Leveraging Electronic Health Records and Machine Learning to Tailor Nursing Care for Patients at High Risk for Readmissions.
Heather Brom, J Margo Brooks Carthon, Uchechukwu Ikeaba and 1 others
PMID 31136529WHAT IT FOUND
The model found the highest 30-day readmission risk in patients with an emergency department visit, 9 or more comorbidities, Medicaid insurance, and age 65 years or older.
Key findings
01The highest-risk group identified by the model had an emergency department visit, 9 or more comorbidities, Medicaid insurance, and age 65 years or older.
02The decision tree’s c-statistic was 0.74, compared with 0.83 for logistic regression.
03Patients who were readmitted had on average 7.6 comorbidities and 2.0 emergency department visits, compared with 5.0 comorbidities and 0.4 emergency department visits among patients not readmitted.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study was retrospective and used only a single quarter of data from an academic hospital. It counted only readmissions to the study hospital, so patients readmitted elsewhere may have been missed. Important social information, such as housing status and substance use, was not available in the data store. The model was not tested for whether it changed nursing care or reduced readmissions. An earlier EHR readmission flag alone did not change readmission after implementation.
Declared interests
The authors declared no conflicts of interest. The paper does not report external funding.
The easy way to misread this
Do not conclude that this machine-learning profile will reduce readmissions. The study described a risk profile from past EHR data and did not test the planned nurse-led care pathway.