Ask the Expert: No Simple Solution for Readmission Risk

Hospitalists are under increasing pressure from hospital administrators to do their part in reducing preventable readmissions. But while there are some proven programs aimed at improving the discharge planning process, there are still plenty of unanswered research questions.

Dr. Devan Kansagara of the Oregon Health and Science University, Portland, and of the Portland VA Medical Center has been studying hospital readmissions and transitions of care for the past several years. As the director of the evidence synthesis program at the medical center, he led a review of readmission risk prediction models and found that most of them perform poorly (JAMA 2011;306:1688-98). He published a study in the Journal of Hospital Medicine, looking at the use of a brief, scripted phone-based needs assessment in chronically ill Medicaid managed care patients (2012;7:124-30).

In an interview with Hospitalist News, Dr. Kansagara detailed some of the gaps in knowledge and what may be ripe for future evaluation.

Hospitalist News: Do we know what interventions work best to reduce readmission rates?

Dr. Kansagara: There are some interventions that have worked. Eric Coleman’s care transitions program, Mary Naylor’s discharge work, and Project RED (Re-Engineered Discharge) have all been proved to work in randomized controlled trials. But in terms of knowing which of these interventions worked best and in what populations they should be used, I don’t think we know that yet. All of these interventions have some similar elements. For instance, there is a bridging component that involves educating patients both in and out of the hospital. That may be a component that’s helpful. But there was a recent review in Annals of Internal Medicine looking at several different types of transitional care interventions and many of them didn’t work (2011 Oct. 18;155:520-8). I don’t know that it’s entirely clear that we’ve pinpointed the elements that will bring down readmission rates.

HN: What research questions are still unanswered when it comes to preventing readmissions?

Dr. Kansagara: We really don’t know which components of the interventions are most important. We also don’t know how many readmissions are actually preventable and what makes them preventable. Another really interesting and completely understudied area is at the other end of the care continuum. What happens as people are getting sick and on their way into the emergency department or the hospital? People haven’t looked broadly at what is happening in the community and the outpatient practice setting.

HN: How important is it to assess a patient’s risk for readmission?

Dr. Kansagara: A lot of people have tried to develop models to predict the risk of readmission, and most of these models don’t work very well. We don’t have a gold standard way of identifying patients at high risk for readmission. But rather than focusing too much energy and attention on quantifying a patient’s risk for readmission, it’s probably more important to develop a system for identifying patient factors that can contribute to readmission risk that might not be readily collected in the medical record. These would be things like housing status, health literacy, substance abuse issues, distance to follow-up care, and the feasibility of them following up.

Hospitalists are under increasing pressure from hospital administrators to do their part in reducing preventable readmissions. But while there are some proven programs aimed at improving the discharge planning process, there are still plenty of unanswered research questions.

Dr. Devan Kansagara of the Oregon Health and Science University, Portland, and of the Portland VA Medical Center has been studying hospital readmissions and transitions of care for the past several years. As the director of the evidence synthesis program at the medical center, he led a review of readmission risk prediction models and found that most of them perform poorly (JAMA 2011;306:1688-98). He published a study in the Journal of Hospital Medicine, looking at the use of a brief, scripted phone-based needs assessment in chronically ill Medicaid managed care patients (2012;7:124-30).

In an interview with Hospitalist News, Dr. Kansagara detailed some of the gaps in knowledge and what may be ripe for future evaluation.

Hospitalist News: Do we know what interventions work best to reduce readmission rates?

Dr. Kansagara: There are some interventions that have worked. Eric Coleman’s care transitions program, Mary Naylor’s discharge work, and Project RED (Re-Engineered Discharge) have all been proved to work in randomized controlled trials. But in terms of knowing which of these interventions worked best and in what populations they should be used, I don’t think we know that yet. All of these interventions have some similar elements. For instance, there is a bridging component that involves educating patients both in and out of the hospital. That may be a component that’s helpful. But there was a recent review in Annals of Internal Medicine looking at several different types of transitional care interventions and many of them didn’t work (2011 Oct. 18;155:520-8). I don’t know that it’s entirely clear that we’ve pinpointed the elements that will bring down readmission rates.

HN: What research questions are still unanswered when it comes to preventing readmissions?

Dr. Kansagara: We really don’t know which components of the interventions are most important. We also don’t know how many readmissions are actually preventable and what makes them preventable. Another really interesting and completely understudied area is at the other end of the care continuum. What happens as people are getting sick and on their way into the emergency department or the hospital? People haven’t looked broadly at what is happening in the community and the outpatient practice setting.

HN: How important is it to assess a patient’s risk for readmission?

Dr. Kansagara: A lot of people have tried to develop models to predict the risk of readmission, and most of these models don’t work very well. We don’t have a gold standard way of identifying patients at high risk for readmission. But rather than focusing too much energy and attention on quantifying a patient’s risk for readmission, it’s probably more important to develop a system for identifying patient factors that can contribute to readmission risk that might not be readily collected in the medical record. These would be things like housing status, health literacy, substance abuse issues, distance to follow-up care, and the feasibility of them following up.

Hospitalists are under increasing pressure from hospital administrators to do their part in reducing preventable readmissions. But while there are some proven programs aimed at improving the discharge planning process, there are still plenty of unanswered research questions.

Dr. Devan Kansagara of the Oregon Health and Science University, Portland, and of the Portland VA Medical Center has been studying hospital readmissions and transitions of care for the past several years. As the director of the evidence synthesis program at the medical center, he led a review of readmission risk prediction models and found that most of them perform poorly (JAMA 2011;306:1688-98). He published a study in the Journal of Hospital Medicine, looking at the use of a brief, scripted phone-based needs assessment in chronically ill Medicaid managed care patients (2012;7:124-30).

In an interview with Hospitalist News, Dr. Kansagara detailed some of the gaps in knowledge and what may be ripe for future evaluation.

Hospitalist News: Do we know what interventions work best to reduce readmission rates?

Dr. Kansagara: There are some interventions that have worked. Eric Coleman’s care transitions program, Mary Naylor’s discharge work, and Project RED (Re-Engineered Discharge) have all been proved to work in randomized controlled trials. But in terms of knowing which of these interventions worked best and in what populations they should be used, I don’t think we know that yet. All of these interventions have some similar elements. For instance, there is a bridging component that involves educating patients both in and out of the hospital. That may be a component that’s helpful. But there was a recent review in Annals of Internal Medicine looking at several different types of transitional care interventions and many of them didn’t work (2011 Oct. 18;155:520-8). I don’t know that it’s entirely clear that we’ve pinpointed the elements that will bring down readmission rates.

HN: What research questions are still unanswered when it comes to preventing readmissions?

Dr. Kansagara: We really don’t know which components of the interventions are most important. We also don’t know how many readmissions are actually preventable and what makes them preventable. Another really interesting and completely understudied area is at the other end of the care continuum. What happens as people are getting sick and on their way into the emergency department or the hospital? People haven’t looked broadly at what is happening in the community and the outpatient practice setting.

HN: How important is it to assess a patient’s risk for readmission?

Dr. Kansagara: A lot of people have tried to develop models to predict the risk of readmission, and most of these models don’t work very well. We don’t have a gold standard way of identifying patients at high risk for readmission. But rather than focusing too much energy and attention on quantifying a patient’s risk for readmission, it’s probably more important to develop a system for identifying patient factors that can contribute to readmission risk that might not be readily collected in the medical record. These would be things like housing status, health literacy, substance abuse issues, distance to follow-up care, and the feasibility of them following up.