simply as total expenditures divided by the number of consumers grew primarily in the five years between and (Figure 2-1). In , expenditures per consumer averaged $6,633, including $5,842 on average in services purchased for individual consumers and $791 in services purchased for groups of consumers. Before , expenditures per consumer grew no more than three percent per year. However, between and , annual growth rates ranged from 8 to 11 percent, with the largest jump between and Figure 2-1: Annual Expenditures per Consumer (Adjusted for Inflation) $14,000 Average Expenditure per Consumer by Fiscal Year Before Cost Containment After Cost Containment $12,000 $10,000 $833 $815 $623 $601 $765 $737 $691 $711 $8,000 $6,000 $4,000 $2,000 $791 $5,842 $761 $800 $775 $6,069 $6,036 $6,239 $802 $6,761 $841 $7,559 $798 $8,298 $9,103 $9,650 $9,926 $10,001 $10,013 $10,067 $ Consumer-based expenditure Contract-based expenditure Habilitation-based expenditure Much of the growth in expenditures in the late 1990s was the result of specific policy changes driven by influences outside of DDS. First, the Coffelt settlement (1994) and the Olmstead (1999) decision, as well as legislation, required DDS to place consumers in the least restrictive settings that is, move them out of developmental centers and to provide sufficient case management and other community services to make these moves feasible. Second, federal audits and compliance review identified quality issues in the provision of services paid through the Medicaid Home and Community Based Services waiver (HCBS). Third, increases in the state minimum wage during these years required rate increases in service categories that were highly dependent on minimum wage labor. Table 2-1 reports key budget augmentations between and , as documented through Budget Change Proposals and other budget sources. 5

11 Table 2-1: Budget Changes Affecting Per Capita RC Expenditures Prior to the Cost Containment Measures Budget Change Amount (millions) Description : Prior to Cost Containment, Period of Rising RC Expenditures FY 1997/1998 Community Care Facilities (CCF) $11.7 3% rate increase for community care facilities FY 1998/1999 Supported living wage increase $5.1 Salaries, wages, and benefits Department of Rehabilitation/Habilitation Services $16 Department of Rehabilitation services are waiver-eligible services Institutional deeming $1.7 Costs shifting from DHS to RCs due to admissions freeze for Medicaid Waiver program Self-Determination Pilot Projects $.07 Restoring Case Management Services $31.1 Restore the 1:62 case management ratio, increase salaries, fund key RC positions Training for CCF Staff $19.2 Staff training and wage increase after completion Community-Based Day Program and $27.4 In-Home Respite Services Agencies Rate System Restructure Community Care Facilities Rates $14.8 3% rate increase to CCFs In-Home Respite Workers Salary and $7.2 Reflects changes in minimum wage level Wage Increase Funds to Renegotiate Supported Living $5.1 Rates FY 1999/2000 Training for CCF Staff $56.1 Year 2 Salary Increase for CCF Staff $22.7 Increase rates of CCF by 10.3% FY 2000/2001 Shift Nursing Rate Increase $2.5 Adjustment to the fall Budget Change Proposal based on new information about in-home nursing rates : Passage of Cost Containment Measures, Period of Level Per Capita Expenditures FY 2002/2003 FY2004/2005 Implementation of Measures Listed in Table 1-1 Sources: Documentation provided by DDS, including Budget Change Proposals; Finance Letters and May Budget Revisions This rapid growth in expenditures per consumer was largely isolated in the period between FY96-97 and FY By , expenditures per consumer leveled off. Since , the year cost containment measures were enacted, expenditures have been nearly constant, except for the shift of habilitation services into the DDS budget. 6

