BACKGROUND/RATIONALE:
Operating rooms are expensive and account for a substantial amount of hospital costs. Thus, optimizing their efficiency is of paramount importance. To date, scheduling cases in the Operative Suite (OS) has been crude and is based on either some sort of human estimate or broad means of operation types, which are typically averaged across hospitals and multiple surgeons.

OBJECTIVE(S):
Objective 1. Compared the impact of a predictive modeling system and a traditional scheduling system on operating room utilization. Outcome measures were the scheduling imprecision and the operative throughput.

Objective 2. Assessed the impact of the PMS on Operative Suite (OS) personnel satisfaction. Outcome measure was the score in the Maslach burnout inventory that was administered to surgeons, anesthetists, and nursing personnel.

Objective 3. Assessed the impact of PMS utilization on important perioperative outcomes. Outcome measures were mortality, myocardial infarction and wound infection rates as captured in the morbidity and mortality (M&M) conference.

METHODS:
This was a randomized blinded controlled trial comparing the scheduling precision of two operative scheduling methodologies - the predictive modeling system (PMS) in which operative cases were scheduled with their length based on a regression modeling methodology, vs. the traditional scheduling system (TSS) according to which cases were scheduled based on historical means. Data was collected from 100 operative days per arm. TSS vs. PMS. The TSS represents the most commonly used method of case scheduling today, and served as the control. This methodology utilized historic means for the calculation of anesthetic, operative, and room clean up times. Historic means were calculated as the simple arithmetic average of duration for a particular case type over the previous three years. This is the typical methodology for scheduling used by most of VA and private facilities. The PMS represents the intervention, and used parameter estimates from case specific regression models to calculate the anesthetic and operative times.

The room turnover times were the same in both methodologies. The regression models used to calculate the time lengths of the PMS were updated every six months, to add information from the interventions performed during that time interval and mainly account for the effect of increasing surgeon experience on the model.

Randomization. A randomized block design was employed with the operative week as the unit of randomization. According to this design, cases on any given week were randomly scheduled using either PMS or TSS. Randomizing weekly blocks as opposed to individual days more efficiently took into consideration the natural tendency of many surgical services to schedule particular types of cases on a particular day of the week in order to meet logistical and other constraints.

FINDINGS/RESULTS:
Primary endpoint: The mean imprecision in predicting the end of operative day was higher with the Traditional Scheduling System (TSS) versus the Predictive Modeling System (PMS) approach (30.8 vs. 7.2 minutes, p=0.024).

Secondary Endpoints: The PMS was associated with higher throughput (379 vs. 356 cases scheduled over the course of the study, p=0.04) and lower overutilization rate (34 vs. 51% of days with overutilization, p = 0.004). Mean length of overutilization (58 vs. 93 minutes, p=0.002) and underutilization (42 vs. 65 minutes, p=0.013) were superior in the PMS arm.

The composite rate of adverse 30-day events was similar (2.2 vs.3.2%, p=0.44).

With respect to Operative Suite (OS) personnel job satisfaction, the mean depersonalization score was higher (3.2 vs. 2.0, p=0.044), and mean personal accomplishment score was lower during TSS weeks (37.5 vs. 40.5, p=0.028); there was no difference in the emotional exhaustion score (11.8 vs. 10, p = 0.223).

Taken together, the results of this randomized trial indicate that the proposed data-driven predictive modeling scheduling system improves multiple measures of OS efficiency and surgery staff satisfaction without adversely affecting outcomes.

IMPACT:
We anticipate that data from this project will contribute to improving operating room utilization, decrease operative\backlogs, and substantially improve veteran access to healthcare services

---

External Links for this Project

NIH Reporter

Grant Number: I01HX000998-01A1
Link: https://reporter.nih.gov/project-details/8594571

---

PUBLICATIONS:

Conference Presentations

1. Sharath S, Pathak A, Garcia A, Chen M, Barshas NR, Ramsey D, Tiwari V, Berger DH, Kougias P. A data-driven surgical case scheduling system improves multiple measures of operative suite efficiency. Results of a randomized controlled trial. Poster session presented at: VA Association of Surgeons Annual Meeting; 2016 Apr 10; Virginia Beach, VA. [view]
2. Chen M, Pathak A, Sharath S, Barshas NR, Kougias P. Analysis of long-term outcomes after carotid interventions. Poster session presented at: VA Association of Surgeons Annual Meeting; 2016 Apr 10; Virginia Beach, VA. [view]

DRA: Health Systems Science
DRE: Treatment - Comparative Effectiveness
Keywords: none
MeSH Terms: none