For a healthcare system where all patients receive care adjusted to their personal needs.

Pacmed helps healthcare providers by providing them with tools to make data driven decisions. It combines medical expertise with machine learning and learns from large volumes of data.

Challenge

Determining the optimal moment of discharge for a patient from an Intensive Care Unit (ICU) is a complex decision. A patient in the ICU generates about 30,000 data points per day. This concerns values of vital functions, but also lab tests and results. Together they indicate the chance that a patient would be readmitted if they were discharged - and both early and late discharge have negative consequences.

Approach

Pacmed had built an algorithm to predict how necessary it is for people to stay in the ICU. Eli5 built a dashboard that would process this analysis to provide ICU doctors with a clear overview of all required information to prevent unnecessary readmissions and long stays in the ICU. The dashboard needed to make the data visual, so doctors could use it on their screen.

Insight

Pacmed Critical is a discharge software that uses 14 years of data from 16,000 IC admissions in Amsterdam. The machine learning algorithm make predictions based on self-found patterns and connections. This supports physicians in making important decisions by sharing information about what has worked with similar patients.

Impact

Pacmed predicts the probability of a patient being re-admitted or dying within 7 days if they are discharged and predicts recordings with an accuracy in line with scientific research. Based on retrospective scenario analyses, the software can reduce the number of readmissions by 10-15%. The average length of stay may be reduced by 1-5%. In 2019 Pacmed Critical won the Computable Award for best healthcare project.