12 B Per Consumer Expenditure by Consumer Characteristics Before we turn to the specific services that represent the increased per capita expenditure, we briefly review the changes in per capita expenditures by key consumer characteristics: age, diagnosis, and former DC residence. Per Consumer Expenditures: Former DC Residents As noted above, in the 1990s, DDS faced increased pressure for consumers to be served in the community with regional center support rather than residing in developmental centers. Although our analysis focuses on individuals living in integrated community settings, we are able to identify consumers who previously resided in developmental centers (DC) based on residence as recorded in the CMF. As seen in Figure 2-2, there was a jump in the number of former DC residents living in integrated community settings following the Coffelt settlement. Since this initial transition, additional individuals have moved from the DCs, albeit at a slower pace. Figure 2-2: Number of Former DC Residents among Consumers Served in the Community 3,500 3,000 2,500 Number of Consumers 2,000 1,500 1, Fiscal Year Although the movement out of developmental centers was concentrated early in the period, expenditures per former DC resident grew dramatically throughout this period. In FY93-94, an average of $2,147 per month was spent in POS expenditures on each former DC resident (Figure 2-3). 7 By FY05-06, this amount had increased to $4,707 per month, a 119 percent increase. 7 Note that Figure 2.4 and subsequent figures examine expenditures per month rather than expenditures per year. 7

13 Our analysis did not examine whether this growth was concentrated amongst those newly transitioning out of DCs or experienced by all former DC residents. Figure 2-3: Monthly Expenditures per Consumer among Former DC Residents $5,000 $4,707 $4,500 $4,000 Average Monthly Expenditure $3,500 $3,000 $2,500 $2,000 $1,500 $2,147 $1,000 $500 $ Fiscal Year Per Consumer Expenditures by Age Group and Diagnosis Although former DC residents experienced the greatest increase in per capita expenditures, they still account for only seven percent of overall expenditures, because they represent such a small share of the caseload (two percent in FY05-06). There were also other changes in the caseload, especially a rising number of children and a rising share diagnosed with autism. Between FY93-94 and FY03-04, the caseload under age 19 grew 153 percent, compared to a 59 percent growth among over 19. By FY05-06, individuals over 19 years old represented 45 percent of all RC consumers, down from 55 percent of consumers in FY The share of all POS expenditures accounted for by adults did not fall as rapidly as their share of the caseload, because adults experienced somewhat higher per consumer expenditure growth. As shown in Figure 2-4, monthly expenditures per consumer rose from $1,115 to $1,695 among adults between FY93-94 and FY05-06, a 52 percent increase in real terms. In comparison, expenditures per consumer grew 37 percent among children 3-19 years old and 25 percent among children under 3. In FY05-06, individuals over 19 years old represented 66 percent of expenditures, down from 75 percent of expenditures in FY Although expenditures per adult grew rapidly, virtually all of this growth occurred prior to FY Since then, growth has been minimal only two percent over three years. 8

14 Figure 2-4: Monthly Expenditures per Consumer, By Age Group (Inflation Adjusted) $1,800 $1,600 $1,400 Less than 3 yrs old Between 3 and 19 yrs old Over 19 yrs old $1,695 Average Monthly Expenditure $1,200 $1,000 $800 $600 $1,115 $637 $602 $822 $795 $400 $200 $ Fiscal Year Rising expenditures per consumer in part reflect changing needs of the consumer population. Over this time period, Early Start consumers and consumers with autism represented increasing shares of the consumer population, while consumers with mental retardation (but without autism, cerebral palsy or epilepsy) are a declining share. Figure 2-5 illustrates this change. Figure 2-5: Percent of Population by Diagnostic Group 60% Under 3 50% Autism 40% Cerebral Palsy, no Autism 30% Mental Retardation, no Autism, no Cerebral Palsy, no Epilepsy Not in CDER 20% Other Diagnosis, in CDER 10% Epilepsy, no Autism, no Cerebral Palsy 0% Fiscal Year 9

