As previously described, Colorado, Minnesota, and Rhode Island have each created a unique payment model for their ACOs. All three states, however, consider behavioral health performance in ACO payments and have established policies that foster consideration of behavioral health performance in payments to individual providers.
ACO Payment
Payment is considered to be the driving force to ensure ACO accountability for performance. All three states have tied ACO payment to behavioral health performance. Colorado’s RAEs may earn two types of incentive payments for performance. A RAE may earn incentives of up to $4 PMPM in key performance indicator (KPI) payments, which are primarily tied to the performance of the PCMP network. The RAE may also earn up to 5 percent of its behavioral health capitation rate (about $30 million) in behavioral health incentive (BHI) payments, which are primarily tied to the performance of the behavioral health network.
In both Minnesota and Rhode Island, ACOs may earn a share of the savings they produce — and are sometimes required to also share in any losses. Both of these states calculate savings and losses using a TCOC methodology. Both also include behavioral health services in the TCOC calculation, and Minnesota considers behavioral health services when calculating the monthly population-based payment that each of its IHPs also receives.
Regardless of their specific ACO payment model, all three states tie ACO payment to performance on a set of metrics that includes behavioral health measures, along with measures of other types of performance. In Colorado, performance on the metrics governs the amount of the incentives the RAEs earn. In Minnesota, performance on the metrics governs whether Track 1 IHPs may continue to participate in the program and whether Track 2 IHPs earn the full portion of their contracted share of any savings they produce. In Rhode Island, performance on the metrics governs what portion of savings are paid to the AE and how much the MCO may retain (and for those AEs that have been approved to accept downside risk, what portion of the losses they are required to pay to the MCO).
“Measures [in Rhode Island] were shaped by high-level goals of the organization and we wanted to make sure that there are mechanisms for balanced, useful data.”
There is little commonality between the behavioral health measures these states chose to factor into payment (Table 1). Only one measure is used by all three states – follow-up appointments within X days after an inpatient hospital discharge for a mental health condition, although even on this measure, states have made different choices as to when to specify seven-day follow-up and when to specify 30-day follow-up. Although there was little commonality between the specific measures used, there was some commonality regarding the domains to be measured. All three states included measures of engagement and follow-up.
When asked why they chose their specific measures, state interviewees uniformly reported that the choices were based on their goals for the program and stakeholder input. They also reported that the choices were informed by previous efforts.
- Rhode Island, for example, drew its measures from a common measure set created as part of the state’s State Innovation Model (SIM) initiative.
- Colorado emphasized that it considered areas that it knew had “room for improvement” and used measurement to create incentives for the RAEs to change that part of the system. For example, the state tied payment to “follow-up after a positive depression screen” in order to incent behavioral health and primary care integration. Further, Colorado sought to align its behavioral health (BHI) and primary care (KPI) payments to incentivize the same behavior across the program. As one official explained, “behavioral health is not in a silo, and requires the engagement of physical health as well.”
- Finally, this measure set is not static, Minnesota, for example, is currently phasing out its medication measures.
Table 1: Behavioral Health Measures that Factor into ACO Payments
|
Colorado |
Minnesota |
Rhode Island |
| Screening and Assessment |
| Screening for depression and follow-up plan |
|
Some IHPs* |
Optional |
| Behavioral health screening or assessment for children in the foster care system |
BHI |
|
|
| Developmental screening in first three years of life |
|
|
Optional |
| Engagement and Follow-up |
| Initiation and engagement of alcohol and other drug abuse or dependence (AOD) treatment |
|
Required |
|
| Engagement in outpatient substance use disorder (SUD) treatment |
BHI |
|
|
| Behavioral health engagement: Percentage of members who received a behavioral health service delivered either in primary care settings or under the capitated behavioral health benefit within a 12-month evaluation period |
KPI |
|
|
| Follow-up appointment within 7 or 30 days After an inpatient hospital discharge for a mental health condition |
BHI; 7-day |
Track 1: 7- or 30-day
Track 2: 30-day |
Required, opt for 7- or 30-day |
| Follow-up appointment within 7 days after an emergency department (ED) visit for a substance use disorder |
BHI |
|
|
| Follow-up after a positive depression screen |
BHI |
Some IHPs* |
|
| Medication |
| Antidepressant medication management: acute and continuous |
|
Required |
|
| Adherence to antipsychotics for individuals with schizophrenia |
|
Required |
|
Acronyms: BHI=behavioral health incentive payment; KPI=key performance indicator payment; IHP=Integrated Health Partnership.
