| Home / Blog
Medicare Readmission Rates Showed Meaningful Decline in 2012

Medicare Readmission Rates Showed Meaningful Decline in 2012

Posted by-Lawerslog
Member Since-29 Dec 2015


The Centers for Medicare & Medicaid Services have made it a top priority to reduce re-hospitalizations among Medicare beneficiaries. Readmissions to hospitals are an indicator of quality care and they account for billions in Medicare annual spending. CMS has taken several initiatives over the past several years to decrease readmissions in Medicare fee-for service (FFS) patients. This includes reporting hospital readmissions through Hospital Compare, funding hospital-level improvements through Partnership for Patients, and changing payment policies through Hospital Readmissions Reduction Program. There are also shared savings initiatives. Many hospitals and other organizations have also employed strategies to decrease readmissions. These include enhanced patient education, post-discharge follow up care, and better coordination with outpatient providers.

Partnership for Patients (P4P), a CMS-led public-private partnership, was launched in 2011 with the goals of improving patient care as well as reducing costs. Nearly 3,500 hospitals are part of the Partnership, which excludes children's hospitals. They account for about 700,000 or 75 percent of index admissions for FFS Medicare beneficiaries each monthly. The Partnership aims to lower the 30-day hospital readmission rate of 20% by 2013 and improve transitions of care.

This analysis examines the unadjusted monthly readmission rates of the nation within Dartmouth Hospital Referral regions (HRR) and compares participating and non-participating hospitals to the P4P program (overall, and by the number of inpatient beds at each facility). To examine long-term trends, we estimate the national readmissions rates between 2007 and 2012. To examine the geographical patterns in the country, we calculate the readmission rate within HRRs. We also compare rates between hospitals that are part of the P4P program. The goal of the program is to reduce readmissions.

Data and Methods

We used claims data from Chronic Condition Data Warehouse to calculate the all-cause rate of readmission. This warehouse contains 100 percent of Medicare claims for beneficiaries enrolled in FFS. All Medicare beneficiaries who are enrolled in Part A (including those below 65) and not enrolled under Medicare Advantage during the year constitute the study population.

Our observation unit was an inpatient stay at an acute hospital. This included critical access hospitals. Readmissions were defined as inpatient hospital stays that began within 30 days after discharge from an index admission. Readmission rates are the sum of the number readmissions for a given period divided by the index stays (here, one month). Inpatient stays in which the beneficiary dies were not included as index admissions. However, such stays can be considered readmissions. Same-day transfers were considered part of one stay. The 30-day period was started at the end. If the beneficiary died during the stay, readmission stays were treated as index admissions.

The readmission rate includes all clinical diagnoses. It also includes beneficiaries who were readmitted from a hospital other than where they were discharged. Index admissions are those in which a patient was discharged against medical advice. These results include readmissions that may have been planned ahead of time or could be considered unavoidable.

We present data that are not adjusted for age or health status. We also looked at rates across age cohorts (unreported). The results were not significantly affected by standardizing according to age.

Calculating readmission rates based on incomplete information

All Medicare inpatient hospital claims can take up to a year to process and finalize. Rates for months after March 2012 were based on data we have adjusted to account for this lack of information. We performed an extensive analysis of the claim maturity patterns from 2007 to 2011. This allowed us to calculate readmission rates. We used this analysis to calculate weightings for readmission stays and index stays. Additionally, we compared the interim data with final data for each metric each month that claims were being processed (processing Months). We multiplied interim monthly data with the appropriate weightings to estimate readmission rates for months that had less-than-complete information.

We did not have enough data to accurately and reliably determine the readmission rates in the two most recent months that we had claims data for. We can estimate the readmission rate of a month using three months worth of claims data. This is done by weighting historical claims data with the claims processed up to that point. Our analysis shows that information on nine percent of readmission stays is not processed in the third month after an index admission. We adjust the claims data that we have at that time to reflect this shortfall.

Our estimates of readmission rates for the most recent months are subjected to uncertainty because there are variations in how claims are processed and finalized. These months are based on our estimations of readmission rates. We compute a range. Based on claims maturity patterns from 2007 to 2011, this range should give us a range that covers about 95 percent of the monthly rates. After the seventh month of processing, more than 99 per cent of inpatient claims have been submitted. Variation has decreased to the point that we do not show readmission rates.


The FFS Medicare beneficiaries' 30-day all-cause hospital readmission rate remained at 19% from 2007 to 2011. The monthly readmission rate fell to 18 percent in October 2012 and was 18.4 percent on average for the year. This is more than half the percentage point less than the average rate over the past five years. This reduction in readmissions amounts to about 70,000 less readmissions than if the rate was unchanged at 19%. A t-test revealed that the 2012 average readmission rate (assuming September-December rates are correct) was significantly lower (P.0001). However, the same result was not achieved when t-tests were used for comparing rates on a year to year basis (e.g. The 2008-2007 period was compared to the 2007-2011 period.

Partnership for Patients Hospitals

The readmission rates of hospitals that participate in the P4P program are generally lower than those at non-participating facilities. The average Medicare readmission rate at all P4P hospitals was 19% from January 2010 to December 2011. This was compared to 19.2 percent for non-participating hospital hospitals. According to claims processed so far, both the participating and non-participating hospitals had lower readmission rates in 2012. They averaged 18.4 percent for participating hospitals and 18.6 per cent for non-participating. Similar to the national-level data readmission rates decreased at comparable levels for all age groups at both non-P4P and P4P hospitals.

Hospital Referral Region Rates

We also looked at readmission rates for the 306 HRRs as defined by Dartmouth Atlas. Unadjusted readmissions rates can vary greatly across the country. This is evident, we found. The readmission rates for 2011 were lower in the Mountain West than in the Pacific Northwest, as shown in. The rates were highest in the Mid-Atlantic, eastern Midwest, certain parts of the South, and the West Coast.

The national readmission rate reductions observed in 2012 were generally consistent across all age groups. Age-controlling did not produce significantly different results. There was a widespread drop in readmission rates across the country. Most HRRs saw rates fall by more than 1% in 2012, compared to 2011. Readmission rates were lower in 2012 than they were in 2011. However, the absolute number and readmissions per beneficiary have been declining over the six year period. The readmission rates of hospitals that participated in the P4P program were on average lower than those at non-participating institutions, except for the very small and largest hospitals.


Our results for 2007-2011 match other studies that have examined hospital readmission rates in the past decade. The differences between our and their results regarding readmission rates in this time period are more likely due to methodological differences in the way rates were calculated than significant differences in trend.

There could have been many factors that contributed to the observed reduction in readmission rates. The observed drop in readmission rates in 2012 could be explained by payment reforms and other initiatives that aim to reduce avoidable readmissions. This is proving to be a significant factor in improving care and provider behavior. The current analysis of readmissions rates doesn't show whether these strategies or policies are having an impact on rates in any meaningful way. However, overall readmissions rates for 2012 were lower than in the past five years.

One reason why readmission rates are declining is because more beneficiaries are receiving care in non-inpatient settings, such as observational stays or emergency departments. Although we didn't control for changes in health status, it's possible that readmission rates have been affected by changes in mortality or severity. However, the size of the increase in 2012 suggests otherwise. It does not appear that a new influx of younger enrollees is driving rates down. Rates fell in all age groups by the same amount.