IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS The Texas Medical Transportation Program provides nonemergency medical transportation to covered medical appointments for Medicaid clients who have no alternative means of transportation. Medicaid members in Texas made an estimated 3.6 million trips using this program in fiscal year 2015 at an All Funds cost of $211.9 million. Before fiscal year 2009, the Medical Transportation Program was operated by the state, which paid transportation providers a fee for each trip provided. The Eighty-first Legislature, 2009, directed the Texas Health and Human Services Commission to implement a capitated broker model in areas of the state where it could be sustained. A capitated broker model pays a set rate per eligible client in the region, whether the program is used or not. To mitigate operational risk concerns, the agency developed a regional pilot. However, statute was amended in 2013 to implement the broker model statewide before the results of this pilot were available. Implementation of the broker model was intended to improve the cost-effectiveness of the program. To date, the broker model has not achieved this goal. Costs per client served have increased, access to transportation services has decreased, and complaints from clients have increased. Since implementation of the broker model, costs per trip have increased from $22.31 to $59.40. Conversely, the percentage of Medicaid clients that are served by the program has decreased by one-half from fiscal years 2009 to 2015. In regions that converted to broker management in fiscal year 2015, substantiated complaints increased 43.4 percent. These cost and quality issues have been due in part to procurement and contract management failures. The Health and Human Services Commission did not, as required, seek competitive bids based on price in the most recent expansion of the model. The agency also selected vendors with performance risks rather than vendors recommended by program staff. As a result, contracts were awarded to companies with histories of poor performance. The Health and Human Services Commission considers almost all the vendors in the program to pose a high risk to the state and to Medicaid clients. The agency has also canceled two contracts due to client safety risks, contract violations, and financial issues. Increasing external oversight and evaluation and authorizing the Health and Human Services Commission to LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 use the most cost-effective model would ensure that the program provides access to services at a lower cost per trip. FACTS AND FINDINGS  In fiscal year 2009, one in four Medicaid clients younger than age 21 had unmet transportation needs. Although comparable measures are not available for recent fiscal years, the percentage of Medicaid clients served by the Medical Transportation Program has decreased from 7.7 percent of clients in 2009 to 3.8 percent in fiscal year 2015. Program changes made by the Health and Human Services Commission may have had the greatest effects on access, however, the conversion to a broker model also coincided with decreases in the number of medical transportation users.  Since implementation of broker management, costs per trip have increased from $22.31 to $59.40. As a result, Legislative Budget Board staff estimate that the 3.6 million trips provided in fiscal year 2016 cost approximately $120.2 million more in All Funds compared to what operations through fee-for-service would have cost in fiscal year 2016. CONCERNS  The goal of the Medical Transportation Program is to provide cost-effective transportation to clients who have no other means of transportation to medical appointments. However, no mechanism exists to track unmet transportation needs and determine if the program is achieving this goal.  The Health and Human Services Commission established conflicting contract performance reporting requirements, failed to establish data sources, and has not collected the appropriate data. For example, the agency does not systematically measure whether clients are transported on time to medical appointments. Additionally, surveys intended to measure the quality of services produce unreliable data due to differences in methodologies across survey firms. LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 191 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS   The Health and Human Services Commission has not incorporated cost per trip in its monitoring of the cost performance of the Medical Transportation Program. As a result, the agency does not adequately monitor the cost-effectiveness of services provided by the program. It is unclear whether a competitive market of qualified brokers is available to provide more cost-effective services than direct payment to providers through a fee-for-service model. According to the Health and Human Services Commission, contracts for 11 of the 12 broker operated regions pose a high risk to the state. Both spending per trip and complaints have increased in the broker model. Many of these complaints relate to failures to transport clients to medical appointments. The agency has also cancelled two contracts due to client safety risks, contract violations, and financial issues. OPTIONS  192 Option 1: Include a rider in the 2018–19 General Appropriations Bill requiring the Health and Human Services Commission to notify the Legislative Budget Board and relevant standing committees of the Legislature if the percentage of Medicaid clients with unmet transportation needs rises above levels reported in calendar year 2012. The agency would be required to develop a corrective action plan to remediate increases in unmet transportation needs.  Option 2: Amend statute to require the Health and Human Services Commission to hire a single, independent vendor to conduct surveys of customer satisfaction and unmet transportation needs.  Option 3: Include a rider in the 2018–19 General Appropriations Bill requiring the Health and Human Services Commission to measure and report the average costs per trip of the program at the end of each fiscal year to the Legislature.  Option 4: Amend statute to authorize the Health and Human Services Commission to use the most costeffective model of delivering medical transportation services. The agency would be required to conduct a needs assessment before any reprocurement of broker services and establish a cost threshold for accepting bids. Include a contingency rider in the 2018–19 General Appropriations Bill to require the agency LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 to seek Legislative Budget Board approval for this cost threshold. Contracts would only be awarded to brokers in regions where qualified vendors bid at prices that meet or fall below this cost threshold. DISCUSSION The Texas Medical Transportation Program (MTP) provides nonemergency medical transportation (NEMT) to covered medical appointments for Medicaid clients who have no alternative means of transportation. Medicaid clients may experience difficulties securing transportation to medical appointments for a variety of reasons. For example, some clients may not have the means to purchase a car or pay fuel costs for trips to medical appointments. Others may have disabilities and health impairments that limit their abilities to drive or use public transportation. Additionally, the type of treatment clients receive can prevent them from safely transporting themselves. For example, clients might not be able to safely transport themselves after receiving dialysis treatment, chemotherapy, or surgery. Access to NEMT could be required for these clients to safely continue treatment. Figure 1 shows the types of visits for which Medicaid clients in Texas received NEMT in fiscal year 2015. Without access to NEMT services, these clients may miss medical appointments. Several national research studies have shown that transportation difficulties are associated with adverse effects on long-term health and quality of life. Delays in access to care due to a lack of transportation have also been shown to increase the need for emergency care and preventable hospitalizations. According to a 2007 survey of Texas Medicaid clients conducted by the Public Policy Research Institute at Texas A&M University, approximately one-third of children enrolled in Texas Medicaid missed medical appointments as a result of transportation barriers. Half of these, or one-sixth of children enrolled in Texas Medicaid, missed visits for routine care appointments. Among caregivers whose children missed medical appointments, 35.0 percent reported needing emergency care services as a result of missing these routine appointments. According to the Public Policy Research Institute, these findings indicate that healthcare costs could be reduced through improved access to transportation. In Texas, a variety of NEMT services are available through MTP, which is operated by the Texas Health