Length of stay declines persist following the initial COVID surge

August 30, 2021

The following analysis uses Olive’s own data to better understand how healthcare works. By analyzing hundreds of thousands of claims, eligibility, and operations, we hope to elucidate what happens in healthcare in (close to) real time. Our first discussion considers how length of stay has changed in the last 18 months. Future analyses will center on the role of BMI or diabetes in care costs and how telehealth is (or is not) impacting the IP admissions.

In the past 18 months, hospital executives focused on optimizing patient discharge to create capacity for COVID cases and avoid hospital-acquired conditions. Olive analyzed data from seven large health systems (65 sites) in our customer database to see if COVID was an accelerant in the existing efforts to lower surgical length of stay (LOS).

In the months following the COVID surge (defined by these healthcare systems as April 1, 2020 to June 1, 2020), the hospitals in our sample saw a decline in their surgical LOS of 0.1 days (2 hours) relative to the same time period pre-COVID.


How durable are recent LOS declines?
Length of stay has declined substantially over the past 50 years. Technological advances, improved discharge plans, and the introduction of the DRG payment system have all led to a consistent average length of stay (ALOS) of 4.5 days. Lower LOS allows providers to increase bed turnover, improving patient flow and margins while letting patients recover at home.

It is not surprising that providers wanted to accelerate LOS declines given the capacity constraints they were under during COVID surges and potential staffing shortages. The question is: Did they?

To faithfully compare changes in LOS, we grouped each surgical encounter into a “cohort” (clinically relevant groups of encounters) that excluded sicker, high-risk, or more complex cases. Data scientists defined these cohorts with not just procedure and diagnosis codes, but information gleaned from the operative notes that determined if cases are truly clinically similar. It is possible that lower risk individuals chose not to receive surgery given the pandemic. However, given the 12+ month timeline evaluated, it is unlikely they forewent surgery indefinitely making the groupings reasonably similar.

Of the 76 cohorts identified, 26 saw a decline of at least 0.1 days. However, some of these declines were present before the pandemic, indicating a continuation of a trend rather than a new trend. For example:

  • OB/GYN surgeries markedly declined in 2020 and 2021
  • Orthopedic procedures continued to have lower LOS (likely due to payment adjustments)
  • General surgery procedures remained at a relatively low LOS and didn’t decline further

Average length of hospital stay, by service line
7 health systems, 2017-2021


In particular, C-section (part of OB/GYN) and GI (general surgery) cohorts had lower LOS than they did in the pre-COVID period.

Home health supports a lower LOS
These LOS reductions would be inconsequential or even harmful if they resulted in a greater number of patients returning to the hospital or if it resulted in a lower level of patient care or patient outcomes. Looking at just those procedures that had a lower LOS (at least 0.3 days), there is not an increase in hospital readmission rates within 30 days.

30-day hospital readmission rates
All non-emergency department returns related to previous surgery


In fact, some of LOS declines may be attributable to greater home health use. A greater percentage of patient discharges in the “LOS declined” group had home health relative to the general population.


Even with patients’ hesitancy to have home health care teams in their homes due to the pandemic, we still see almost 20% of discharges recommended to have home health involvement as part of the discharge plan.

What does this mean?

It means hospitals can do it. Despite several years of stalled progress, the past 12 months have shown that providers can lower LOS for up to a third of surgeries. They can better use home healthcare providers, talk to service line leaders and work with nurses on discharge planning.

Hospitals should re-examine their own data to see if they’ve experienced a decline in length of stay — and if so, is it a continuation of a trend, an acceleration of a trend or a new trend? In reviewing their data, hospitals can identify the material changes that will allow them to maintain this efficiency as they work through any elective procedure backlog and see volumes rise.