Intervention analysis and classification: key to health outcomes optimization

Area of Science:

• Healthcare Management
• Health Services Research
• Biostatistics

Background:

• Healthcare outcomes are critical and influenced by interventions.
• Characterizing and categorizing interventions can facilitate optimal selection and deployment.
• Provider productivity, measured by patient visits per hour, is a key metric.

Purpose of the Study:

• To characterize and categorize a specific healthcare intervention using multi-axial statistical analysis.
• To evaluate the impact of patient schedule manipulation via overbooking on provider productivity.
• To propose a framework for uniquely identifying, comparing, and selecting interventions for maximum desired outcomes.

Main Methods:

• Interrupted time series (ITS) and robust-ITS models were applied to an 18-month dataset.
• Change-point analysis segmented data into pre-intervention, transitional, and post-effect periods.
• An 'intervention moments' table was constructed using parameters like time-to-effect, mean change velocity, and effect decay time.

Main Results:

• The overbooking intervention significantly altered patient visits per hour (p=0.0001).
• Post-intervention variance increased, while correlation improved.
• Key metrics included a 4-month time-to-effect, 11-month decay time, and a transition velocity of 0.35 visits per hour/month.

Conclusions:

• A multi-axial statistical analysis enabled the profiling and categorization of the overbooking intervention.
• The proposed 'intervention moments' table facilitates comparative analysis and selection of interventions.
• This approach can contribute to optimizing healthcare outcomes through informed intervention deployment.