and Human Services Commission (HHSC). Like services in most states, Texas’ MTP provides NEMT services through mass transit and dispatched vehicles. Dispatched vehicle services, typically LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 1 MEDICAL TRANSPORTATION PROGRAM TRIPS BY TYPE OF MEDICAID-COVERED APPOINTMENT, FISCAL YEAR 2015 Primary Care Physician (9.3%) Dentistry (1.9%) Specialists (22.8%) Cancer Treatment (1.2%) Mental Health (7.7%) Other (7.4%) Therapy (24.3%) Dialysis (28.5%) Hospital or Clinic (1.1%) Laboratory (0.9%) Obstetrics/Gynecology, Family Planning (0.9%) Surgery (0.6%) Pharmacy (0.6%) Miscellaneous (0.1%) S඗ඝකඋඍ: Texas Health and Human Services Commission. referred to as demand response, are provided by individual providers. This service is available when mass transit is not available or feasible. Historically, the MTP demand response category has represented the majority of trips and spending. In fiscal year 2015, as shown in Figure 2, demand response represented 73.6 percent of trips, and mass transit represented 2.0 percent of trips. In the Houston, Dallas, and Austin regions, the percentage of trips using mass transit ranged from 2.0 percent to 5.0 percent. In most regions of the state, few, if any, of trips were provided using mass transit. FIGURE 2 MEDICAL TRANSPORTATION PROGRAM TRIPS BY MODE OF TRANSPORTATION, FISCAL YEAR 2015 Individual Transportation Provider Program (24.4%) Demand Response (73.6%) Mass Transit (1.9%) Airfare (0.1%) S඗ඝකඋඍ: Texas Health and Human Services Commission. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 Family members and friends of Medicaid clients can enroll as volunteer drivers to provide transportation services through the individual transportation provider (ITP) program within MTP. Texas also offers reimbursement for meals and lodging for overnight and extended medical stays. Certain clients with cystic fibrosis and families with children may be eligible to receive meals, lodging, and mileage reimbursement in advance through the advance funds benefit. The advance funds benefit was established in response to class action litigation filed against HHSC in 1993, Frew vs. Commissioner of Texas Health and Human Services Commission. According to the plaintiffs, access to MTP in Texas was problematic in part due to poor customer service by HHSC and its vendors. HHSC has been party to a consent decree with the plaintiffs since 1996. The consent decree includes a variety of Medicaid program functions, including NEMT services from 1996 to 2015. By participating in the consent decree, HHSC agreed to conduct surveys of access and demonstrate improvement in these surveys. Since the 1990s, HHSC has focused on complying with the Frew lawsuit while attempting to mitigate program integrity risks identified in state and federal audits. Although the class action lawsuit is specific to Texas, many states experience competing priorities among access, cost, and program integrity. According to the U.S. Government Accountability Office, NEMT programs represent a high-risk area of LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 193 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS government operations. This risk is in part due to the size, complexity, and nature of the services. For example, providers often have difficulty complying with documentation requirements. Other states have also reported challenges controlling costs while ensuring access to NEMT services. STRATEGIES TO ADMINISTER MEDICAL TRANSPORTATION States use a variety of strategies to provide NEMT services, including operating the program directly using state agency staff. Alternatively, states can outsource management to external vendors. When states directly operate the program, state employees take requests for transportation, review eligibility, schedule trips, process claims, and develop provider networks with transportation providers. This model was used by Texas until fiscal year 2015 and is typically referred to as a fee-for-service (FFS) system because the state reimburses transportation providers directly. agency and MCO. These third parties, typically described as brokers or managed transportation organizations (MTO), can be public transit agencies, nonprofit entities, or for-profit companies that specialize in providing brokerage services. In determining how to award contracts to brokers, some states use a single statewide NEMT contract with one broker, while others, such as Texas, use a regional broker model. As shown in Figure 3, transportation providers are reimbursed by brokers in this model. These brokers are paid a rate per person per month based on the number of people within their region. This fixed rate is known as capitation. States also contract for NEMT management, and sometimes outsource a limited number of functions. Before MTP’s transition to the broker model, HHSC outsourced claims processing to an administrative services contractor in fiscal year 2012. In other cases, states may choose to outsource most program management functions. In 2003, 41.0 percent of states used a broker model. Until 2005, states were required to seek a federal waiver to operate NEMT through a broker model. The federal Deficit Reduction Act of 2005 removed this requirement and authorized states to amend their Medicaid state plan through a standardized process. The rules to implement the NEMT provisions of the Act were finalized by the U.S. Center for Medicare and Medicaid Services (CMS) in January 2009. Following these changes, the number of states using broker models expanded. By 2014, according to the Texas Transportation Institute, 63 percent of states were using a broker model for NEMT services either in-part or in whole. Two options are available for broadly outsourcing program management. A state can integrate NEMT within its managed care contracts for health services. In this model, managed care organizations (MCO) are responsible for health services and for NEMT services. Alternatively, a state can outsource most NEMT program management functions to third parties, which operate independently of the state The effects of this transition to NEMT broker models has been unclear. Interviews and surveys of state agencies responsible for NEMT administration typically suggest that brokers have achieved cost savings without compromising quality. According to several research institutions, however, minimal academic peer-reviewed research from independent sources supporting these assertions has been conducted. FIGURE 3 NONEMERGENT MEDICAL TRANSPORTATION MODELS, FISCAL YEAR 2016 Fee-for-service Model State agency responsible for medical transportation Payment for services provided Transportation providers Broker Model State agency responsible for medical transportation Payment for each Medicaid client residing in the broker’s region (capitation) Brokers Payment for services provided Transportation providers N඗ගඍ: The managed care organization (MCO) model works similarly to the broker model. The capitation rate in an MCO model would include payments for transportation and for health services. S඗ඝකඋඍ: Legislative Budget Board. 194 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS TRANSPORTATION BROKERS IN TEXAS As shown in Figure 4, until 2012, Texas MTP was implemented using a FFS system, which experienced a number of problems. In the 1990s, the former Texas Department of Health was responsible for operating the program. According to research conducted by the federal Transit Cooperative Research Program, the Department of Health focused on maximizing competition at the expense of quality. The standards for drivers and vehicle safety, for example, were lower than the standards required of public transit operators. As a result, higher-quality public transit vendors were priced out of the Texas NEMT program during the late 1990s. A 2002 audit conducted by the Texas State Auditor’s Office (SAO) indicated negative effects from this strategy. The audit found that the Texas Department of Health failed to consider any factors other than price in its procurement. This procurement process led to failures to deliver services by unprepared vendors. In one region, the agency canceled a contract after eight weeks due to the contractor’s inability to deliver required services. In some cases, although price was the sole factor, the agency did not negotiate rates or achieve maximum cost savings. The audit found that, overall, the agency did not adequately manage the procurement or program. Based in part on these audit findings, in fiscal year 2003, the Texas Comptroller of Public Accounts (CPA) recommended transferring operations from the Department of Health to HHSC. CPA also recommended outsourcing management of MTP to either a single statewide broker or regional brokers. During fiscal year 2003, the Seventy-eighth Legislature transferred MTP to HHSC as part of a system wide consolidation of health and human services. HHSC then delegated operational responsibility of the