Automated traffic monitoring of neurosurgical operating room

Abstract

Background:

Operating room (OR) traffic disrupts airflow and increases particle count, which predisposes patients to surgical site infections, particularly in longer surgeries with hardware placement. The aim of this study is to evaluate the rate of traffic during neurosurgical procedures, as well as reasons for and perceptions of OR traffic.

Methods:

This is a single-center, multimethod study monitoring neurosurgical OR traffic through direct observation, automated monitoring, and interviews. Traffic was observed between the skin incision and closure. Personal interviews with OR teams including surgeons, anesthesia, and nurses were conducted to evaluate their perceptions of the frequency of OR traffic and reasons for OR traffic.

Results:

Direct observation reported OR door opening an average of 18 times, with 20 people entering or exiting per hour. The exact reason for traffic was not verified in all traffic cases and was able to be confirmed in only a third of the cases. Automated monitoring resulted in an average of 31 people entering or exiting the OR per hour. The procedure length was significantly associated with the number of people entering or exiting the OR per hour (P < .0001). Interviews highlighted that OR teams reported traffic to be significantly lower than observed and automated monitoring results, with approximately < 6 people entering or exiting per hour.

Conclusions:

OR traffic is higher than staff expected, and updated processes are required to reduce the number of times the OR door opens. Implementing automated observation of OR traffic could reduce the OR traffic and the risk for surgical site infection.

BACKGROUND

In an 85-year lifespan, the average person will undergo 6 procedures in the operating room (OR). According to the National Center for Health Statistics, in 2009, 51.4 million inpatient procedures were performed¹ and about 2% to 5% of these surgeries will develop a surgical site infection (SSI).² SSIs result in significantly increased morbidity and mortality as they are associated with a 2 to 11-fold increase in the risk of death.^3,4 SSIs are associated with 9.7 additional hospital days and an increased cost of $20,842 per admission.⁵ Accounting for 20% of all health care–associated infections, SSIs are the costliest, with an estimated $3.3 billion in additional costs.⁶

OR traffic has been associated with an increase in SSIs.⁷ Major guidelines for SSI prevention recommend limiting OR traffic, as it has a negative effect on OR air quality.⁸ Frequent door openings correlate with increased bacterial counts and contribute to the increasing risk of SSIs.⁹ When adjusting for the length of the procedure and the number of people in the OR, every time the OR door opened increased the likelihood of the cfu/m³ being higher than 20 (recommended maximum is 20 cfu/m³ within 30 cm of incision) by 5%.¹⁰ Air particulate counts increased by 13% when the door to the OR was opened, and large particles correlated to bacterial size were significantly elevated.¹¹ OR door opening was associated with a significant increase in particles larger than 0.5 μm.¹² Opening the door to the OR also disrupted the positive pressure environment, which is vital for preventing airborne particle transmission.¹³ When the door is opened with a frequency of once per 2.5 minutes, it results in an overall elevation of contaminant level of about 7 cfu/m³ of OR air.¹³ One study found it takes 4 minutes for the particle concentration to decline and reach a steady-state level.¹³

Limiting the number of OR personnel and their traffic is important to reduce the dispersion of skin particles in the OR.¹⁴ OR traffic occurs for many reasons, such as clinical consultation, gathering supplies, emergencies, and changes in OR staff.

Directly observing staff behaviors regarding OR traffic is time-consuming and not feasible on a large scale for most facilities. Automatic devices can be used to overcome some of these issues. Other studies have used video systems to evaluate door openings and movements inside the OR during procedures with success in maintaining long-term reduction in OR traffic.^3,15 The aim of this study was to measure the OR traffic during neurosurgical procedures and explore perceptions of OR traffic among surgical teams.

METHODS

This prospective observational study was conducted at an academic medical center with an average of 1,100 neurosurgical procedures annually, of which about 800 are followed National Health Care Safety Network procedures. Observations were conducted between August 2022 and April 2023. This study used direct observation, an automated monitoring method, as well as interviews with staff to explore current OR traffic during neurosurgical procedures.