15 This changing diagnosis profile matters because there are consistent differences in average expenditures per consumer by diagnosis, shown in Figure 2-6. For consumers with mental retardation, the largest share of consumers, average expenditures are moderate, falling just above consumers with cerebral palsy. In FY93-94, consumers with autism had the highest expenditures per consumer. As this group has grown, the average expenditure per consumer with autism has fallen somewhat, reflecting the number of children in this group. Since FY00-01, consumers with epilepsy but no autism or cerebral palsy have accounted for the highest average expenditures per consumer. Other than the autism group, each diagnosis group showed increasing expenditures per consumer from FY95-96 through FY For those not in CDER, average expenditures per consumer dropped sharply after the FY03-04 policy changes. Early Start expenditures per consumer flattened after FY02-03 and then experienced a small uptick starting in FY05-06, but otherwise, per consumer expenditures tended to rise in recent years, albeit at slower rates than in earlier periods. Figure 2-6: Monthly Expenditures per Consumer by Diagnosis for Eligibility $1,800 $1,600 Average Monthly Expenditure $1,400 $1,200 $1,000 $800 $600 $400 $200 $0 Early Start Autism Cerebral Palsy, no Autism Mental Retardation, no Autism, no Cerebral Palsy, no Epilepsy No CDER Available Other Diagnosis, in CDER Epilepsy, no Autism, no Cerebral Palsy

16 3. Cost Containment and Per Consumer Expenditures on Services by Type After growing substantially in the late 1990s, per consumer expenditures have held steady since FY Effective July 1, 2003, a number of rate freezes and other cost containment strategies were implemented that affected specific service types. In this section, we review the services affected by these cost containment strategies and the trends in expenditures on these services over the analysis period. A Services Affected by Cost Containment We focus on services affected by cost containment measures, including: Residential services including CCF (savings measure #3) Day programs (savings measure #2), look-alike day programs and supported living (savings measure #4) In-home respite and respite related services (savings measures #2 and 9), and Transportation (savings measure #4). Although the range of services covered by the cost containment measures is fairly broad, there are some closely related services that are not directly covered by the measures. 8 If rate freezes have an impact on availability of services, one possible response is to substitute similar services not covered by rate freezes. The most important area to consider potential substitution is in the area of residential services. During the time period covered by this analysis, new specialized residential services (habilitation and health based) became available as options. Supported living services also became a more common selection for DDS consumers. B Trends in Per Consumer Expenditures on Residential Services Services grouped as residential services is the only category where we can directly compare average expenditures per consumer for cost contained and non-cost contained services. (In considering expenditures per consumer by service type, we count only consumers using these services.) Among CCFs, ICF, specialized and other residential services, only CCFs were subject to cost containment. As shown in Figure 3-1, average expenditures per consumer were level or falling for CCFs after FY Over the same period, per consumer expenditures grew for other residential types. (Note: service types defined as specialized residential health and habilitation did not exist before FY Similarly, while other residential services were available prior to FY96-97, we only report average expenditures per consumer for services 8 We do not estimate the cost avoidance brought by the habilitation services rate reduction because habilitation services (defined by service codes 950, 952, 954) were not financed by DDS until Thus, we do not have any expenditure information on these services until after the rate reduction was enacted and cannot project the rate reduction s impact. 11

17 received by more than 50 consumers per year. The other residential services category did not exceed 50 consumers until FY ) Figure 3-1: Average Monthly Expenditures per Consumer on Residential Services $8,000 $7,000 CCF-Cost contained ICF-Not cost contained Spec. Residential (Habilitation)-Not cost contained Spec. Residential (Health)-Not cost contained Other Residential-Not cost contained $6,000 Average Expenditures per Consumer $5,000 $4,000 $3,000 $2,000 $1,000 $ Fiscal Year When comparing the growth in other residential service expenditures following cost containment to the flat growth in CCF expenditures, we cannot simply assume that CCF expenditures would have risen the same amount in the absence of cost containment. First, CCF expenditures had already flattened prior to the saving measures imposed in FY Second, the rates in the other residential service categories may be driven up by increased demand for these services, offering higher reimbursements than CCFs. The flat expenditures in CCF are particularly striking because of the movement of consumers across types of CCFs. Within CCF ARMs, service levels are scaled from level 2 to level 4, and within level 4 from 4A to 4I, where level 2 offers care, supervision, and training for consumers with some self-care skills and no major behavior problems and level 4I offers the most supervision, targeted to consumers with more severe disability levels or behavioral 9 The Other Residential Services category includes family home agencies, housing services, crisis intervention facilities, out-of-state residential treatment and geriatric facilities. 12