program to the Texas Department of Transportation (TxDOT). According to a 2007 SAO audit, TxDOT addressed the previous procurement problems in the program. In the procurement reviewed by SAO, TxDOT considered both price and vendor qualifications in making awards. However, the audit identified a number of internal control and oversight weaknesses and concerns about the quality of services. Before the results of this audit were published, the Legislature discontinued delegation of MTP operations to TxDOT. This decision was based on authorization included in the federal Deficit Reduction Act of 2005 to amend Medicaid state plans to use the broker model. Senate Bill 10, Eightieth Legislature, 2007, authorized HHSC to contract for MTP using public and private transportation providers or regional transportation brokers. Although Senate Bill 10, 2007, gave HHSC the authority to change the MTP model, the federal rules authorizing state plan amendments were not effective until January 20, 2009. After the federal rules were finalized, the Eighty-first Legislature, General Appropriations Act (GAA), 2010–11 Biennium, required HHSC to use a capitated regional broker model “in areas of the state that the Commission finds can sustain a regionalized model.” HHSC’s Office of General Counsel subsequently determined that the 2010–11 GAA did not require use of the broker model statewide. An August 2009 memo to the HHSC FIGURE 4 POLICY CHANGES RELATED TO THE TEXAS MEDICAID MEDICAL TRANSPORTATION PROGRAM CALENDAR YEARS 2005 TO 2015 Administration transferred from TxDOT to HHSC > @ 2010–11 GAA requires HHSC to use brokers SB 8, 2013, requires statewide use of brokers 2008 2009 2010 2011 Federal rules authorizing states to use brokers are finalized 2012 2013 Broker model starts in Dallas and Houston areas 2014 HHSC expands broker model statewide 2015 2016 HHSC releases RFP to expand broker model statewide N඗ගඍඛ: (1) TxDOT = Texas Department of Transportation; HHSC = Texas Health and Human Services Commission; RFP = request for proposals; GAA = General App[ropriations Act. (2) SB 8 refers to Senate Bill 8, Eighty-third Legislature, Regular Session, 2013. S඗ඝකඋඍ: Legislative Budget Board. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 195 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS Executive Commissioner documented concerns about the potential effects of a broker model on access and quality. HHSC analysis showed that Texas’ benefit package was different from what transportation brokers were accustomed to managing. In particular, the advance funds benefit used by some to receive NEMT services was established in response to the Frew decree, and it was unclear how a broker model based on capitation would account for these costs. As a result of these concerns, HHSC staff recommended that the broker model be piloted. The intent of the broker pilot was to give HHSC information to evaluate the effects of the model before expanding it to other areas of the state. The agency estimated that two to four years of operations would be necessary to determine the effects of the conversion. HHSC awarded contracts for broker pilots in the Houston and Dallas regions, and operations started in March and April 2012, respectively. Before the results of this pilot were available, the Eighty-third Legislature, Regular Session, 2013, passed Senate Bill 8, which required HHSC to operate MTP using a regional broker model across the state. The legislation required the agency to seek a federal waiver to authorize transportation brokers to directly own transportation providers. As a result, the agency did not rely on the discretion authorized by the Deficit Reduction Act of 2005. Instead, HHSC was required to amend its state plan and to receive a federal waiver. CMS authorized HHSC for a waiver effective September 1, 2014, to August 31, 2016. PROCUREMENT OF THE BROKER MODEL The expansion of the broker model in fiscal year 2015 resulted in HHSC issuing a request for proposals (RFP) to procure vendors to operate in 11 regions of the state. During this procurement, HHSC did not include price as a factor or properly disclose evaluation criteria in the RFP. HHSC also did not award contracts to vendors in accordance with internally developed performance criteria. These procurement issues have had negative effects on MTP. CONSIDERATION OF PRICING The statewide expansion of brokers in fiscal year 2015 was subject to several mandates requiring price competition. Requirements to include price in the procurement for the 2015 expansion of MTP brokers included: • state plan disclosure requirements linked to federal law – CMS required HHSC to assure in writing that the agency could prove upon request that 196 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 brokers were selected through a competitive bidding process. The process was required to include cost as a consideration to be consistent with federal rules. The cited federal rules distinguish between competitive bidding and competitive proposals. Competitive bidding, as described in these rules, requires firms to compete on price; • state statutes requiring competitive bidding – Senate Bill 8, Eighty-third Legislature, Regular Session, 2013, required the agency to develop a capitated rate system using competitive bidding to procure the contract. The Texas procurement manual, State of Texas Contract Management Guide, and common legal definitions of bid are all based on vendors submitting prices; and • HHSC’s administrative rules – At the time of the procurement, HHSC rules required the agency to include price unless the agency documented reasons for disregarding price. However, the RFP that HHSC issued did not require vendors to submit pricing information and the agency did not document a reason for excluding pricing from its procurement. Instead, vendors were required to submit audits of their firms to help evaluate their operational efficiency. As a result, preliminary awards were granted based on criteria other than price, including indicators of probable performance and the effects that the proposals would have on HHSC’s productivity and resources. After this preliminary award, HHSC disclosed the capitation rates, and the selected vendors had to agree to these rates to receive contracts. All vendors received the same rate except for adjustments for the demographic and utilization history of each region. Legislative Budget Board (LBB) staff interviews with HHSC staff in fiscal year 2016 indicated that the agency determined that capitation precludes bidding. HHSC does not require medical MCOs to compete on price and carried this practice forward to the transportation program despite indicators from vendors that this was not a standard practice for NEMT procurements. After the RFP was posted, HHSC received questions from vendors. One vendor asked HHSC why it was not required to bid on price in the following question to the agency: Unlike any other Medicaid RFP we have encountered, the State is requesting that contractors turn in a proposal without knowing what the capitated rate will be? This LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS concept that any reputable organization would go to the trouble of putting a proposal only to find later that the State’s capitated rate is not sufficient is at best questionable. … Every other RFP in U.S. we are familiar with asks bidders to propose a capitated rate, why not here? LBB staff reviewed NEMT procurements in 12 other states, all of which required vendors to submit bids with prices, including states that use actuarially sound rates. HHSC’s response to the vendor’s question advised the vendor to decide whether to execute the contract after HHSC set the rates. During the RFP solicitation, the agency received similar questions from other vendors. One vendor asked, “How will the HHSC determine the ‘long-term cost’ of the vendor, if the HHSC is determining the capitated rates?” HHSC responded that MTO’s cost performance would become a factor in subsequent rates; therefore, HHSC was seeking vendors with plans for reducing costs. The agency was also asked how best value would be determined without submission of cost proposals. The agency stated that this information is not disclosed during active procurements. This practice conflicted with state contracting regulations and with HHSC’s rules. The State of Texas Contract Management Guide, which was also in effect at the time of procurement, requires agencies to provide best value and proposal evaluation criteria within solicitations. The Guide also requires agencies to explain to potential vendors how best value will be determined and how proposals will be evaluated. In solicitations, agencies should include the weights used for each criterion. Additionally, HHSC’s administrative rules required the agency to publish the criteria used to determine best value. VENDOR SELECTION PROCESS HHSC established specific weights internally with a detailed rubric to evaluate vendor proposals. An evaluation team composed of HHSC staff reviewed each proposal and scored it against this rubric. Despite this internal procedure, several inconsistencies occurred throughout the vendor selection process. In North Texas (Region 4), the RFP evaluation team determined that one “respondent did not demonstrate even a basic understanding of nonemergency medical transportation services or program requirements.” Additionally, the vendor failed to demonstrate a financial capacity to provide broker services to HHSC. This vendor was ranked last among four proposals, but it was awarded an MTP contract in one region. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 HHSC awarded contracts to the highest-scoring vendor in four of 11 regions. One vendor with a history of performance problems was awarded contracts for multiple regions. According to a 2007 audit by SAO, 16.0 percent of that vendor’s contracted transportation providers had invalid driver licenses, and 19.0 percent had criminal histories. The evaluation team determined that this vendor “did not clearly address all elements of the RFP.” The reasons for these deviations were not documented or justified during the procurement process. According to LBB staff interviews with HHSC staff in fiscal year 2016, these decisions may have been made to increase the diversity of contracted vendors. If the agency had followed its evaluation criteria, the agency would have awarded contracts in most regions to a single vendor. As Figure 5 shows, in regions where the highest-scoring vendor won, complaints and access trends improved compared to the other regions. In regions where the highestscoring vendor was not awarded a contract, complaints increased and access to services decreased. Poor performance from underqualified vendors resulted in HHSC terminating contracts for the North Texas (Region 4) and South Central Texas regions (Region 8). These terminations were made seven months and 15 months, respectively, after contracts were awarded due to client safety issues and possible financial fraud. In North Texas, the termination was specifically related to concerns that the RFP evaluation team raised 15 months previously. LBB staff has been working with HHSC to remediate these problems. On April 4, 2016, LBB staff recommended to HHSC staff that the agency reprocure the MTP contracts using competitive bidding as soon as possible. In April 2016, HHSC indicated that it would reprocure the contracts. However, HHSC subsequently renewed the broker contracts through August 31, 2018. It is unclear if the agency intends to re-procure the contracts in the future using price as a competitive factor. In addition to procurement problems, MTP has experienced increases in cost and decreases in access to services in the FFS and broker models. These concerns are not explained solely by procurement issues. ACCESS TRENDS As shown in Figure 6, the percentage of Medicaid clients served by MTP was 7.7 percent for fiscal year 2009. At this level of utilization, 23.9 percent of Texas children enrolled LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 197 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 5 COMPARISON OF MEDICAL TRANSPORTATION PROGRAM VENDOR SCORES TO CHANGES IN ACCESS AND COMPLAINTS FISCAL YEARS 2014 TO 2015 REGION GAP IN PROBABLE PERFORMANCE SCORE PERCENTAGE DECREASE IN UNIQUE CLIENTS SERVED (5.45) (25.6%) 50.8% 6. Southwest Texas PERCENTAGE CHANGE IN COMPLAINTS 4. North Texas (5.21) (11.2%) 167.6% 1. Panhandle (3.21) (21.0%) 110.4% 3. Northwest Central (3.21) (23.3%) 112.2% 10. South Texas (3.21) (27.1%) 23.4% 5. East Texas (2.93) (21.8%) 228.7% 9. Southeast Texas (2.93) (30.3%) 146.0% 2. West Texas 0 (2.9%) 41.8% 8. South Central Texas 0 (5.0%) (26.6%) 11. Northeast Central 0 (10.2%) (29.9%) 7. Central Texas 0 (14.0%) (32.7%) N඗ගඍ: The gap in probable performance shows the difference from the request for proposals (RFP) evaluation team’s probable respondent performance score for the winning vendor to the vendor with the highest score for all RFP criteria. A score of zero indicates that the Texas Health and Human Services Commission selected the highest-scoring vendor. S඗ඝකඋඍ: Texas Health and Human Services Commission. FIGURE 6 PERCENTAGE OF MEDICAID CLIENTS SERVED BY THE MEDICAL TRANSPORTATION PROGRAM, FISCAL YEARS 2001 TO 2015 12% 10% 9.8% 8% 6% 4% 3.8% 2% 0% 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 S඗ඝකඋඍඛ: Texas Health and Human Services Commission; Texas Department of Transportation; Texas Department of Health. in Medicaid had unmet transportation needs. This unmet need was calculated in a report produced by the Public Policy Research Institute at Texas A&M University for the Frew lawsuit. The report evaluated unmet transportation needs through a survey of caregivers for Medicaid children. The report calculated the percentage of Medicaid clients 198 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 younger than 21 years old who did not use MTP and had either a difficult or very difficult time obtaining transportation to medical appointments during the previous year. The study authors concluded that NEMT services were underutilized in Texas, even though past evaluations LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS concluded that many Medicaid clients could potentially benefit from the MTP services. Since this time, the percentage of Medicaid clients served by MTP has decreased by 50 percent. Therefore, more Medicaid clients likely have unmet transportation needs than are served by the program. There are 154,266 children and adults currently served by the program. Based on the estimate of 23.9 percent of children in Medicaid with unmet transportation needs and the current number of children in STAR, 646,671 children may have unmet transportation needs. Most of the decrease in utilization occurred during the previous use of the FFS model. However, statewide expansion of the broker model in fiscal year 2015 coincided with an additional decrease in the number of MTP users. By fiscal year 2011, the percentage of Medicaid clients that MTP served had risen from 7.7 percent to 9.8 percent of Medicaid clients. According to documents filed by HHSC, this increase in access was a result of corrective actions taken by the agency from fiscal years 2008 to 2012 to resolve the Frew lawsuit. Corresponding to this increase in the percentage of Medicaid clients using MTP, the calendar year 2012 survey found unmet need had decreased from 22.0 percent to 17.0 percent in urban areas and from 23.0 percent to 14.0 percent in rural areas. These improvements in MTP access and other system improvements resulted in the court vacating the Frew Corrective Action Order for NEMT services in fiscal year 2015. However, Figure 6 shows that these gains reversed starting in fiscal year 2012. The number of clients using MTP decreased from 346,147 clients in fiscal year 2011 to 154,266 clients in fiscal year 2015, despite Medicaid caseload growth. In fiscal year 2015, MTP served 3.8 percent of Medicaid clients. HHSC’s program changes affected access to MTP before the adoption of the statewide broker model. In 2012, HHSC began to implement controls in response to audit deficiencies (see the Program Integrity Risks section). The agency revised eligibility requirements for the advance funds and ITP benefits, established controls for the types of reimbursable expenses, and required individuals to submit compliance documentation. According to HHSC, the changes to the advance funds benefit produced immediate cost savings due to decreases in fraud and misuse. From fiscal years 2012 to 2014, HHSC also took steps to enforce a parental accompaniment rule. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 As shown in Figure 7, before MTP’s conversion to a broker model in fiscal year 2015, the number of users in FFS regions decreased by 128,620 Medicaid clients from fiscal years 2011 to 2014. This decrease represented 67.0 percent of the total decrease in MTP users from fiscal years 2011 to 2015 across all regions. These figures suggest that antifraud-related efforts and other changes that HHSC made affected the number of MTP users, independent of the state’s conversion to a broker model. FIGURE 7 MEDICAL TRANSPORTATION PROGRAM USERS BY REGION, FISCAL YEARS 2011 TO 2015 220,000 190,000 160,000 130,000 100,000 70,000 2011 2012 2013 Houston and Dallas 2014 2015 Other Regions N඗ගඍ: Pilots of the broker model in the Houston and Dallas regions started in March and April of fiscal year 2012. The broker model was expanded statewide at the beginning of fiscal year 2015. S඗ඝකඋඍ: Texas Health and Human Services Commission. Although HHSC’s program changes may have had the largest absolute effects on access, the conversions from a FFS model to a broker model also coincided with further decreases in the number of MTP users. After the fiscal year 2015 statewide expansion, a decrease of 18.3 percent occurred in those regions in addition to the decrease that occurred during use of the FFS model. These decreases in the number of users coincided with a decrease in the reported number of trips per user. As Figure 8 shows, the fiscal year 2015 trip counts suggest a decrease in the number of trips per users compared to fiscal year 2011. Given historical access and unmet need, the current volume of MTP utilization may correspond to a large volume of unmet transportation needs for Medicaid clients. This trend cannot be explained by procurement problems and instead points to overall performance issues within MTP. The goal of MTP is to provide cost-effective transportation to clients who have no other means of transportation to medical LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 199 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 8 ONE-WAY TRIPS PER UNIQUE MEDICAL TRANSPORTATION PROGRAM USER PER YEAR, FISCAL YEARS 2001 TO 2015 35 30 29.2 25.1 25 21.0 20 15 10 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 N඗ගඍ: Trip counts may have been calculated differently in certain years. Pilots of the broker model in the Houston and Dallas regions started March and April of fiscal year 2012. The broker model was expanded statewide at the beginning of fiscal year 2015. Transportation brokers have underreported claims data to the Health and Human Services Commission, and some trips may be missing, especially for fiscal year 2015. Inflating reported trips for fiscal year 2015 to account for this missing data would add 2.0 trips to the average number of trips per user. S඗ඝකඋඍඛ: Texas Health and Human Services Commission; Texas Department of Transportation; Texas Department of Health. appointments. To ensure the program is providing transportation needs to clients as intended, Option 1 would amend the 2018–19 General Appropriations Bill to include a rider requiring HHSC to notify the LBB and relevant standing committees of the Legislature if the percentage of Medicaid clients with unmet transportation needs exceeds the levels in calendar year 2012. In calendar year 2012, the last year in which the percentage of unmet transportation needs was determined, the level of unmet transportation needs was 17.0 percent in urban areas and 14.0 percent in rural areas. The agency would be required to develop a corrective action plan to remediate any unmet transportation needs in any instance in which survey data reveals unmet transportation needs exceed calendar year 2012 levels. QUALITY TRENDS In addition to the downward trend in MTP utilization, indicators show decreasing program quality. Statewide complaints relative to the number of MTP users have more than tripled since fiscal year 2011. Many of these complaints have been substantiated and relate to failures to deliver clients to medical appointments. Historical survey data indicates that each expansion of the broker model has been followed by an increase in complaints about the quality and reliability of services in MTP. 200 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 A survey fielded from October to December 2012, when HHSC conducted its pilot, sampled MTP users in both the broker and FFS models. The timing of this survey provided some Medicaid clients six to eight months of experience with the broker model in the Houston and Dallas regions. The survey found that broker users and FFS users rated the usability of NEMT services equivalently. According to survey respondents, travel time among broker users was more reasonable than among FFS users, although this result may have been related to the more urban composition of broker users. However, responses regarding timeliness and reliability of service indicated differences between the broker and FFS systems. Within the Houston and Dallas regions, broker users were more likely to miss appointments because of MTP transportation problems than users of the FFS system. Among broker users, 34.0 percent reported missing either a few or almost all medical appointments, compared to 27.0 percent of FFS users. Broker users were also more likely to be dissatisfied with wait times for pickup after medical appointments. Medical providers in this survey also rated performance in broker regions as less reliable. Despite the differences in missed appointments, overall satisfaction was rated almost the same among users of the two systems. Among broker users, 73.0 percent reported that they were LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS satisfied with services, compared to 75.0 percent in nonbroker regions. Transportation providers in broker regions reported that they were less satisfied with the program than providers in FFS regions. The urban composition of broker users compared to FFS users varied and may have contributed to these outcomes. However, complaint data during this period supports the conclusion that NEMT users experienced problems with transportation to medical appointments after the transition to the broker model. After statewide expansion in fiscal year 2015, negative effects related to timeliness and service quality continued to be associated with transportation brokers in Texas. However, in three regions of the state in which the vendor with the highest score based on RFP criteria is operating the program, the percentage of complaints declined between fiscal years 2014 and 2015. Figure 9 shows increases in the ratio of complaints to yearly unique MTP users (the complaint ratio). These increases coincided with each expansion of the broker model. For fiscal years 2011 and 2012, the complaint ratio equaled approximately 3.5 percent. For fiscal year 2013, the first full year of broker operations in Houston and Dallas, the complaint ratio increased to 9.4 percent. During fiscal year 2013, more complaints were filed in the Houston and Dallas regions than in all regions in fiscal year 2012 combined. For fiscal year FIGURE 9 RATIO OF COMPLAINTS TO MEDICAL TRANSPORTATION PROGRAM USERS, FISCAL YEARS 2011 TO 2015 12.9% 9.4% 3.5% Fee-for-service Broker statewide Broker pilots 2013, the rate of total complaints across regions remained steady. When statewide expansion of the broker model occurred during fiscal year 2015, the complaint ratio increased to 12.9 percent. This increase was caused by a 62.7 percent increase in the volume of complaints in the regions that converted to the broker model. Complaints in the Houston and Dallas regions remained steady for fiscal year 2015. From fiscal years 2011 to 2015, total complaints increased as the number of MTP users decreased. Additionally, the volume of complaints in relation to the number of people using the program more than tripled. Figure 10 shows some of the reasons for which complaints were filed in the regions that converted to a broker model in fiscal year 2015. In these regions, overall complaints substantiated by HHSC increased from 3,041 complaints to 5,379 complaints. Substantiated complaints that clients were never transported to their medical appointments increased by 76.9 percent from fiscal years 2014 to 2015. Complaints about the overall timeliness of service represented the most common substantiated complaints. Substantiated complaints also increased related to the individual transportation provider program and customer service problems. Changes in complaint volumes can occur for reasons unrelated to program quality. For example, if brokers were more transparent in the complaint filing process, then complaints may have risen in part due to greater awareness about a grievance process. This awareness could increase the likelihood of clients seeking to resolve issues. However, the following findings suggest that changes in the quality of services contributed to increases in complaint rates from fiscal years 2011 to 2015: • 2012 surveys of MTP users in the Houston and Dallas regions showed similar systematic differences in the quality of service in the broker model compared to the FFS model – survey data is not subject to the same type of reporting bias that complaint data could be, and, as such, when survey data and complaint data show consistent findings, this correlation increases the confidence about the reliability of complaint data; • 2011 2012 2013 2014 2015 N඗ගඍ: Pilots of the broker model in the Houston and Dallas regions started in March and April of fiscal year 2012. The broker model was expanded statewide at the beginning of fiscal year 2015. Multiple complaints may have been filed by a single user. The ratio shows a general gauge of complaints relative to the number of unique nonemergency medical transportation users each year. S඗ඝකඋඍ: Texas Health and Human Services Commission. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 complaints rose in eight of the 11 regions converted to a broker model – one vendor consistently reduced complaints across two regions and had the highestrated probable performance score by the RFP team among all vendors; this vendor also experienced a lower-than-average decrease in the number of unique users, suggesting that vendor performance has a distinct effect on complaint volumes; and LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 201 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 10 SUBSTANTIATED COMPLAINTS FOR REGIONS THAT CONVERTED FROM FEE-FOR-SERVICE MODEL TO BROKER MODEL FISCAL YEARS 2014 TO 2015 2014 (FFS) 2015 (BROKERS) Client was not transported at all 866 1,365 Program policy or procedure 79 188 Client was not picked up within 1 hour of request 575 685 MTO ITP enrollment process N/A 177 COMPLAINT COMPLAINT 2014 (FFS) 2015 (BROKERS) Client arrived late to appointment 538 528 Operator issues 80 139 MTO ITP reimbursement N/A 448 Operator mistreated client 71 125 Dispatcher failed to control or monitor service delivery 269 370 Client unable to reach MTO via dispatch number N/A 101 23 366 Client not called day before appointment to schedule pickup time 81 101 N/A 280 Total Substantiated Complaints 3,041 5,379 Customer service MTO scheduling error N඗ගඍඛ: (1) FFS = fee-for-service model; MTO = managed transportation organization; ITP = individual transportation provider. (2) Complaints exclude those related to transportation brokers in the Houston and Dallas pilot program regions. S඗ඝකඋඍ: Texas Health and Human Services Commission. • the scale of the increase in complaints was large, complaint volumes relative to usage more than tripled from fiscal years 2011 to 2015. Additional data to clarify the effects of brokers on service quality are either missing or unreliable. HHSC has established performance standards in contracts for the timeliness of service and the percentage of clients receiving appropriate services. But the agency has not identified a way to measure or track these standards. As a result, the agency is largely dependent on complaints as an indicator of service quality. Transportation brokers are also required to hire and pay vendors to conduct customer satisfaction surveys, but these surveys are problematic. According to HHSC, the MTP contract has conflicting requirements regarding whether these broker-led surveys are to be conducted quarterly or annually. For the first three quarters of fiscal year 2015, four of the 11 regions that converted to a broker model did not provide surveys to HHSC. To resolve any confusion regarding the requirements, HHSC required brokers to conduct only annual surveys. Results of these broker-led surveys are difficult to compare because each MTP vendor conducted its surveys differently. For example, different scales are used to rate satisfaction. In one region, the annual survey did not calculate a total satisfaction score, despite this score being a key performance measure in the MTP contract. Surveys may also not always include the appropriate mix of clients. In at least one instance, the survey was conducted only among clients who were 202 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 transported, potentially excluding clients whose MTO did not arrive for pickup or who were unable to schedule services. This lack of reliable data makes it difficult to verify quality trends within MTP. Option 2 would amend statute to consolidate responsibility for surveying MTP users and Medicaid clients with HHSC, rather than with each broker. HHSC would be required to contract with a single, independent vendor to ensure highquality, consistent survey data. This consolidation would align MTP survey practices with those in place for other Medicaid programs. For instance, HHSC contracts with a single external quality review organization (EQRO) to survey Medicaid clients regarding medical benefits provided by MCOs. This singularity ensures independence and comparability of results. Surveys of MTP users and Medicaid clients could be incorporated into the contract with the EQRO or performed by another qualified survey firm. The agency would be required, at a minimum, to estimate MTP client satisfaction as well as unmet transportation needs among Medicaid clients. COST EFFECTS OF BROKER MODEL MTP costs have increased by approximately $316.5 million cumulatively in All Funds since the introduction of transportation brokers in fiscal year 2012. For fiscal year 2016, LBB staff estimate the cost difference between the broker model and what operations using FFS would have cost was approximately $120.2 million in All Funds per year. This cost difference is due to increases in the cost of LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS administering the program ($46.3 million) and for paying transportation-related expenses ($73.9 million). a 76.7 percent increase from fiscal year 2011. This increase occurred despite stable fuel costs from fiscal years 2011 to 2013. Despite fuel cost decreases in fiscal years 2014 and 2015, actual costs were $59.40 per trip for fiscal year 2015, after adjusting for missing claims from some brokers. Based on costs for fiscal year 2011 and subsequent fuel cost changes, the FFS trend indicates that the cost per client would have been approximately $26.16 per trip. With an estimated 3.6 million trips in fiscal year 2015, the broker model costs were $118.6 million greater than the FFS trend. The cost difference for each year is shown in Figure 12. A comparison of the actual MTP costs between the broker model and an LBB staff estimate of what MTP would have cost if the FFS model had continued shows the relative expense of converting to a statewide broker model. MTP contracts were historically adjusted each year based on changes in the fuel cost index. To establish a comparable baseline from the broker model to the FFS model, LBB staff adjusted trip costs using fuel cost changes. Figure 11 shows adjustments to the actual MTP cost per trip in fiscal year 2000 using subsequent fuel cost changes. Adjusting for fuel costs accounts for nearly all variations in the actual annual cost per trip through fiscal year 2011. Variations in the cost of transportation fuel establish a trend of what the program might have cost if it had continued using the FFS model through fiscal year 2016. Total actual spending on MTP for fiscal year 2016 decreased due to a decrease in capitation rates associated with HHSC’s parental accompaniment rule. MTP actuaries decreased rates for fiscal year 2016 to account for decreasing utilization in South Texas (Region 10) observed during fiscal year 2015. However, fuel costs were also decreasing simultaneously. This fuel cost decrease was not incorporated into the fiscal year 2016 rates. As a result, the estimated cost difference grew in fiscal year 2016 despite the rate cut. In total, from fiscal years 2012 to 2016, the broker model cost approximately $316.5 During fiscal year 2011, the final year of FFS-only operations, the program spent $22.31 per trip. During fiscal year 2013, the first full year of broker operations in the Houston and Dallas regions, costs in the program rose to $39.38 per trip, FIGURE 11 MEDICAL TRANSPORTATION PROGRAM COST PER TRIP, FISCAL YEARS 2000 TO 2016 $70 $60 $50 $52.35 $40 $30 $20 $22.31 $10 Fee-for-service Broker statewide Broker pilots $0 2000 2001 2002 2003 2004 2005 2006 Actual Cost Per Trip 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Modeled trip cost using 2000 base year with fuel cost adjustments N඗ගඍඛ: (1) Modeled costs use fiscal year 2000 as the base year with adjustments for cost per client based only on fuel cost changes in subsequent years. Each year is adjusted based on fuel costs changes using a one-year lag. (2) Pilots of the broker model in the Houston and Dallas regions started during fiscal year 2012. The broker model was expanded statewide at the beginning of fiscal year 2015. (3) Onetime settlement costs from the Frew vs. Executive Commissioner of Texas Health and Human Services Commission lawsuit in fiscal year 2011 were excluded from these calculations. (4) Brokers underreported claims in fiscal year 2015. Adjustments to these trip totals are shown using an inflation factor estimated based on actuary estimates of missing claims. Trip volumes for fiscal year 2016 were assumed to be equivalent to fiscal year 2015 for this analysis. S඗ඝකඋඍඛ: Texas Health and Human Services Commission; Texas Department of Transportation; Texas Department of Health; U.S. Bureau of Labor Statistics. