Last week we included a comparative performance chart similar to the one below. Today’s chart bears some explanation.
In our forthcoming ACO Performance report, one of the ways we examine the data is by looking for connections among quality metrics. For example, do ACOs with high scores for the Diabetes Composite metric have lower hospital admissions for heart failure? We cannot make a statistical association here as we would need the individual patient records to do this. But we can look for patterns and visualize the data in other ways.
In the chart below, we see that ACOs that have a high percentage of physicians qualifying for either a Medicare or Medicaid EHR Program performance payment tend to have lower readmission rates. This makes intuitive sense: the use of EHR technology should help providers keep track of their difficult patients and prevent problems before they happen.
But three of the worst performing Pioneers for hospital readmissions—Michigan, Montefiore and Steward—were among the top performers for shared savings. How can this be? Are they sacrificing quality for savings?
“Measuring success by shared savings can lead down a garden path,” one source told us, because it’s really a measure of how well the ACO performed based on the forecasting methodology, which is based on a national trend. Also skewing the numbers are reimbursement and pricing adjustments, which factor in the cost of care for low income residents and for the cost of teaching residents.
This is where we turn to the program evaluation report, which uses a control group of similar nearby providers and accounts for preceding trends in its forecasting model. According to the report, Montefiore had savings in year 1 and losses in year 2, but neither was statistically significant—which means to date effectively they are a wash for shared savings. Steward and Michigan, on the other hand, had statistically significant savings in both years.
And this is why using readmission rates alone to evaluate performance may not tell the complete story. Montefiore did in fact improve its readmission rate from year 1 to year 2 by nearly a full percentage point. Montefiore simply may have a more difficult population to treat than the average Pioneer.
Conversely, Appleton, Wis.-based Bellin-ThedaCare fared slightly worse on readmissions from year one to year two, but its rate is still below the national average. It had shared savings in both years too, measured by the raw numbers and by using the evaluation report’s forecasting model. Bellin-ThedaCare is also the top performing ACO on quality metrics: it ranks in the top five for each measure more than any other Pioneer.
Considering our readership—primarily pharma/biotech and medical device executives—it is fair to ask why we are obsessing over ACO performance metrics. For one, it is important for these executives to track the progress of these CMS initiatives because if they continue to grow, more resources will be required to cover them. The inverse of this is also true.
Second, many of our clients focus existing resources by targeting on the strengths and weaknesses of each customer. For example, a company with a strong diabetes product portfolio may seek to partner with ACOs and Integrated Health Systems with a substantial diabetic population, or with those struggling to get their diabetic population under control. Understanding how best to allocate scarce market access resources is one of the things we do for our clients.
In the evaluation report there are several key findings to date, which we conclude with here:
Pioneers that had shared savings in both years tended to have larger reductions in inpatient stays, tests, imaging and procedures compared to those that had no savings or savings in only one year. All Pioneers reduced ER visits for the first two performance years.
Researchers found no evidence of a systemic reduction in unplanned 30-day readmissions for the Pioneers versus controls.
For all Prevention Quality Indicators, there were statistically significant reductions in admissions in both years, especially for COPD, asthma and heart failure in year 2. There was an unexplained increase in admissions for heart failure in year 1.
Pioneers had better post-discharge physician followup than controls.
Data sharing continues to be a problem for Pioneers: only six reported interoperability across EHRs and most are challenged by multiple EHR systems. This finding is consistent with our own research within Darwin’s ACO panel.