18 challenges. The freeze on service level changes prevents a facility from switching which type of CCF it is, but does not prevent consumers from shifting among types of CCFs. The number of consumers served in ARM level 2 facilities has fallen every year since 1993, and the number served in level 4 facilities has increased in every year (Figure 3-2). Consumers in level 3 facilities grew slightly in the 1990s, but fell back a bit recently. Use of other residential services, including specialized residential facilities has also grown as these options became more available. Figure 3-2: Number of Consumers with Community Care Facility or Specialized Residential Facilities by Fiscal Year 14,000 12,000 10,000 8,000 6, ,000 2,000 - CCF2 CCF3 CCF4 ICF/SNF Other /Spec CCF1/unknow n Between FY03-04 and FY05-06, 1,675 consumers left level 2 CCF facilities. As Table 3-1 shows, the largest share of these received no residential services in FY05-06 (for example, moved into supported living). Of the remaining, the most common destination residential type was level 3 CCF; level 4s were relatively uncommon. Another one in seven were no longer active on the caseload in FY05-06; these would include consumers who moved, died, or no longer received services from the RC system. At the same time, the number of consumers served in level 4 facilities increased for all facility types, although the most dramatic increase occurred in level 4I facilities, where the number of consumers in these residence types increased from 275 in FY93-94 to 3,145 in FY05-06 (Figure 3-3). Between FY03-04 and FY05-06, 950 consumers entered a level 4I CCF. Of these, 579 were either not active consumers in FY03-04 (i.e. new entrants) or were not in residential care in FY Another 310 resided in lower level CCFs in FY Only five moved from ICFs or SNFs and 56 from other residential facilities. Among consumers 13

19 Table 3-1: Residential Type in For Consumers who Left Level 2 CCF Facilities Between and Residential Type in Number Percent No Longer on DDS Caseload % CCF % CCF4A 14 1% CCF4B 10 1% CCF4C 56 3% CCF4D 19 1% CCF4E 6 0% CCF4F 12 1% CCF4G 18 1% CCF4H 5 0% CCF4I 37 2% ICF/SNFs 7 0% Other Residential incl Specialized 79 5% No Residential Services % All Exits % Figure 3-3: Number of Consumers with Community Care Facility Level 4 Expenditures by Fiscal Year 3,500 3,000 2,500 2,000 1, ,448 3,145 1, CCF4A CCF4B CCF4C CCF4D CCF4E CCF4F CCF4G CCF4H CCF4I Number of Consumers 14

20 entering CCF4s, those entering level 4I were even more likely to be newly entering residential care than those entering lower levels within the CCF4s. For example, 38 percent of those entering level 4A were new to residential care compared to 48 percent in level 4I. Another 38 percent of those entering level 4A came from lower level CCFs, compared to 32 percent entering level 4I. Taken together, it appears that per consumer expenditures on CCFs have held essentially steady, even though there has been a shift out of lower level CCFs (especially into nonresidential services such as supported living) as well as a shift into CCF level 4 s, with the greatest growth in the most expensive level 4I. We explore the effect of the controls on CCF in greater detail in Chapter 4 below. C Trends in Per Consumer Expenditures on Other Services Affected by Cost Containment We next consider together day programs, look-alike day programs and supported living, which only became a commonly used option around FY All three of these service types were subject to rate freezes. For look-alike day programs, the rate freeze appears to have dampened the relatively rapid rate of increase in per consumer expenditures from FY00-01 to FY03-04, while per consumer expenditures in real dollars fell for day programs (Figure 3-4). Figure 3-4: Average Monthly Expenditures per Consumer on Day Programs and Look- Alike Day Programs $1,200 $1,000 $800 $600 $400 Day Programs-Rate freeze and DDS established Look-Alike Day-Rate freeze and negotiated $200 $