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 203 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 12 MEDICAL TRANSPORTATION PROGRAM COST TRENDS, ALL FUNDS, FISCAL YEARS 2011 TO 2016 (IN MILLIONS) COST SCENARIO 2012 2013 2014 2015 2016 TOTAL Actual Expenditures $206.4 $190.8 $159.7 $211.9 $186.8 $1,181.3 FFS Trend $221.0 $136.7 $121.5 $93.3 $66.6 $639.1 Cost Difference ($14.6) $54.1 $38.2 $118.5 $120.2 $316.5 N඗ගඍඛ: (1) Fiscal year 2011 is used as the base year to estimate the fee-for-service (FFS) trend and is adjusted for fuel cost and utilization in subsequent years. (2) Pilots of the broker model in the Houston and Dallas regions started in March and April of fiscal year 2012. The broker model was expanded statewide at the beginning of fiscal year 2015. (3) Onetime settlement costs of $8.7 million from the Frew vs. Executive Commissioner of Texas Health and Human Services Commission lawsuit in fiscal year 2011 were excluded from these calculations. (4) Subsequent refunds due to profit caps (known as experience rebates) reduced net expenditures for fiscal years 2014 and 2015 and are accounted for in these totals. Fiscal years 2015 and 2016 totals were adjusted to account for underreported claims, estimated to be 10.0 percent of trips. Fiscal year 2016 estimates assume the same number of trips as during the previous year. S඗ඝකඋඍ: Legislative Budget Board. million more than estimated expenditures would have been using the FFS model. administering the program increased from $22.9 million to $60.4 million from fiscal years 2011 to 2016. This cost difference is likely to continue in subsequent years. The proposed rates for fiscal year 2017 include a change equivalent to a $3.2 million overall decrease to the rates for brokers in fiscal year 2016. This change is based on fiscal year 2016 enrollment and excludes the North Texas region, which has temporarily converted to a fee-for-service model after the vendor contract was terminated. Actuaries include a 5.0 percent inflation factor in each year’s rates. Sometimes, such as for fiscal year 2017, this inflation adjustment is offset by other adjustments. Therefore, the cost difference shown in Figure 12 is likely to continue, considering the decrease in rates for fiscal year 2017 and the customary 5.0 percent inflation factor. This cost increase occurred during a period of decreasing numbers of users and claims. As a result, total costs for administering the program increased from 11.1 percent of claims in fiscal year 2011 to 47.8 percent of claims in fiscal year 2016. Fixed costs for providing MTP services serve as a base to overall administrative costs for the program. As a result, an increase in administrative costs relative to claims may have occurred in accordance with FFS operations as utilization decreased. However, from fiscal years 1999 to 2011, the ratio of administrative costs to claims never exceeded 21.1 percent. As Figure 13 shows, if the program had maintained an administrative expense ratio of 21.1 percent relative to claims, HHSC could have avoided $48.7 million in MTP administrative costs for fiscal year 2016. For fiscal year 2017, actuaries set administrative costs using the same percentage and fixed fee as the previous year. This scenario indicates that administrative costs associated with the model in accordance with actuarially set rates are likely stable and inherent to the structure of the model and contract. Cost growth in the broker model can be attributed to two components: administrative cost growth and an increase in rates paid to transportation providers. LBB staff estimates for fiscal year 2016 indicate that costs for the broker model were $120.2 million greater than the cost of operating the program in the FFS model. The administrative cost growth represents approximately $48.7 million of this cost difference. As a result of the transition of NEMT services to the broker model, HHSC decreased the number of full-time-equivalent (FTE) positions administering the program from 355.3 positions in February 2011 to 108.0 positions in September 2015. However, the reduction of 247.3 FTE positions at HHSC was offset by 580.0 new administrative positions at the transportation brokers (nonstate staff). Consistent with this net increase of 332.7 positions, total costs for 204 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 Although $48.7 million of the $120.2 million cost difference for fiscal year 2016 is attributable to administrative costs, the remaining $71.4 million is due to costs paid directly to transportation providers and Medicaid clients for NEMT services. During fiscal year 2015, for example, brokers paid subcontractors 2.1 percent more than was paid in the FFS model during the previous year. This increase occurred despite a 35.9 percent decrease in the cost of fuel during the same period. In the FFS model, rates paid to transportation providers would have accounted for decreases in fuel costs. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS FIGURE 13 COMPONENTS OF COST GROWTH IN THE MEDICAL TRANSPORTATION PROGRAM, FISCAL YEAR 2016 (IN MILLIONS) COST SCENARIO RATIO OF ADMINISTRATIVE COSTS TO CLAIMS HHSC ADMINISTRATION BROKER ADMINISTRATION ADMINISTRATION SUBTOTAL CLAIMS TOTAL $11.6 N/A $11.6 $55.0 $66.6 21.1% $7.6 $52.7 $60.4 $126.4 $186.8 47.8% ($4.0) $52.7 $48.7 $71.4 $120.2 26.7% Estimated 2016 FFS Costs Estimated 2016 Broker Costs Cost Difference N඗ගඍඛ: (1) HHSC = Health and Human Services Commission; FFS = fee-for-service Medicaid model. (2) Estimated fiscal year 2016 broker costs are based on the capitation rates paid to brokers and the number of clients using Medical Transportation Program (MTP) services. Fiscal year 2016 estimates assume the same number of clients use MTP services as during the previous year. S඗ඝකඋඍ: Legislative Budget Board. However, in the broker model, rates are not adjusted for fuel cost changes unless the brokers make rate reductions paid to their transportation providers. Additional reasons for cost growth may include increases in reimbursement rates for other types of services and adjustments in the mix of transportation modes provided. According to HHSC, brokers “cost less than FFS when considering all operating costs.” This determination is based on a comparison of MTP costs to the number of people enrolled in all Medicaid programs, which is the methodology used to establish rates for the transportation brokers. This cost measure for MTP tracks the average cost per recipient per month. The methodology used to calculate this cost is shown in Figure 14 and is based on Medicaid case growth, which regularly increases, and does not reflect the actual volume of MTP services provided. Instead, it divides costs for 154,266 MTP clients into a caseload of 4.1 million Medicaid enrollees. FIGURE 14 MEDICAL TRANSPORTATION PROGRAM (MTP) COST ANALYSIS METHODS, FISCAL YEAR 2016 PROPOSED COST PER MTP TRIP METHOD COST PER CLIENT METHOD MTP Expenditures / MTP Trips Provided = Cost Per MTP Trip MTP Expenditures / Medicaid Clients = Cost Per Client S඗ඝකඋඍ: Legislative Budget Board. HHSC does not regularly track and report the cost efficiency of the program using cost per trip. Option 3 would include a rider in the 2018–19 General Appropriations Bill to require HHSC to report cost per trip annually to the Legislature and post a copy of the report on the agency’s website. This cost would be calculated by dividing MTP expenditures by the LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 number of MTP trips provided. When considered with fuel cost changes, cost per trip provides a reliable way to measure the cost efficiency of the program and facilitates benchmarking performance against historical trends. PROGRAM INTEGRITY RISKS As HHSC has transitioned to a broker model, many brokers have had problems properly submitting documentation regarding claims paid to HHSC. In addition, HHSC has not collected adequate performance data for key risks to clients and the state. Without this data, HHSC cannot identify key program integrity risks. Although HHSC has taken some steps to mitigate risks when providing NEMT services, risks remain. Audits conducted in the FFS system repeatedly identified noncompliance with program rules. In a federal audit based on field work conducted during fiscal year 2011, for example, most claims had a deficiency. The audit sampled claims and found examples of the following: • providers unable to verify vehicles used for transportation and whether the vehicles had current state registrations and inspections at the time service was provided; • children were transported without a parent or legal guardian; • an inability to verify that computerized criminal history background checks, drug testing, and driver history checks on drivers had been completed; • Medicaid-covered healthcare services were not provided to a beneficiary on the transportation date; • cancellation of a transportation request before the service was received; and LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 205 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS documentation was not available to support the NEMT services. as of January 2016, contracts in 11 of the 12 regions operated by brokers pose a high risk to the state. The federal audit recommended that the state refund $30.4 million to the federal government for these improper payments. An internal audit at HHSC conducted during a similar period corroborated these conclusions. HHSC’s audit report found that the program did not have adequate controls to ensure services were provided only to eligible clients. Another recurring issue in audits of the program has been the extent to which clients received transportation services unrelated to medical appointments. LBB staff provided recommendations