ICU Data Quality: Why Heart Rate Records Differ Across Hospitals

Researchers looked at two big ICU databases to see how often nurses write down vital signs. They checked heart rate, oxygen levels, and breathing rate in the first day of care. The goal was to find out if recording habits change based on the hospital, the unit, or the data source. It turns out the source matters a lot.

Nurse-charted records showed big differences between hospitals and units. Some places logged heart rate over 50 times a day. Others barely hit 12. That low number became a quality floor. If a hospital falls below it, their data might not be reliable for research or safety checks.

Monitor streams told a different story. Machines that auto-record vital signs were way more consistent. They didn't care which hospital or unit they were in. The numbers stayed steady across the board. That makes them better for comparing sites fairly.

The study suggests we shouldn't treat all ICU data the same. Nurse notes reflect local workflows. Monitor feeds reflect machine logic. Mixing them up can lead to wrong conclusions. Researchers need to know the source before they trust the numbers.

This matters for anyone using ICU data to build models or set standards. A model trained on one hospital's nurse charts might fail elsewhere. But monitor data could work everywhere. Knowing the difference helps build better tools for patient care.