Blue calculated and compared to the ascertained
Blue Cross Blue Shield (BCBS) ofMassachusetts (BCBSM), a private insurance provider, launched the AlternativeQuality Contract (AQC) in 2009. Originally encouraged by the Affordable CareAct (ACA) and beginning with 7 participating provider organizations, the goalof the AQC was to improve quality, lower costs, and slow the exponential growthof healthcare spending.1As of now, the AQC only covers BCBS HealthMaintenance Organization (HMO) patients and Point-of-Service enrollees.
Providerorganizations who give primary care for at least 5,000 enrollees in the BCBSMHMO or POS plans can enter into this voluntary 5-year contracts as “AQC groups”.AQC groups decide an annual budget for their attributed patients which includesall health care costs the patient will receive that year. The budget varies pergroup and is initially based on attributed patients’ prior total healthcare costs.After the first year, the budget is adjusted for patients’ health status andinflation, again varying by group. Providers receive fee-for-service payments(FFS), as in a traditional healthcare model, however, at the end of the year, totalfees are calculated and compared to the ascertained target. If total spendingis below budget, the group can receive a shared savings payment from the insurer.However, if spending is above target, since this is a shared-risk model, theAQC must pay BCBSM a share of the overdraft. During the first two years, AQCgroups can choose to take on 50-100% up-side or down-side risk and are eligiblefor an additional 10% of their total spending target if certain qualitymeasures were met.
In year three, the shares of savings and losses isdetermined based on previous performance on quality measures where highperformance allows for a greater share in savings and a smaller share of risk; the10% bonus is removed in year three. Sixty-four quality measures (QM) are inplace as a complimentary performance metric to spending. These QMs adjust the size of shared savings orpenalties owed. Thus, quality cannot suffer at the expense of cost.1For a patient, by designating aprimary care provider (PCP) who is a part of the ACQ, a patient’s care becomesattributed to that physician’s budget. Enrollee participation is not voluntaryor decided by them; in fact, BCBS members are not told if their PCP is even enrolledin the AQC. However, their benefits should remain identical to those of otherpatients in an HMO or POS – able to receive care from any in-network providersif referred.
This is could be classified as a “commercial HMO” model.1In agreeance with the ACA, the AQChas met its goals of cost reduction without sacrificing quality. It is now oneof the largest private payment reform initiatives in the U.S with 90% ofphysicians in BCBSM’s HMO network participating.
However, the plan has hadfar-reaching impacts that have affected multiple stakeholders differently. Whatfollows is an analysis of how the AQC has affected the cost and quality of servicesas they pertain to patient-consumers, AQC physicians and AQC hospitals. Spill-overeffects on CMS, a different payer population, will also be examined.
The effect on providers has beenvaried due to the diversity of participating groups, but the LA Times captured the sentiment well inan interview with a participating physician: “That kind of attention has alwaysbeen good medicine. For Dr. Folch, 59, it’s now good business.”3 (page 2)The quote above comes from an AQC PCP office in suburban Boston that will notlet a patient with high cholesterol leave the office without a scheduled eyeexam which is key prevention. To review, providers were given an annual budgetbased on attributed patient’s historical per member per month spending withbonuses the first two year for meeting quality metrics. BCBSM provides regularreports on quality, service use and spending. Some participating groups areexperienced due to previous contracts with global-risk whereas others are newto this arrangement.
Regarding both cost and quality, AQC groups prioritizequality improvement so as to increase financial returns. Furthermore, becausetargets for quality measures are absolute numbers, rather than a relativeranking compared to peers, providers face no disincentive to share bestpractices with each other. A 2006 review of 27 FFS programs saw the averagephysician performance bonuses were only 2.3% of their total payments whereasthe AQC offers up to 10% of total spending.