to HHSC to address these issues. In recommending reprocuring MTP services, LBB staff included recommendations to ensure that past performance is adequately factored into the evaluation of vendor proposals. LBB staff also recommended that HHSC determine whether evaluation criteria should include preferences for diversifying the number of vendors awarded contracts and develop a contingency plan for regions where qualified vendors are not available. Based on these audits, HHSC established a process to match transportation claims with medical claims data. This process facilitates the identification of potential fraud and misuse of program resources, and HHSC has applied it to broker claims data. For the fourth quarter of fiscal year 2015, transportation claims matched medical claims at a rate varying from 87.3 percent to 97.6 percent at each broker. Brokers were required to sample a portion of the unmatched claims and provide proof that the transportation was related to a covered healthcare event. ESTABLISH FLEXIBILITY TO USE MOST COST-EFFECTIVE MTP MODEL The statutory requirement to use a broker model for MTP statewide limits HHSC’s ability to develop the program to control costs and ensure quality services. As discussed previously, LBB staff analysis shows that the broker model costs approximately $120.2 million per year more than FFS trends. This cost difference exists at a time of historically low utilization; from fiscal years 2011 to 2015, the number of program users decreased from 346,147 to 154,266. Complaints from MTP users have more than tripled since fiscal year 2011, and many clients are likely having difficulty arriving at medical appointments due to transportation barriers. • However, this analysis was partly impaired by the amount of data missing from broker submissions. The contract with brokers states that “the state will not recognize or pay services that cannot be properly substantiated by the MTO and verified by HHSC,” and includes requirements to maintain claims records. In June 2015, however, the actuaries found that brokers were not submitting encounter data correctly. As a result, the data could not be used to set rates. Therefore, the actuaries relied on older FFS data from fiscal year 2014 to set fiscal year 2016 broker rates. This missing data limits the agency’s ability to identify potential fraud in the medical claims matching process. Brokers have also had difficulty complying with provider credentialing and documentation requirements. In August 2015, HHSC fined every vendor for contract violations related to network adequacy and client safety issues. The letters assessing liquidated damages sometimes indicated that deficiencies were placing clients at risk of imminent harm. Additionally, four of the five brokers did not meet driver training, drug testing, or sex offender registry check requirements. These problems resulted in some contract terminations. HHSC canceled two contracts due to client safety risks, nonperformance, and financial issues. According to HHSC, 206 LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 The outcomes of a future reprocurement are uncertain. It remains unclear whether a competitive broker market exists for cost-effective, high-quality services in Texas. Three considerations suggest potential limitations in the market for broker services: 1. According to HHSC, as of January 2016, contracts for 11 of the 12 regions pose a high risk to the state in performance and viability; 2. According to HHSC, there is a risk that requiring vendors to bid on price will result in higher prices. A number of states have moved away from competitive bidding for medical contracts to rely on state-set rates. According to the U.S. Medicaid and CHIP Payment and Access Commission, states may rely on directrate setting to achieve lower rates than they may be forced to accept from bidding; and 3. MTP contracts with a variety of vendors; however, one vendor has consistently reduced complaints across regions. This vendor paid substantially higher rates to transportation providers in most of its regions, suggesting a negative cost impact to the state for any increased reliance on this vendor. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 IMPROVE THE COST-EFFECTIVENESS OF MEDICAL TRANSPORTATION SERVICES FOR TEXAS MEDICAID CLIENTS Therefore, even if the MTP contract is reprocured with vendors competing on price, it is unclear if vendors will be able to bid less than the current rates while providing quality services. Option 4 would amend statute to authorize HHSC to use the most cost-effective model for administering the program with a sufficient level of service in each region of the state. This authorization would give the agency the ability to use either the broker model or FFS model. The optionwould require the agency to evaluate the cost-effectiveness of each model during future reprocurements to manage MTP services. This would align MTP with current requirements for medical services. To implement Option 4, the agency would be required to conduct a needs assessment before any reprocurement of broker services. This needs assessment would establish a cost threshold to determine whether to accept bids from qualified vendors. This threshold would be used in conjunction with an assessment of quality to evaluate cost-effectiveness. Option 4 would also include a contingency rider in the 2018–19 General Appropriations Bill requiring the agency to obtain LBB approval for this cost threshold during the biennium. If the agency, for example, determined that soliciting broker contracts presented an option for costeffective delivery of services in a region, the agency would be required to determine a cost threshold for bids from vendors. This threshold would be reviewed and approved by the LBB. After approval, the agency could solicit proposals and bids. If bids from qualified vendors failed to meet or fall below the pre-specified cost threshold, HHSC would not award a contract for the region and could instead operate the region through a FFS model. Option 4 would establish whether transportation brokers are capable of meeting or exceeding the performance of the FFS system in each regions of the state. It would also ensure that HHSC utilizes the most cost-effective MTP model for Texas Medicaid. FISCAL IMPACT OF THE OPTIONS Option 1 would require HHSC to report if there are increases in the percentage of Medicaid members with unmet transportation needs and develop a corrective action plan for these increases. If unmet transportation needs increase and HHSC implements a corrective action plan, this could affect utilization of the program. Over time, utilization increases may also increase costs but these costs could be offset by avoiding emergency care or preventable hospitalization. LEGISLATIVE BUDGET BOARD STAFF – JANUARY 2017 Option 2 would require HHSC to consolidate MTP surveys. HHSC would be required to directly contract for an annual survey. Multiple surveys are conducted quarterly by vendors that contract with brokers. This option is not expected to have a significant fiscal impact. HHSC has the option to reduce premium rates paid to vendors that conduct surveys to offset any costs the agency incurs to assume this responsibility. Option 3 would require HHSC to measure and report on the cost per trip of the program to the Legislature. This option would have no significant fiscal impact. Option 4 would amend statute to authorize HHSC to use the most cost-effective model of delivering MTP services. The agency would also be required to evaluate the costeffectiveness of the broker model and establish a cost threshold for any bids in a broker procurement that would be approved by the LBB. This evaluation would ensure that a reprocurement can meet the historical cost-effectiveness of the FFS system. Estimates indicate that expenditures using brokers cost approximately $120.2 million more than expenditures would be using FFS in all regions of the state. State administration of the program could result in reduced administrative costs. Rates paid to transportation providers could also be adjusted to align provider rates with fuel costs. Due to potential negative effects on network adequacy, it would likely be necessary to incrementally implement provider rate adjustments in a transition to a FFS system. The timing of these reductions would impact the cost savings each year. Considering the cost difference between FFS and brokers, brokers would need to bid at rates lower than the current costs per trip to meet the requirements of Option 4. Because it is not possible to anticipate how many or which regions may be transitioned to a FFS model, or what the cost threshold for a broker would be, the fiscal impact of Option 4 cannot be estimated. However, it is expected this option would reduce costs per trip. The House introduced 2018–19 General Appropriations Bill includes a rider implementing Options 1 and 3. LEGISLATIVE BUDGET BOARD STAFF REPORTS – ID: 3729 207