In theory it makes sense. A sick child seen in the emergency department isn’t ready to go home yet but isn’t quite ill enough to be admitted to the hospital. Maybe they need a few more hours to wake up after a seizure, or a night of intravenous hydration. They just need observation.
Unfortunately, the Medicare rules that established the “2-midnight rule” as the basis for defining observation (OBS) vs admit in 2013 for adults have been inconsistently applied in the pediatric setting. Some hospitals went “all in” requiring every child to be placed in observation then converted to an inpatient status if the length of stay (LOS) was going to exceed 48 hours. Other hospitals didn’t place any children in observation as the guidance was for Medicare patients. In the first of two articles in this month’s Pediatrics looking at observation status in 52 children’s hospitals that are part of the Pediatric Health Information System database, Gay et al (10.1542/peds.2020-0120) found the percent of total discharges that were OBS ranged from 0.0-60.3%. This extreme variability leads to numerous challenges for health services researchers.
The first challenge noted by Gay et al is LOS calculations. For hospitals that do not use an observation category, their inpatient LOS is significantly lower than hospitals that place fewer ill children in observation status. As Dr. Naomi Bardach from UCSF points out in an accompanying commentary (10.1542/peds.2020-028530), this negatively impacts the LOS comparison measurements for hospitals that use observation status and can affect “hospital ranking status.” In a related article, Synhorst et al (10.1542/peds.2020-003954) use the same database and highlight the effect on readmission rates—another important quality indicator. Observational readmission rates are not surprisingly lower than inpatient status, and similar to LOS measurements—i.e. hospitals that use observation status are thus again negatively impacted by an artificially higher readmission rate for inpatients than hospitals that don’t use observation status.
So, what does all this mean? Are we just worried about hospital rankings? While children’s hospital CEOs certainly may be, these studies shed light on the challenge of applying adult rules to pediatric settings. We know that we have to adjust medication doses to children sizes and kids aren’t “little adults.” It seems like we should develop consistent pediatric appropriate rules for inpatient and observation status. Dr. Bardach’s commentary includes a helpful table highlighting the impact of different drivers that impact observation status. Additionally, we need to better understand appropriate pediatric benchmarks for indicators like LOS and readmission rates. Until that can happen, the variability of observation status implementation should raise significant concerns about interpreting LOS and readmission quality indicators. Or, even better, as Dr. Bardach suggests, we should just get rid of the observation status for pediatrics and compare apples to apples.