21 Over the same period, use of supported living grew rapidly, rising from 1,576 consumers in FY97-98 to 7,350 consumers in FY As the number of consumers using these services rose, the average expenditures per consumer also rose, even with the rate freeze, in part due to the rising popularity of this option for consumers with greater need for support (Figure 3-5). Figure 3-5: Average Monthly Expenditures per Consumer on Supported Living Services $3,500 $3,000 $2,500 $2,000 $1,500 $1,000 $500 $ As with day programs, we cannot compare respite programs subject to cost containment with those not subject to cost containment, since most programs in these service types were affected either by rate freezes or by other cost containment measures, such as requiring families to pay a share of costs to participate. Figure 3-6 shows the trends in these two types of respite programs. At the time the cost containment measures were imposed, per consumer expenditures on in-home respite care were already falling, while expenditures on other respite services had grown substantially. Despite the rate freeze, per consumer expenditures continued to rise after FY03-04, indicating greater intensity of use per consumer. 16

22 Figure 3-6: Average Monthly Expenditures per Consumer on Respite and Transportation $700 $600 In-Home Respite - Rate Freeze Other Respite and Related Services - Cost Contained Transportation - Rate Freeze and Negotiated $500 $400 $300 $200 $100 $ Fiscal Year For simplicity, we also include transportation on this graph as a baseline for the final service category we examine in our cost containment analysis. As this shows, transportation costs per consumer were rising since FY97-98, although these costs are lower than costs in other service types. 17

23 4. Estimated Cost Avoidance from Cost Containment Measures The findings in Chapter 3 suggest that the cost containment efforts have been successful in holding down expenditures in CCFs, day programs and look-alike day programs, in-home respite care and transportation. They also suggest that cost containment has been less successful for supported living services and, to some degree, other respite care. However, as we saw with residential services, caseload changes across different service types and the substitution of one service for another make it somewhat difficult to determine the impact of cost containment. In this chapter, we take a different approach to measuring the impact of cost containment. A Overview of the Estimation Approach To determine the impact of the cost containment, we estimate what costs would have been in the absence of the change and compare it to the actual costs observed. This difference provides a measure of the costs avoided by the implementation of cost containment. Our projections of the costs in the absence of cost containment are called counterfactuals because they are not directly observed in the data. For example, in the case of a rate freeze imposed in FY03-04, we project what expenditures and utilization from FY03-04 to FY07-08 would have been for each service category if none of the payment rates were frozen in FY Our counterfactual estimates are based on trends observed in prior years. Because we cannot be certain that a growth rate will persist, we determine three counterfactuals high, medium and low for each consumer group in each fiscal year. The high and low values can be interpreted as the range and the medium value as the preferred estimate. In building the counterfactuals, we attempt to account for both direct and indirect effects of the cost containment. Cost containment can affect growth in total expenditures by preventing an increase in expenditures per consumer. This would be considered the direct effect of the measure. However, cost containment can also have indirect effects. It can reduce the utilization of services because the supply may fall if reimbursements do not keep up with costs in the absence of rate increases. Further, cost containment can indirectly affect the use of other comparable services, as consumers shift to other options. The magnitude of these indirect effects will depend on how well reimbursements cover the costs of services, the sensitivity of the supply to changes in reimbursement and the availability of comparable, non-cost contained services. Finally, our counterfactuals allow for differences in utilization and expenditures by consumer characteristics, including age and diagnosis. For example, when we consider what the utilization would have been for in-home respite care, we consider to what degree these services are used for consumers of different ages and diagnoses, and then account for the changing size of these groups (such as growing numbers of children) in the population. A more detailed methodology is available in the appendix. B Estimated Effect of Rate Freezes and Related Policy Changes For each of the rate freezes included in the cost containment measures, we examine the trends in the utilization of services that is, the share of consumers using these services and the trends in the total expenditures. To address the spillover effects of cost containment in one 18

