About one of my psychiatric readmissions posts, a reader wrote me,
Your comparison to a readmission for another broken bone seems pretty inappropriate. I don’t know enough about physical medicine/disease to find a “perfect” analogy, but mental health readmissions seem to be more analogous to something like diabetes or certain types of heart disease – chronic conditions that have a high risk of rehospitalization if left untreated, but should be able to be well controlled (and thus few rehospitalizations) if appropriate treatment (such as medication) is determined and adherence is high. Based on having a family member who has been hospitalized several times for anxiety (which MAY be different from diagnoses like bipolar, in this context – but may not be), it seems that a readmission within a couple of years may well mean that the mental health after the first discharge never returned to “100%”, but it was also not “bad enough” during the interim to require rehospitalization. (Medication compliance has been high – but that alone has not been a “cure.”) If you’re looking for multiple readmissions, even 10 years might be reasonable.
Like the reader, I too am not expert in this area, so I offer these points as speculative and invite subject matter experts to weigh in:
- It’s essential to be explicit on what use one has in mind for readmission rates. A typical use is as a quality indicator. That’s what I have in mind. What factors of treatment are modifiable such that readmissions (somehow measured) could be reduced? I expressed this more clearly here.
- My broken bones case did assert a chronic condition of a sort, “bones more prone to fracture than is typical.” In fact, I was getting at the very type of chronic issue that the reader raises. So, I agree that diabetes and cardiovascular conditions apply. Even with that, I’m less certain that multi-year readmission rates are reasonable quality indicators for mental or physical health.
- As suggested, it is reasonable to expect variation in potentially avoidable and unavoidable readmissions across mental health diseases and even within diseases, by severity and other related factors.
- The literature supports the sensible idea that medication adherence is a relevant factor. You’ll see this factor mentioned in my recent round-ups (here and here). Is poor adherence a patient or provider factor? At first blush it seems like a patient factor, but consider also attention to side effects and outpatient follow-up and monitoring. So, if adherence is poor — poor enough to lead to a readmission — is this an indication of poor provider quality or not? This is a tricky question. I invite comment.
- More generally, and acknowledging all the other social and prognostic causal factors, are readmissions and indication of poor quality of either the index hospitalization or subsequent outpatient care? If so, what’s the maximum interval between index and subsequent hospitalization for which it is reasonable, on average, to infer a quality problem?
About the questions in the last bullet, I refer you to the OECD report, “Selecting Indicators for the Quality of Mental Health Care at the Health Systems Level in OECD Countries” (PDF). The authors considered readmissions at 7 and 30 days but not longer. Unless I missed something, it’s not evident in the report why longer intervals were not considered, but as I wrote, the case for (poor) quality-driven rehospitalization becomes weaker as the interval lengthens. That’s not to say rehospitalizaton at one, two, or ten years later is not related to quality of care. It’s just far more likely that if rehospitalizations do indicate poor quality at all, more proximate rehospitalizations are a stronger indicator. How many other good reasons are there that a patient should return to the hospital in a week or month after discharge? All other things equal, is a hospital with a higher 7- or 30-day rehospitalization rate offering worse inpatient care or transitions to worse follow-up, outpatient care?
(There are so many issues that I’m not raising here. For a good discussion how the use of rehospitalizations as quality indicators can be problematic see Karen Joynt and Ashish Jha.)
The OECD report authors write,
Hospital readmission rates are widely used as proxies for relapse or complications following an inpatient stay for psychiatric and substance use disorders. Since they indicate premature discharge or lack of coordination with outpatient care, high readmission rates have led some inpatient facilities to examine factors associated with readmissions, including patient characteristics, length of stay, discharge planning and linkages with outpatient care. Given the high cost of institutional care, reducing readmission rates can have a substantial effect on mental health spending.
They also write,
Studies of the association between readmissions rates and other indicators of quality have been mixed. Rosenheck et al. (1999) found small but significant correlations between 180-day readmission rates and other measures of outcome in a cohort of veterans with service-related posttraumatic stress disorder (PTSD), but had non-significant findings for 14 and 30-day rates. Lyons et al. (1997) found no association between psychiatric readmission rates (30 and 180-day) and clinical measures of outcome in a more diverse cohort of inpatients with psychiatric disorders.
That’s pretty weak support for a 7- or 30-day readmission rate quality indicator. Apart from the Rosenheck et al. study, it isn’t that clear that there is a large body of work that suggests longer is necessarily better. But, like I said, this is not an area of expertise for me. Can anybody speak more expertly to these issues?