Direct observations of procedures were selected through convenience sampling during the 12 weeks prior to the start of automated monitoring. Ten hours of observation were completed in either 1- or 2-hour increments of 7 spinal fusion, laminectomy, or craniotomy procedures. Observations were made from a discrete area in the hallway outside the OR, where the OR door and adjacent area were in clear view. Detailed observation notes were recorded each time the door opened, the reason for staff movement if observable, and the number of staff entering or exiting.

Automated traffic monitoring was performed using 3 Pearl Wireless People Counters by SMS Store traffic¹⁶ as seen in Figures 1 and 2. Data was continuously collected every 15 minutes counting the number of times the sensor was crossed in each direction. The devices were validated to ensure they were accurately counting and to determine if any external factors would influence them. Influencing factors were identified as equipment crossing the sensor, hesitation at the door, and placement of notes on the door. We aimed to monitor 100 neurosurgical procedures which included spinal fusions, laminectomies, and craniotomies. We did not record between 1 and 28 minutes of the beginning or end of each procedure, averaging 13.98 procedure minutes without data available. We noted the staff entering or exiting during lunch time and change of shift. Statistical analysis was conducted using SAS statistical software. Continuous data was analyzed using 2-way t test, sign t test, or Kruskal-Wallis test. Linear regression was performed to analyze the association between the length of the procedure and the number of people entering or exiting the OR. The threshold for statistical significance used was α = 0.05.

Fig. 1.

General layout of OR with location of the sensors.

Fig. 2.

Automated monitoring device (circled) outside doors to 2 ORs.

Staff interviews were performed to explore their perceptions and behaviors related to entering and exiting the OR during a procedure. The reason for entering or exiting an OR, and the perceived frequency of entering or exiting an OR during a procedure was asked. Interviews took place within the department and took 2 to 3 minutes to complete. Staff were sampled by convenience. Surgeons, nurses, technicians, anesthesiologists, vendors, and housekeeping were included. Interviews were audio recorded after obtaining consent. Audio recordings were then transcribed and analyzed using thematic analysis to identify similar points between staff members interviewed.

RESULTS

Direct observation results

In the 7 procedures observed, the length of the procedure varied from 3 hours to 6 and a half hours long averaging 4 hours and 44 minutes. In over 10 hours of direct observation, the operating room door was opened an average of 18 times per hour with a standard deviation (SD) of 3.59. On average, 20 people entered or exited the OR per hour (SD 5.54). 87% of the time the door was opened, only 1 person entered or exited the OR. The time between door openings ranged from 0 to 24 minutes, with an average of the door opening every 3 minutes and 18 seconds. Although the average door opening occurred every 3 minutes, 67% of the door openings occurred within 0 to 2 minutes of the previous door opening. The door was held open for longer than an average door swing 23 times, with an average of 2.2 times per hour.

Staff entered or exited the OR for many reasons during a procedure, as determined through direct observation (Fig 3). Staff opened the door to retrieve or deliver supplies 18% of the time or had another clear task 16% of the time. Examples of clear tasks were moving imaging equipment, delivering specimens, and emergencies. The staff opened the door to perform hand hygiene 5% of the time. For the remaining door openings, there was no observed task (59%), or no one entered or exited, but the door was opened for no observable reason (2%).

Fig. 3.

Observed reasons for staff opening OR door.

Automated observation results

Spinal fusions with implants were the most common procedure observed (55%) using automated observation. The neurosurgical procedures’ length varied from 31 minutes to 9 hours and 38 minutes, with an average of 2 hours and 48 minutes long. Of the 16,818 procedure minutes, 15,420 minutes (257 hours) of procedures had data on how many people entered or exited the OR. The average length of the procedure was just over 2 hours at 154.2 minutes. No data were collected for an average of 13.98 minutes per procedure due to limitations on the data output of the sensor only reporting traffic in 15-minute intervals. For the automated observations, the sensor was crossed 9,143 times in 257 hours. An average of 91 people entered or exited per procedure and an average of 31 people entered or exited per hour. There was a significant association (P-value of < .0001) between the length of the procedure and the number of people entering or exiting per hour. This relationship is shown in Figure 4. There was a statistically significant difference between the 2 doors in the OR (Fig 1). The staff favored the door exiting to the scrub sink, with more people entering or exiting per hour (5.6 people/hour) on average than the wider patient door. The linear regression of people entering or exiting per hour and the procedure length was statistically significant with a P-value < .0001 and r² of 0.2501.