24 service on the use of other, related services, we group services together. This has the disadvantage of preventing separate calculations of the effect on each piece. However, it allows us to measure a broader impact than looking at an individual service in isolation. Day Programs, Look-Alike Day Programs and Supported Living Services Starting with day, look-alike day and supported living programs, Figure 4-1 shows that utilization after cost containment was lower than we would have predicted from earlier trends. Between FY93-94 and FY02-03, the percentage of consumers using these services increased from 30 percent to 38 percent. Soon after the rate freeze, the share of consumers using these services was between 38 percent and 39 percent. Based on our counterfactual calculations, if the rate freeze was not implemented, we would have expected the percentage of consumers using these services to increase to 42 percent. This is shown by the higher line (in green) which shows the trend under the counterfactual, compared to the actual. Utilization did not fall, but it did not rise as much as we would have expected. Figure 4-1: Utilization of Day Programs, Look-Alike Day Programs and Supported Living Services: Actual/Projected and Counterfactual Rates over Time 60% Percentage of Consumers Using the Services 50% 40% 30% 20% 10% Counterfactual Actual and Projected 0% Fiscal Year Once we have a counterfactual for utilization (accounting for age groups and diagnoses), we consider the trends in expenditures on these services per consumer by consumer group. Using different time periods to consider the past trends, we develop a high, medium and low estimate of expenditures under alternative counterfactuals. Without cost containment, we would have 19

25 expected total expenditures on these programs to be somewhere between the high counterfactual and the low counterfactual. Figure 4-2: Total Expenditures for Day Programs, Look-Alike Day Programs and Supported Living: Actual/Projected and Counterfactual Values over Time $1,600 $1,400 Total Expenditures (Millions) $1,200 $1,000 $800 $600 Low Counterfactual Medium Counterfactual High Counterfactual Actual and Projected $400 $200 $ Fiscal Year The results from the cost avoidance analysis, shown in Figure 4-2, suggests that after the rate freeze in FY03-04, the growth of total expenditures in day program, look-alike day program and supported living slowed. All three counterfactuals are higher than the actual, suggesting that the rate freeze resulted in cost avoidance. Since our counterfactuals show greater increase in expenditures over time, the estimate of costs avoided grows over time. In FY04-05, the medium cost avoidance was estimated to be $85 million followed by $133 million in FY Forecasting these trends forward and accounting for the changing caseload, we estimated that the rate freeze will result in cost avoidance of about $250 million in FY Respite After cost containment measures, the utilization of in-home respite and related services actually fell, as shown in Figure 4-3. The fall was minimal in the first years of the cost containment, from 27 percent in FY03-04 to 25 percent in FY04-05, but it accelerated after This acceleration coincides with the timing of the implementation of the Family Cost Participation Program, although other changes may also be driving the decline in utilization. We estimate that these rates will drop further to 15% in FY Since the utilization measures the 20

26 share of consumers using any service at any point in time during the fiscal year, the fall in utilization suggests that some subset of families preferred going without the respite services rather than paying a share of the costs. This need not imply that families dropped these services; in a growing caseload, some new families may have chosen not to pick up these services. By considering all respite services together, this analysis already accounts for potential shifts across types of respite services. As the counterfactual line suggests, we would have expected modest increases in utilization if prior utilization trends had continued. Had the earlier trend persisted, about 34% of consumers would have used these services. Figure 4-3: Utilization of In-Home Respite and Related Services: Actual/Projected and Counterfactual 100% 90% Percentage of Consumers Using the Services 80% 70% 60% 50% 40% 30% 20% 10% Counterfactual Actual and Projected 0% Fiscal Year When we examine average monthly expenditures per consumer using respite services (Figure 4-4), we see an initial dip in the average monthly expenditures per consumer following the rate freeze, and then an increase after the cost participation requirement. At the same time, we find that the average number of months of respite uptake per year, among those who uptake respite services at all, fell during the post-rate-freeze era (from over 8.5 months in FY03-04 to about 7 months in FY07-08). One possible explanation for the rise in average monthly expenditures is that those who continued to use respite services in the post-rate-freeze era were relatively high cost users during their months of respite utilization compared to those who stopped using these services. Another possibility is that those who continued to uptake respite services increased their monthly respite spending. Consistent with the fall in utilization, the total expenditures on respite services fell, as shown in Figure 4-5. Compared to the medium counterfactual, the fall suggests cost avoidance 21

