PeraHealth Insights

Reducing Preventable Readmissions through Predictive Models



In order to discuss solutions for improving readmission rates, it is first important to distinguish between two classes of readmissions – preventable and non-preventable readmissions. Non-preventable readmissions are related to the nature of a patient’s condition or disease, particularly with progressive diseases and chronic conditions. For example, a patient with congestive heart failure or cancer may have more periodic hospital visits due to complications associated with their disease, which are not preventable.

What can be addressed and improved, however, are the preventable readmissions. And while there is a lack of industry-wide consensus around what percentage of readmissions are truly preventable, it is these that we must work to address. Ultimately, there are two things that hospitals can do to reduce preventable readmissions:

  1. Make better, more informed decisions on when to discharge patients.
  2. Implement targeted post-discharge programs to ensure the proper care is provided once a patient is home.

Searching for Signs

To effectively determine if a patient is ready for discharge and what the appropriate post-discharge care plan is, hospitals need an indication of just how likely that patient is to return to the hospital. A number of factors and models that exist today help hospitals estimate a patient’s likelihood of being readmitted by evaluating a number of factors, ranging from a patient’s total number of hospital visits within the last year and their length of stay each time to general measures of morbidity. Some models also look at a patient’s social variables to assess how stable their home situation is, as these factors can also contribute to readmissions.

Finding Answers in Predictive Models

Simple tests involving these factors can help prevent some readmissions, but the real answer will lie in the use of more predictive models. But more complex predictive models can better indicate when a patient is ready for discharge and also help hospitals provide the best (and most cost-efficient) post-discharge home support, and thus reduce their chances of having to return to the hospital once they’re home.

There is a divergence in opinion as to what fraction of readmissions are truly preventable – and moreover, on what the best strategies are for reducing these preventable readmissions. However, there is no disagreement on the need to address readmissions. The industry-wide goal is to help hospitals reduce all preventable readmissions – in fact, the Patient Safety Movement has a goal of zero preventable patient deaths by 2020. To so do, predictive models can help hospitals better determine when a patient is truly ready to go home with little to no risk of having to return to the hospital, and help them deploy efficient and effective post-discharge support. The right insights will allow hospitals to take smart action both before and after discharging patients and play a key part in reducing the number of preventable readmissions.