Responding to my questions about psychiatric readmissions, Benjamin F. Miller, Director of the Office of Integrated Healthcare Research and Policy in the Department of Family Medicine at the University of Colorado School of Medicine, wrote me. What follows is an excerpt, lightly edited for clarity/typos and with my emphasis (bold) added:
Psychiatric admissions are such a rarity nowadays that it seems to be a useless thing to measure regardless of where you are in the country. While these patients have severe and serious needs (suicidality, etc.) most of the time they are readmitted to the hospital is due to another complicating medical issue. This does not mean that some patients are not routinely hospitalized for their bipolar disorder, etc. however; measuring quality based on when they return and readmitted seems to be a poor proxy for actually improving their care.
There are a couple of other contextual pieces that help us make sense of the problem of measuring quality in mental health. First, when viewed as an isolated condition, mental health alone can be costly. When mental health combined with chronic disease (often a reason for regular admission to hospital) the mental health costs skyrocket. For example, look at MEPS data (http://bit.ly/2j6DJm):
Annual Medical Expenditures for Adults with a Specific Chronic Condition, with and without a Mental Health Condition
Cost without mental health condition Cost with mental health condition Adults w/o chronic condition $1,913 $3,545 Heart condition 4,697 6,919 High blood pressure 3,481 5,492 Asthma 2,908 4,028 Diabetes 4,172 5,559 The other problem with mental health is where these patients present. These folks are mostly in primary care. Consider that 84% of the time, the 14 most common physical complaints have no identifiable organic etiology ); 80% with a behavioral health disorder will visit primary care at least 1 time in a calendar year (http://1.usa.gov/SjbKwY and http://www.ncbi.nlm.nih.gov/pubmed/8381266;50% of all behavioral health disorders are treated in primary care (http://1.usa.gov/SjbWfT); and 48% of the appointments for all psychotropic agents are with a non-psychiatric primary care provider (http://bit.ly/Sjc0MA). What this means is that anything that is about “specialty” mental health (including hospital admissions) are missing the majority of the population with mental health (another take here).
Just to drive home the main points and draw them out a bit:
- “Readmissions” for patients with mental health conditions may not be for mental health treatment per se. Rather, they may be for treatment of other chronic conditions. However, those with mental health conditions may be at greater risk for failing to care for those other chronic conditions without appropriate outpatient support. To the extent this is the case, the readmission may be far less about the quality of inpatient care during the index stay, but about that of outpatient care in the community setting.
- Looking at the chart above, it’s clear that the costs associated with patients with mental health conditions are higher relative to patients without mental health conditions. This is true whether they have a chronic condition or not. Nevertheless, it is consistent with the hypothesis that mental health conditions render care management more challenging.
- No matter what the condition, one is far more likely to have greater contact with primary and outpatient care than inpatient care. Obviously it makes sense to locate the management functions where the patients are and to focus analysis on such patients and settings.
For all that, we still need to measure the quality of the care received, separate from patient and other factors not modifiable by the health system, and to attribute it to the appropriate provider. Also, we still may have reason to be concerned about rehospitalizations even if they are not good measures of quality. After all, they do represent a high level of resource use and a high burden on the patient that, to the extent possible, we’d like to avoid. If things go right (or better) somewhere, do we not expect (avoidable) hospitalizations to come down? Sure. But that doesn’t mean it’s the only worthy measure (if it is even that) or something hospitals should be at risk for.
