Skip To Content

Voices On Value RSS Feed

There’s More Than One Way to Slice an Apple – and to Calculate Readmissions

There’s a lot of data out there, folks. Tons, in fact. One of the statistics you hear a lot about these days is hospital readmission rates. Hospitals have a lot riding on their readmissions rates and many projects across the country focus on reducing readmissions.  Hospitals have access to their own readmissions data, but that’s typically where their information ends. The Colorado All Payer Claims Database now has new ways of looking at readmissions that will give hospitals and health systems deeper insights into readmissions across the state and in their own facilities like never before. Separate Medicaid and commercial insurance readmission rates are also now publicly available for comparison across Colorado – giving us the ability to break down readmissions by coverage type, service line and geography. This level of data transparency is a critical part of eliminating unnecessary hospital readmissions for a safer, healthier, and less expensive health care system.

The 30-day all-cause readmissions (30 day ACR) calculation really entered the spotlight in 2009 when Jencks, Williams and Coleman published results showing that almost one in five Medicare beneficiaries was returning to the hospital less than 30 days after being discharged. Soon after, as a part of the Affordable Care Act under the Hospital Readmissions Reduction Program, CMS started penalizing hospitals with higher than expected readmissions through a negative payment adjustment. As a result, most hospitals and health care facilities have been putting a lot of work into reducing readmissions and improving care transitions to other facilities in order to keep people from coming back to the hospital unnecessarily. In fact, CIVHC is working with CFMC, the Colorado Hospital Association, the Colorado Health Care Association and the Colorado Rural Health Center on a statewide campaign called Healthy Transitions Colorado ( to support readmission reduction efforts across the state.

As the CMS penalties increase, having accurate information about 30-day ACR is critical. Typically, 30-day ACR is calculated by looking at the number of readmissions within 30 days divided by the number of total hospital discharges. In essence, of those who left the hospital alive, how many came back within 30 days? There are a couple of problems with this method:

  1. Most hospitals or systems only have the ability to track whether a patient returned to the same facility or system. If a patient returns to a different hospital, that readmission isn’t recognized as a readmission by Medicaid or commercial payers. Because of this, it is difficult to know how many readmissions are occurring
  2. Efforts to reduce REadmissions frequently have the effect of also reducing ADmissions. When you reduce both the numerator and the denominator of a fraction, the resulting percentage may not change or may underestimate the real reduction in readmissions and in some cases where admissions are declining at a greater rate than readmissions the rate may increase. With CMS focusing solely on readmission percentages, this calculation can actually be a disincentive to reduce hospital admissions and can mask significant achievements and improvements in care.

The newest CO APCD data release addresses these issues and gives us a new way to look at 30-day ACR.The APCD gives us two ways to look at 30-day ACR, and it’s available for the Medicaid population, the commercially insured population, and both of those groups combined. The first method, 30-day all cause readmissions (by admission), shows 30-day ACR percentages much like CMS does (details are available on the website) but includes readmitted patients no matter what hospital they return to. This information is available in aggregate based on location of patient residence (county or ZIP Code 3) on the website, but can be custom ordered to track the patients for a specific hospital or health system regardless of where they go after discharge. This gives hospitals and health systems important information about their patients after discharge that they’ve never had access to before.

As you can see below, this method also shows the percent readmissions by service line (how many cardiology patients are readmitted, how many pulmonology patients are readmitted, etc.), so hospitals can identify the specialties with the highest and lowest readmission rates.  While, this method improves the capture of readmissions across the state, it doesn’t solve the problem with the connection between readmissions and admissions.

In order to look at readmissions without admissions, the CO APCD also contains 30-day all cause readmissions (by population) which shows 30-day readmissions per 1000 insured residents. This method divorces readmissions from admissions by comparing readmissions alone against the total insured member population in the database, rather than against admissions. Using the readmissions per 1000 population metric allows us to look at how readmission efforts are working and see how those efforts are reducing readmissions, regardless of their impact on admissions. This is the same way admissions are reported in the APCD, by the population of insured individuals in the area.

As you can see below, this report also shows the breakdown of readmissions by service line – what percentage of readmissions came from each area. Higher readmissions rates are highlighted red, and lower readmission rates are colored green to allow for quick identification of outliers. Again, the per population method is an excellent way to get an idea of the impact of community interventions, but readmission rates can be diluted or overinflated due to large or small population sizes. 


To get a better idea of how readmissions vary across the state, there is also a state map of readmissions. The map displayed below shows readmissions for both commercially insured Coloradans and those on Medicaid combined, but isolated views for both payers are also available.

There’s no one perfect way to look at readmissions, but by combining these two methods, communities can get a clearer picture of the progress that is happening locally and across the state. Health care facilities and systems can order custom reports to analyze where their patients are going and if there have been readmissions sliding under the radar because of changing locations. This kind of comprehensive, detailed information is what will allow Colorado hospitals and facilities to make real changes in the quality of care that patients receive. Please contact us if you have any questions.

About the Author: Kristin Paulson is CIVHC's Senior Manager of Policy and Initiatives. Contact her at

Share on: DiggFacebookRedditDeliciousTwitter


Catherine Morrisey
Hello - I am wondering if there is additional insight into the methodology of how indigent populations are counted in this database. I am specifically trying to look at indigent populations with behavioral health issues, but do not have private or Medicaid insurance. These expenses would be absorbed by the individual hospitals (I believe). Do we know if these individuals are counted in the numbers?
12/13/2013 11:24:53 AM

Leave comment


 Security code