Remote Patient Monitoring Program Dashboard

This dashboard was developed to support the monitoring and analysis of patient health data collected through a Remote Patient Monitoring program. The goal of the dashboard is to transform raw monitoring data into structured insights that help clinicians and program managers identify patterns in patient health conditions and assess potential risks of hospital readmission.

Remote patient monitoring programs collect continuous physiological data from patients using home-based medical devices. In this project, monitoring data from three primary devices (Blood Pressure Monitor, Blood Oxygen Monitor, and Weight Scale) were analyzed to evaluate how abnormal readings and alert events may relate to future readmission.

In addition to device monitoring metrics, the analysis incorporates patient demographic and clinical characteristics, including age, gender, and Congestive Heart Failure (CHF) group membership. These variables help provide a more comprehensive understanding of patient risk factors and allow for more informed interpretation of monitoring data.

One key metric used in this analysis is the Alert Ratio, which measures the proportion of alerts generated relative to the total number of device readings. Higher alert ratios may indicate deteriorating patient conditions or irregular health patterns that could increase the likelihood of readmission.

1. Record the number of readings and alerts for each monitoring device individually, covering at least two completed weeks.
2. Specify weeks with at least an accurate alert.
3. Calculate the Alert Ratio for each device by dividing the total number of alerts by the total number of readings during the weeks specified in step 2. If there is no week with an accurate alert for a device, the Alert Ratio is equal to 0.

1. Readings and Alerts of medical devices including Blood Oxygen Monitor, Blood Pressure Monitor and Weight Scale are recognized as the most important factors in predicting the readmission status.
2. Younger patients face higher readmission risk compared to older patients which is counterintuitive and needs more investigation.
3. Readmission risk is sensitive to the gender and males experienced a higher chance of readmission.
4. Patients who are members of the Congestive Heart Failure Group are less likely to be readmitted compared to patients who are not.

1. It would be helpful if two flag variables are formed to determine accurate readings and related alerts. So, it would be much more convenient to compute Alert Ratios correctly and predict the readmission status.
2. As the number of monitored weeks increases, models will be able to capture a more in-depth representation of patient health trends, thereby improving prediction reliability.