Medical Economics

Dying To Reduce Hospital Readmissions

20% of people who get admitted to the hospital require readmission within 30 days of being discharged. Medicare has taken steps to reduce these readmissions, but the unintended consequence of these steps appears to be an increase in mortality rates.

Readmissions are expensive and cost Medicare about $17 billion per year. Conventional dogma holds that these readmissions are largely avoidable and that if we reduce them, we can save Medicare a lot of money. So, Medicare’s solution to the readmission problem… fine the hospitals for readmitting patients in the first 30 days after a discharge. The way the “Hospital Readmissions Reduction Program” works, is that Medicare tracks 6 diagnoses: heart failure, myocardial infarction, pneumonia, COPD, knee & hip replacement, and coronary artery bypass surgery. If a hospital has an excessively high 30-day readmission for any of these diagnoses compared to Medicare’s target, then the hospital’s total Medicare payments for that year are reduced. This applies to all of the hospital’s Medicare payments, not just for the 6 diagnoses that they track. The maximum penalty is 3% of Medicare payments.

Medicare recently announced the 2018 readmission penalties. 2,574 hospitals in the U.S. were penalized for a total of $564 million in penalties. 48 hospitals got the maximum 3% penalty and the average penalty was 0.73%. I was relieved to find that our hospital had only a 0.05% penalty – one of the lowest in the country.

There are a lot of positive things that hospitals can do to reduce unnecessary readmissions: better medication reconciliation, better communication to primary care physicians, improving patient health literacy through education, improve access to medications after discharge, and improve outpatient follow up. This was the intention of the Hospital Readmission Reduction Program. But as with many well-intentioned programs, there have been unintended consequences.

This month, an article in JAMA Cardiology examined the effect of the readmissions program on heart failure mortality. It looked at Medicare patients admitted with heart failure from 2006 – 2010 (before the readmissions penalties) and compared them to Medicare patients admitted with heart failure from 2010 – 2012 (during the implementation of readmissions penalties) and 2012-2014 (after implementation of readmission penalties). Studied were 115,245 patients at 416 hospitals. The penalty program was successful in reducing readmissions for heart failure: the 30-day readmission rate fell from 20.0% to 18.4% and the 1-year readmission rate fell from 57.2% to 56.3%. However, the 30-day mortality rate increased from 7.2% to 8.6% and the 1-year mortality rate increased from 31.3% to 36.3%. If you extrapolate the data and do the math, this is about 10,000 excess deaths in the United States per year.

The study was retrospective so all we can really do is correlate mortality with readmission and not imply direct causality between the reduction in readmission and the increase in mortality. So, it is possible that there were other factors involved, such as American physicians suddenly got a lot worse at treating heart failure (unlikely!). So why would reducing readmissions increase mortality? We can only speculate at this time but I have a few ideas:

  1. Hospitals put pressure on their physicians to keep heart failure patients at home when they get sick rather than send them to the hospital. The message to the physicians was to try outpatient treatment, even for those patients who were in danger of dying, rather than send the patients to the hospital. It is very easy to tell a patient over the phone to double their Lasix in hopes of keeping them out of the hospital – sometimes that works, but sometimes the patient really needs more intensive treatment. The message the physicians received was to try everything at all possible in order to keep the patient out of the hospital.
  2. Transitional care can result in wrong care. One of the tactics hospitals employ to reduce readmission is transitional care programs, such as having a nurse navigator call patients after discharge or have a home health nurse visit patients. One of the down sides to these programs is that it can be very tempting for the nurse to provide advice to patients who are deteriorating to try to make them better (reduce salt intake, reduce fluid intake, etc.) when what they really need is a change in medications (which the nurse by themselves can’t do). Also, patients being seen by a nurse may feel reassured about their symptoms and think that that they are getting all of the correct treatment that they need and delay going to the ER.
  3. Emergency departments were put under pressure to put heart failure patients into outpatient observation status rather than admit patients for care. Observation status equates to spending < 2 midnights in the hospital so there was pressure to make those return visits to the hospital short so they would not have to be changed from observation status to regular admission status and thus add to the readmission rates and increase the hospital’s penalty.
  4. The hospitals that were penalized the most were safety net hospitals and academic medical centers. These are hospitals that usually have the least resources and care for the most challenging patients. These hospitals tend to have the slimmest of financial margins. So, penalize them, and they have even fewer resources to treat the most vulnerable patients. Alternatively, the hospitals re-direct resources from actually taking care of inpatients to trying to prevent those patients from getting readmitted. The result, worse care to the sickest patients.

The Hospital Readmissions Reduction Program is one of those things that at the outset seemed like such a good idea: punish the bad guys (i.e., the hospitals with high readmission rates) and you will coerce them into behaving properly, thus saving Medicare a lot of money. The problem was, it was never tested before it was implemented. In the United States, we require new medical treatments to undergo randomized, placebo-controlled trials before those new treatments are approved by the FDA. During those clinical trials, we can find out if the treatment actually works and find out whether there are limiting side effects to the treatment. There was no randomized clinical trial of the Hospital Readmissions Reduction Program, it was just conceived and initiated without being studied. But now we have learned that death is a side effect of the Hospital Readmissions Reduction Program.

November 25, 2017

By James Allen, MD

I am a Professor Emeritus of Internal Medicine at the Ohio State University and former Medical Director of Ohio State University East Hospital