Analytics – process analysis of surgery schedules and the requirements for, and flow of, supporting surgical instruments

Tools – SQL for data analysis, optimization models and algorithms for scheduling surgeries and instrument usage, process mapping, Excel

Project Description

Worked with business partners and a leading orthopedic hospital to determine the impact of surgeon block schedules on instrument usage during the execution of surgeries. The number of such instruments can be substantial; the number of trays requested may be much greater than the number that are used.

Large hospitals may have tens of millions of dollars invested in instruments and trays, and storing trays requires lots of space, can be expensive. All instruments used during a surgery must be sterilized. The daily workload in a sterilization facility is substantial; a large hospital may require a hundred or more persons to sterilize instruments in a timely manner.

Benefits Delivered:

The analysis showed many tray types had significant amounts of excess inventory.  The block schedules created using the optimization methods and heuristics resulted in schedules that greatly reduced variations in workloads for the sterilization processing unit and reduced inventory levels for trays.  We showed that a very large percentage of the trays could be stocked in lower-cost locations without impacting surgery schedules.  Savings per year in reduced overtime and inventories are estimated at well over $500,000.

How it was done:

Optimization models were used to determine the impact of altering block schedules on inventory levels for each tray type and visualization tools to display the consequences of altering surgeon block schedules.

Analytics – Evidence-Based Operational Modeling (EBOM) used to evaluate strategies that can reduce Surgical Site Infections (SSI’s), improve patient care and hospital efficiency

Tools – Statistical analysis, DISCO process modeling software (© Fluxicon, Netherlands, 2016), C-Visual Explorer (CVE), Process Plant Computing Ltd., UK.

Project Description

Multiple modes of data visualization and analysis were used to gain insight into operations and patient outcomes at a major orthopedic surgical hospital. Determined whether there are measurable differences in the routing or timing of patient flow for cases that are complicated by post-operative infection in spine surgery (which may be long and complex cases) compared to hip and knee surgery (which typically are faster and more routinized).

Benefits Delivered:

Administrative data and timepoint analysis may not accurately reflect the true trajectory of patient care.  While univariate statistics using Wilcoxon-Mann-Whitney tests showed statistically significant differences in procedure length for both types of surgery, we demonstrated that process modeling and parallel coordinates analysis can be used in combination with traditional statistical analyses to gain greater understanding of patient activity in hospital settings, and trends in hospital data. 

How it was done:

Cleansed data and calculated univariate statistics, examining patients with and without infections within the spine surgery and hip/knee surgery populations.  Modeled patient flow through the operating room using DISCO process modeling software, and visualized relationships between patient- and process-level variables using C-Visual Explorer parallel coordinates software.  Apart from the first transition (InOR to StartAnesthesia), which happened very quickly in all subjects, the time differences for hip/knee patients between infected and non-infected cases were smaller than the time differences for spine cases.