Pioneer ACO model saves $385M in first two years

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - ACO

An accountable care organization (ACO) model saved Medicare approximately $385 million in its first two years, according to an independent report released on May 4.

Lead author David J. Nyweide, PhD, of the Centers for Medicare & Medicaid Services (CMS), and colleagues also released their findings online in the New England Journal of Medicine.

CMS launched the Pioneer ACO model in 2012 as a payment model focused on making providers collaborate to cut costs and increase the quality of care. Participants who spend below a certain amount of money are eligible to receive a bonus and share in the savings based on their performance on 33 quality measures.

This year, more than 600,000 Medicare beneficiaries are served through Pioneer ACOs.

In this analysis, the researchers examined the 32 Pioneer ACOs that participated in the program in 2012 and 2013, including nine that stopped participating in the second half of 2013. During those years, more than 1.4 million beneficiaries received care through the Pioneer ACO model.

The researchers found beneficiaries in Pioneer ACOs saved $280 million in 2012 and $105 million in 2013 compared with those who were in fee-for-service programs. They said there was no difference in all-cause, 30-day readmissions, although Pioneer ACO beneficiaries had more follow-up visits after they were discharged from the hospital.

The study had some limitations, according to the researchers. They noted that CMS selected the Pioneer ACO model participants based on their ability to manage patients’ care differently from the typical fee-for-service model. In addition, there was no control group, which may have led to unmeasured differences between the ACOs and other models.

Further, the total spending did not include Medicare Part D prescription drug spending or cost-sharing payments by beneficiaries. The response rate to the ACO survey was also only 52.8 percent. The study did not include data from the nonresponders.