Since groups can direct the bonuseswithin their provider network, they can specifically incentivize physicians.For example, because PCP salaries represent less than 10% of total spending,groups that received a full bonus for ambulatory care quality measure (half ofthe 64 measures, thus 5% of the total 10% bonus) could then increase PCPcompensation by greater than 50% above their existing BCBS fee schedule. Thisoffers a competitive advantage for groups that receive lower rates from BCBSand thus have relatively lower physician salaries. These groups are at risk oflosing physicians to organizations on the higher end of the fee scale, howeverparticipation in the AQC gave them an opportunity to attract physicians withpay. Physicians were given the opportunity to improve quality in real-time withregular feedback on quality scores and data on specific patient follow-upmishaps such as chronic and preventative care services as well as assistance incontacting patients. BCBS pay some groups up to 2% of their annual budget topromote infrastructure development. Along with linking physician compensationto performance, referral management was among the largest changes forphysicians. Groups tried to control for “leakage” or the percentage of patientsreferred outside of the group’s network and directing patients to lessexpensive care facilities.
Some groups even considered expanding their in-housespecialists to expedite referrals. In terms of ambulatory care quality measures,after two years the contract correlated with a 3.7% increase per member peryear among participating members. In a NEJM study comparing 4-year AQC resultsfrom 2009-2012 to those among persons in control states, Song et al. found thatmedical spending on claims grew an average of $62.21 less per enrollee perquarter than it did in a control cohort.
The 2009 medical spending average forgroups in the AQC did increase, it rose by 2% less than average spending forBCBS HMO enrollees with providers not in the contract. Not surprisingly thestrongest performance was accomplished by groups new to global-risk and thushad the most changes to make; these groups medical spending increased by 6.3%less than the average of non-AQC groups. Most savings came from the change in referralpatterns where there was a shift to providers with lower outpatient facilityfees.2,3,4 Asfor patients, the recipients of care, enrollee cost-sharing is not affected bythe AQC or if their provider is in an AQC group.
Therefore, enrollees do nothave financial disincentives to see providers outside of the AQC. However, sinceas discussed above, PCPs may be incentivized to direct referrals in-house or tolower-cost facilities, possibly causing friction between patients and theirprimary providers. Patients are able to switch PCPs at any time and are notlocked in to provider in the network.
In terms of quality, since a key part ofthe AQC is linking quality metrics to savings, patients are now tracked to getthe right screening exams, if they are satisfied with their care, and if theyget the right drugs and so on. This model penalizes physicians who withholdcare in the interest of cost-savings. Patients seen by groups in the AQC werehospitalized less and used fewer expensive services like imaging with noevidence that reduced utilization compromises the quality of care.
The AQC madestrides in quality when it came to coordinating care for high-risk enrollees:those with multiple chronic diseases who are at highest risk for rapiddeterioration and thus costly hospitalization. Patient-care improved withmultidisciplinary approaches to coordinating care, using programs likeStatusOne and Guided Care. To reduce avoidable hospital admissions andreadmissions and emergency visits, one group contacted all patients who weredischarged from the hospital to ensure they understood their instructions, werecompliant with medications, and had access to support services and were notexperiencing any complications.
On a composite of 5 Health Effective Data and InformationSet (HEDIS) measures for Diabetes and Cardiovascular disease, the 2009 AQCimproved health outcomes 12% above the HEDIS national average. An interestingstudy in Health Affairs 2017 by Song et al. found that while quality improvedfor all enrollees whose providers were a part of the AQC, process measuresimproved 1.
2% per year more for enrollees in areas with lower socioeconomicstatus (SES) as when compared to patients of high SES. There were no differencesin outcomes between groups or spending. This suggests that the AQC isunintentionally narrowing disparities among patients. Providers may beincentivized to focus on improving quality for historically medically-neglectedpopulations as they bear the most potential for quality improvement sand savings.
However, in a 2015 report by Avalere, it was found that quality metrics did notimprove on measures not tied to payment, unfortunately demonstrating the needfor incentives.