"The Quick Check accelerated the planning and implementation of the project. The analyses pointed the way for further detailed work on the project."

Prof. Dr. Peter Hahn

Vulpius Clinic

Contact at the AI Innovation Center

Chandan Kumar

chandan.kumar@iao.fraunhofer.de

AI-optimized patient planning

Quick Check

Initial situation

When planning operations, there are always waiting times for patients. Waiting times cause dissatisfaction among patients and medical staff. In order to be able to implement smooth planning, it is important to estimate the duration of the operation as accurately as possible. The best results were obtained by calculating the time from incision patient A to incision patient B rather than the incision-to-suture time. The algorithms tested derive this using various parameters such as the average historical length of an operation, the surgeon's level of training, the patient's age, text in the surgical plan and other parameters.

Procedure

XYZWhen planning operations, there are always waiting times for patients. Waiting times cause dissatisfaction among patients and medical staff, and it is important to estimate the duration of the operation as accurately as possible in order to ensure smooth planning. The best results were obtained by calculating the time from incision patient A to incision patient B rather than the incision-to-suture time. The algorithms tested derive this using various parameters such as the average historical length of an operation, the surgeon's level of training, the patient's age, text in the surgical plan and other parameters.

Result

The following were carried out as part of the Quick Check:

  • Cluster analysis for possible uniform grouping of several features in the context of findings and diagnosis
  • Consideration of the statistical distribution of various features to analyze and derive possible planning heuristics
  • Training of different models, e.g. with gradient boosting method (XGboost)
  • Creation of a simulation routine for testing different strategies

Overall, the reliability of the forecasts could be improved with the aforementioned methods. The simulation routine also makes it possible to test different planning strategies. The newly developed approach is currently being evaluated for the planning of ongoing operations.