On average, risk taking ACOs generated savings of roughly $5M more than non-risk taking ACOs ($11.72M v. $6.65M).  This translated into earned savings of $7.33M v. $2.82M for risk taking and non-risk taking ACOs, respectively, despite the smaller number of risk taking ACOs. 

The number of beneficiaries and total expense were not statistically different between the two ACO types.  If you consider the MedPac estimate of 2% of beneficiary expense in added overhead to operate an ACO, the $5M difference in earned savings becomes quite significant. 

It's clear. Whatever happens in the ACO as a result of the act of taking risk--whether there are structural or process changes--risk taking ACOs save more money than non-risk taking ACOs. 

Reading between the lines, however, strongly suggests that entities working in capitated or global budget arrangements might have even more opportunities for financial returns because the issue of demand destruction in full-risk v. largely FFS with some downside risk environments is different. 

If an ACO is going to generate savings it has to spend less than the budget.  The math is simple.  Since ACOs use existing Medicare payment methodologies under Parts A and B, and these are largely prospective payment systems, generating savings may well mean the providers in the ACO are capturing less volume-driven revenue.  The difference in that demand destruction between risk taking and non-risk taking ACOs was $124 per capita.  In other words, risk taking ACOs “destroyed” $124 more in revenue per beneficiary than non-risk taking ACOs. 

Consider an “average” ACO with 20,000 attributed beneficiaries and a total medical expense of $20M.  If the risk taking ACOs generated on average $7.33M in shared savings but spent $4M to operate the ACO and then reduced per beneficiary expenses by $124 more than non-risk ACOs, the total of $6.2M between those factors essentially leaves the ACO with a $1M surplus. 

Our findings mean the number one driver of ACO success is committing to risk—taking meaningful, downside risk is more likely to drive financial success than spending more time on the “glide path” while dabbling in upside risk.

Our univariate analysis revealed some significant differences between the two ACO types in terms of types of expenses and visits. When it comes to broad expense categories, the risk taking ACOs spent less on IP and OP (defined as HOPD) and more on physician services:

Some of the variation in these expense categories may be attributable to univariate differences in broad encounter types. The risk taking ACOs had fewer “expensive” visit types but more PCP E&M encounters:

In the performance data published in July 2019, risk taking ACOs had a statistically lower number of all-cause unplanned admissions. Medicare did not publish this metric in the January 2020 PUF, updated August 2021. However, the difference on this measure between the ACO types is consistent with the differences in expense and visit categories in the January 2020 PUF.

This finding is consistent with previous research: successful ACOs manage patient care in lower cost settings.

The two ACO types were not statistically different in terms of PCPs and specialists. The overall quality scores were not statistically different either. However, the risk taking ACOs statistically outperformed the non-risk taking positively on the following quality measures:

• Discharges for AHRQ prevention quality indicators (dehydration, UTI, bacterial pneumonia)
• Fall risk screening
• Influenza immunization
• Tobacco use/screening
• Depression screening
• Colorectal cancer screening
• Breast cancer screening
• HbA1c poor control
• Controlling HBP

If the better performance of the risk taking ACOs on these quality measures does not scream primary care, then it is difficult to know what else would. The statistical difference in the mid-year file regarding all-cause unplanned admissions highlights the impact of primary care. Our findings suggest it’s not that successful ACOs have more PCPs. Rather, successful ACOs measure, document, and manage care better in primary care than their counterparts.