*The measure is used if it is appropriate to the topic of the health intervention that the IHP and Minnesota Medicaid jointly select.
Colorado’s RAEs have increased the percent of Medicaid enrollees who receive behavioral health care.
To date, only Colorado has published a summary of ACO performance on the measures that factor into incentive payments. This state has seen some positive results. On the primary care side, payment is tied to behavioral health engagement. During the first year of program operation, the RAEs produced almost a two-percentage point increase in this metric. On the behavioral health side, Colorado established incentives for performance on five measures.
During the program’s first year of operation:
- All seven RAEs met performance improvement goals for engagement in outpatient SUD treatment and follow-up after a positive depression screening;
- Five met the goal for behavioral health assessment for children in the foster care system; and
- Four met the goals for follow-up from mental health inpatient stays and SUD emergency room visits.[11]
Although not yet complete, Rhode Island is conducting an evaluation of its program which, when complete, will shed more light on its program’s impact on behavioral health care.
Minnesota officials report that IHPs perform above average when compared to clinics and/or systems that are not IHPs, on many quality measures, including those related to diabetes, asthma, and vascular care. In terms of behavioral health metrics, IHPs perform better than non-IHPs on ensuring adolescents are screened for mental health issues. Additionally, those IHPs with health interventions with a behavioral health focus have seen some positive results, including high levels of relative improvement across years.
Payment to Individual Providers
Colorado and Rhode Island both sought to ensure that the individual providers, including behavioral health providers, affiliated with their ACOs would benefit from performance payments earned by the ACO. Colorado explicitly requires its RAEs to share any incentive payments the RAE receives with the providers who form the RAE’s health neighborhood, which consists of the providers (including behavioral health providers) and facilities that work with an individual’s PCMP to meet all of the person’s health needs. RAEs are also required to report all payment arrangements, including those with behavioral health providers to the Medicaid agency. Rhode Island does not require its AEs to share savings, but it does require the AEs to have documented relationships with behavioral health provider partners to meet their patients’ needs, and Rhode Island Medicaid envisions that some of those relationships will include shared savings. It also requires the AEs to share 10 percent of any incentive pool payments they receive with a partner organization, which may be a behavioral health provider.
Minnesota encourages its IHPs to implement accountable care partnerships by potentially rewarding the IHP with more favorable financial arrangements. For example, instead of receiving 50 percent of any savings, the IHP might receive 70 percent. An IHP can use this mechanism to also reduce the share of any losses that it must pay, and Medicaid officials report that it is this attribute that is often most attractive to the IHP. The financial arrangements are negotiated with each IHP. Minnesota offered an example of a potential partnership between an IHP and a CMHC. The CMHC agrees to provide behavioral health services to a targeted set of patients who screen positive for a behavioral health need. IHP-affiliated providers will screen patients, refer those who screen positive to the CMHC, and the CMHC will provide any needed services and ‘close the loop’ by providing information back to the referring provider. Minnesota Medicaid also negotiates the specific measures that will be used to judge the success of the partnership. In this case, that might include the number of patients that were screened for behavioral health needs divided by the total number of patients, or it could also include the total patients who received services divided by the total patients referred. Although there is no requirement that partners be paid, eligibility for a more favorable risk arrangement depends on several factors that encourage payment, including how well the partners are working together.
Finally, when it launched its accountable care program, Colorado Medicaid began directly paying PCMPs (i.e., primary care providers) for short-term behavioral health treatment (up to six visits) on a fee-for-service basis. Other behavioral health services are paid for by the RAEs. Colorado Medicaid had anticipated that this change would foster behavioral health integration and support then-existing partnerships between some federally qualified health centers (FQHCs) and CMHCs. As expected, this policy did improve access for patients with more mild behavioral health conditions. However, some FQHCs chose to withdraw from the partnerships and hire staff to provide the short-term services. Colorado Medicaid also found the policy more difficult to implement than anticipated, as the agency had to adjust the payments it made to the RAEs for behavioral health services to reflect that payment for some short-term services that would now be made by the agency.