27 of about $26 million in FY03-04, rising to $95 million in FY Even if we shift to the lower estimate (the low counterfactual), we would expect costs to have been $7 million higher in FY03-04 and $76 million higher in FY05-06, with higher cost avoidance in later years. Figure 4-4: Average Monthly Expenditures per Consumer Using In-Home Respite and Related Services: Actual/Projected and Counterfactual $500 $450 $400 $350 Low Counterfactual Medium Counterfactual High Counterfactual Actual and Projected $300 $250 $200 $150 $100 $50 $ Fiscal Year Figure 4-5: Total Expenditures on In-Home Respite and Related Services: Actual/Projected and Counterfactual $400 Total Expenditures (Millions) $350 $300 $250 $200 $150 $100 Low Counterfactual Medium Counterfactual High Counterfactual Actual and Projected $50 $ Fiscal Year 22

28 Residential Services As we saw in Chapter 3, use of residential services declined throughout the analysis period. The declining trend in the use of these services continued after cost containment, accounting for shifts across the services included in this set. There is no evidence that cost containment changed the rate of decline. Given that many of the new entrants to the caseload were children, some of this decline may speak to a declining need for residential services as a share of the caseload overall. Figure 4-6: Utilization of Residential Services: Actual/Projected and Counterfactual 20% 18% Percentage of Consumers Using the Service 16% 14% 12% 10% 8% 6% 4% Counterfactual Actual and Projected 2% 0% Fiscal Year The cost containment efforts do appear to have held down costs of residential services, even though the trend in the percentage of consumers using the residential service did not change. Here the low counterfactual largely captures the lower growth in residential service expenditures following the suspension of cost statement updates for CCFs in FY In fact, this counterfactual suggests little or no additional cost avoidance from the service level freezes in FY

29 Figure 4-7: Total Expenditures on Residential Services: Actual/Projected and Counterfactuals $1,000 $900 Total Expenditures (Millions) $800 $700 $600 $500 $400 $300 Low Counterfactual Medium Counterfactual High Counterfactual Actual and Projected $200 $100 $ Fiscal Year Transportation Services Like residential services, the rate freezes had only modest impact on the overall expenditures on transportation. In the few years prior to the imposition of the rate freeze, expenditures on transportation were fairly flat in real terms, as shown on Figure 4-8. Expenditures had been rising more in earlier time periods. If we base our counterfactual on this longer period, as in the medium counterfactual, we do see cost avoidance as a result of the rate freezes, equal to about $12 million in FY Because transportation services are purchased through contracts that cover multiple consumers, we cannot distinguish the effects of utilization from the overall change in expenditures for these services. 24

30 Figure 4-8: Total Expenditures on Transportation: Actual/Projected and Counterfactual $250 Total Expenditures (Millions) $200 $150 $100 Low Counterfactual Medium Counterfactual High Counterfactual Actual and Projected $50 $ Fiscal Year C Estimated Effect of Eligibility Standardization Breaking the consumer population into groups based not only on age and diagnosis but also on time since entrance to the DDS caseload, we use a similar simulation approach to test the effect of the eligibility standardization. Figure 4-9 shows the number of consumers who actually entered the RC caseload each year after the eligibility standardization, compared to the number we would have expected based on earlier patterns. By FY06-07, we estimate approximately 3,000 fewer consumers per year were determined to be eligible for DDS services than would have been in the absence of the standardization. Although the difference between the Actual/projected and counterfactuals in Figure 4-9 shows the reduction in the number of newly eligible we would expect in a given year, the effect of people not entering the consumer population has a cumulative effect on total expenditures. The cost avoidance in FY05-06, for example, includes costs avoided for those who would have newly entered the caseload in FY05-06 plus the costs for people who would have entered in FY04-05 plus the costs for those who would have entered in FY With fewer consumers to serve, we estimate that the DDS avoided $1.6 million in FY03-04, $8.8 million in FY04-05, and $25.8 million in FY