• More bad news for Medicare’s hospital readmissions measure

    Jordan Rao reports:

    A study published Tuesday says Medicare may be missing factors that lead to post-hospital health problems because it isn’t counting many discharged patients who come back to the emergency room but aren’t admitted.

    The study in Annals of Emergency Medicine looked at 11,976 patients discharged from Boston Medical Center, the largest safety net hospital in New England, in the first half of 2010. The total number of impatient discharges during the period studied was 15,519, including patients who were readmitted more than once.

    The researchers found that a quarter of those discharges resulted in at least one emergency department visit within a month after the patients left the hospital. But 54 percent of those visits to the emergency department (ED) did not lead to a readmission and thus would not have shown up in the statistics when Medicare calculated the hospital’s readmission rate. […] 

    Similar results were found in a study published in January in the Journal of the American Medical Association. That study calculated that for every 1,000 discharges, there were 98 treat-and-release visits to the emergency room.

    The Annals of Emergency Medicine paper by Rising et al. is here. The JAMA one here. I have not read either, yet.

    Is there a good reason not to count ED visits along with readmissions? Recall that readmissions are down. To what extent can that be explained by an increase in ED visits that don’t result in an admission? I’m not saying that’s what’s happening, but do we know?


    • Unscheduled ED returns after admission – regardless of decision to admit or discharge – represent a (potential) failure in the transition home. But don’t worry, hospital administrators are doing their darnedest to find ways to keep the re admission from happening. Not sure if “observation” status counts towards readmissions, but that would be one trick. Alternatively, resources are finally put in the ED – case management, social work – to find alternatives to admission.

    • I’ve heard it said that health care costs are like balloon – you squeeze one part, another part expands.

      New York University has an algorithm that sorts ED claims into avoidable and non-avoidable visits. Avoidable visit subcategories include those that weren’t actually urgent, visits that were urgent but could have been treated in primary care, and urgent visits that were preventable. Non-avoidable visits could include injuries and sudden infections. (Their classification methods are based on expert reviews of hospital charts. It’s a statistical tool to evaluate access to care at a high level, not an actual triage tool.) I’d like to see that algorithm run on the study population, or to have some clinicians do a chart review. That could shed some light.

      One has to note that the study above was only based on hospital administrative data, not Medicare/Medicaid administrative data. There are a lot of things about these patients that the hospital is not observing. The study didn’t consider Medicaid or dual eligible status and it didn’t consider existing comorbidities. And the hospital is a safety net hospital, so its case mix is probably more acute than your average hospital.

      And yes, observation stay utilization is also something one would have to consider.

    • Is this necessarily a bad thing? An ER visit gets paid only for services performed without facility charges.