A group of neurosurgery researchers at Columbia University has found that when they gave patients a Fitbit after hospital discharge, patients that either didn’t track or intermittently tracked were more likely to be readmitted.
Medicare penalties have made readmission rates a big focus for many hospitals. In this study, presented as an abstract at the Congress of Neurologic Surgeons and published this month in the journal Neurosurgery, investigators gave 23 patients on the neurosurgery service a Fitbit ZipTM prior to discharge. Patients were followed for thirty days after discharge. Four patients were readmitted.
While limited by relatively small numbers, they made an interesting observation. In particular, they saw that the patients who were compliant with tracking weren’t readmitted. Patients that were intermittently compliant or weren’t tracking at all were at significantly higher risk of readmission. And according to these investigators, data integrity was better than the actual step count.
In some ways, it’s pretty intuitive as noncompliance in other areas – medications, coming to clinic – is usually not a good sign either. The notion that noncompliance with health trackers may be an important signal to consider in predicting readmission where further resources, not less, should be directed is interesting. Unfortunately, this data hasn’t been published formally yet and the study itself is quite small so further evaluation is needed here before drawing firm conclusions.
This study does offer a useful reminder that, as we are capturing bigger and bigger piles of data, the signals that matter may not be the ones we expect. And when talking about complex, multifactorial events like hospital readmission, there is no single 100% predictive signal either.
