Reducing readmissions is a priority for healthcare services, but many effective interventions are costly and complex. Predictive models are showing promise in identifying at-risk patients.
Among 14 models tested, accuracy varied, and models correctly predicted readmission between 55% and 80% of the time.
Excerpt: “Reducing 28-day or 30-day readmissions has become an important aim for healthcare services, spurred in part by the introduction of financial incentives for hospitals with high readmission rates in the USA, England, Denmark, Germany and elsewhere.
“Unfortunately, many of the most effective interventions are costly, since they are multimodal and involve several components and multiple healthcare practitioners. Therefore, some healthcare teams are turning to predictive models in order to identify patients at high risk for readmission and focus resource intensive readmission prevention strategies on such ‘at risk’ patients.”
“While doubts remain about the practical value of predictive risk models (for example because it is not clear whether interventions are more effective when targeted at high-risk than low-risk patients5), it is undeniable that many models accurately predict readmission risk. Among the 14 published models that target all unplanned readmissions (rather than readmissions for specific patient groups), the ‘C statistic’ ranges from 0.55 to 0.80, meaning that, when presented with two patients, these models correctly identify the higher risk individual between 55% and 80% of the time. As a benchmark, consider one study6 that asked practitioners to estimate the 30-day readmission risk for patients discharged from their tertiary medical centre in 2008.
“Staff physicians, residents and interns correctly predicted patients who would return to hospital within 30 days with a C statistic of around 0.58 (considerably below the typical target for acceptable discrimination of at least 0.7). Nurses and case managers performed little better than chance (with C statistics of 0.55 and 0.50, respectively) in predicting readmissions at the time of discharge.6 It is possible that the predictions of healthcare practitioners have improved since 2008, due to the many insights since published in the literature regarding the causes of readmissions. Nonetheless, it seems likely that some predictive models outperform clinicians when it comes to discriminating between patients at high and low readmission risks.”
“Now that the technical feasibility of predictive modelling has been demonstrated, it is timely to ask where next. In this editorial, we argue that the priority lies with developing logic models that link the outputs from these models to the decisions practitioners need to make regarding the care of individual patients.”
WBB Take: 30-day Hospital readmission represents a flag for missed opportunities, quality issues, and increasingly incur payment penalties. Correctly identifying at-risk patients prior to discharge would enable hospitalists to seize opportunities to improve health outcomes, lower morbidity and mortality risks, and reduce financial penalties. However, staff tend to be no better at predicting readmission risk than chance, and the predictive models range in accuracy between awful (very slightly better than flipping a coin), to useful (correct 80% of the time or better). Few models are ready for the role of real-time decision support.
The “Holy Grail” is to predict with higher than 85% accuracy that a patient is at high risk of post-discharge issues prior to making the decision to hold or discharge them. This information must be presented to clinicians through a usable interface, and in a way that it supports discharge planning and improves the clinical plan in real time.
While models are in general still a distance from being at that point, machine learning, and process mining may eventually yield models that can provide real-time operational support prior to discharge.
Cited by Matthew Loxton