Using telephony data to facilitate discovery of clinical workflows

Area of Science:

• Healthcare Informatics
• Systems Engineering
• Telecommunications Analysis

Background:

• Identifying clinical workflows for redesign using Lean and Six Sigma is challenging.
• Voice over IP (VoIP) telephone call pattern analysis can visualize institutional telecommunications, complementing EMR and observation methods.

Purpose of the Study:

• Develop an analytic framework to map high-volume VoIP call patterns to clinical units in a large medical center.
• Utilize enterprise unified communications server logs to visualize communication patterns with graphical networks.

Main Methods:

• Parsed call detail records from a unified communications server to correlate call patterns and map numbers to cost centers.
• Constructed hashed data structures for edge and node files representing high-volume call patterns.
• Employed an open-source graph network tool for visualization.

Main Results:

• Analyzed one week of data: 912,386 calls totaling 23,186 hours.
• Half of all calls between number pairs averaged under 60 seconds, with an average duration of 27 seconds.

Conclusions:

• Cross-correlation of phone calls by clinical cost center enables graphical displays of enterprise communications.
• Short call durations present opportunities for automation or workflow redesign.
• Extensive VoIP usage indicates that call pattern analysis offers valuable insights into core clinical workflows.