Decreasing length of stay leads to fewer re-admissions

A recent study of all Veteran’s Affairs hospitals in the United States over a 14-year period showed that decreasing length of stay (LOS) has not resulted in a commensurate increase in 30-day re-admissions – in fact the opposite appears true (Kaboli PJ, 2012).

This, and previous studies (Baker DW, 2004), ( Rosenthal GE, 2003), should give hospital administrators and practitioners confidence that efforts to improve patient flow and decrease length of stay are also likely to have a positive impact on quality of care.

In the study, which analyzed data over 14 years, mean risk-adjusted LOS at 129 hospitals (representing 4,124,907 patients) decreased by 1.46 days (from 5.44 to 3.98 – a 26.8% relative decrease). Patients with acute myocardial infarct experienced the greatest decrease of 2.85 days.

Meanwhile, 30-day re-admission rates declined relatively by 16.4% over the same period (from 16.5% to 13.8%). Furthermore, all-cause mortality, both in and out of hospital, also decreased by 3.4% and 3.0% per annum respectively.

These results would seem to indicate that shortening LOS can increase safety, quality and reduce re-admissions. There is, however, an important caveat. In the study they found increased readmission rates at hospitals that discharged patients sooner than expected (based on illness severity and general trends in LOS). This would suggest, consistent with common sense, that it is possible to discharge patients too early with the result that they must return for medical care later.

The authors point to several factors that may have contributed to these findings at Veteran’s Affairs (VA) hospitals:

  • Several reports had indicated significant inefficiencies at VA hospitals at the beginning of the study period that likely contributed to excessively long LOS, particularly compared with similar private-sector hospitals.
  • The VA system initiated a Flow Improvement Inpatient Initiative in 2006 that may have improved quality of care within the hospitals and at discharge (Davies M, 2007).
  • VA hospitals undertook efforts to improve medication reconciliation, which has been shown to reduce re-admissions ( Schnipper JL, 2006).
  • Hospitalist programs were implemented, which have been shown to reduce inpatient LOS by up to 15% ( Wachter RM and Goldman L, 2002).

In their discussion, the author’s note that there may be unintended consequences associated with a focus on reducing 30-day readmission rates. They point to chronic diseases such as COPD and cancer where repeated admissions may represent appropriate care and improve quality of life for patients. Despite concern about the negative impact of re-admissions, recent studies suggest that rates of inappropriate re-admissions may in fact be as low as 8.8% ( Frankl SE, 1991).

In a previous article I suggest that cycle time (total time of all admissions within 30-days) provides a powerful and simple way to measure overall performance, and avoids an inappropriate focus on either LOS or 30-day readmission rates in isolation ( Woodhouse D, 2013). Furthermore, cycle time reduces the need to try and categorize re-admissions as avoidable or unavoidable (which the author’s suggest) by focusing instead on the total time in acute medical care which in most cases is a more relevant measure.