Fig. 4.

People entering/exiting per hour by the length of the procedure.

Table 1 summarizes factors associated with increased OR traffic. There were significantly more people entering or exiting per hour if the surgery was longer than 120 minutes (10 people, P-value < .0001) or occurred during lunch or relief (8 people, P = .007). Differences in OR traffic were observed by procedure type, with spinal pump insertions associated having the highest number of people entering or exiting per hour (38.09) and stimulator insertions associated with the lowest number of people entering or exiting per hour (20.87). Age, sex of patient, surgeon, and which OR the procedure was performed in were not significant for differences in OR traffic. Differences in OR traffic were not observed between procedures starting before or after 12:00 PM or between surgeons.

Table 1.

Procedures relationship to the number of people entering/exiting per hour

	Sample size	Mean number of people entering/exiting per hour (Standard Deviation)	Range	P-value
Length of procedure*	46	26.01 (11.73)	4.00–60.80	< .0001†
< 120 min
> 120 min	54	35.92 (10.16)	18.77–69.17
Start time‡	60	33.00 (12.93)	4.00–69.17	.1376
Before 12:00
After 12:00	40	28.89 (9.91)	7.20–49.68
Staff change for lunch or relief during procedure‡	54	35.00 (11.44)	6.67–69.17	.0007†
Lunch/Relief
No Lunch/Relief	46	27.09 (11.16)	4.00–60.80
Difference between doors in OR9§	55	20.69 (7.34)	5.00–35.16	.0001†
OR 9 Scrub Door
OR 9 Patient Door	55	15.01 (6.73)	3.00–39.33
Surgeon**	47	33.15 (13.16)	4.00–69.17	.2463
A
B	44	29.08 (10.54)	6.00–52.25
C	9	33.15 (10.91)	18.00–51.64
Procedure type**	25	36.39 (8.44)	22.29–58.63	.0020†
Lumbar or thoracic fusion
Cervical fusion	19	32.88 (13.45)	14.86–69.17
Crani (All types)	16	29.12 (10.07)	6.67–44.00
Extreme lateral interbody fusion	11	35.24 (11.31)	16.00–52.25
Laminectomy	7	25.98 (3.95)	21.00–31.50
Stimulator Insertion	7	20.87 (8.86)	7.20–32.00
Pump Insertion	5	38.09 (16.06)	17.60–60.80
Other	10	22.96 (14.59)	4.00–51.64

^*Student t test.

^†Significant differences in the mean number of people entering/exiting per hour. P-values below the alpha level of .05 are significant.

^‡Kruskal-Wallis test.

^§Sign t test.

^**Welch’s ANOVA.

Interview results

Staff interviews reported perceived reasons for traffic as gathering and delivering supplies, medications, or equipment, providing assistance, taking lunch or breaks, teaching, conducting a patient assessment, and attending a team check-in. Each staff member reported having personally entered the OR during a procedure between 1 and 3 times. OR teams understood that OR traffic should be kept as minimal as possible, and recognized the traffic could increase in urgent cases. The majority responded that an estimated 1 to 5 people entered or exited during a surgical case. The majority of OR personnel thought that it was “shocking” that OR traffic was > 30 times per hour.

Staff members also provided suggestions to reduce OR traffic, such as using phones for communication during the procedure, having supplies prepared in the OR, increasing staff awareness, door signage, and adjusting break schedules. During the interview, 2 staff members who had worked at other hospitals said there was less traffic in and out of the OR because it was closely monitored or part of the culture. Another staff member echoed this and said that reducing OR traffic is “challenging because of the culture in place.” The most suggested intervention was increased use of the phones within the OR to communicate, reevaluating the supplies needed for these procedures so they can be in the OR before the procedure starts, and coordinating breaks.

DISCUSSION

The average OR traffic with direct observations was 20 people per hour and via automated monitoring 31 people per hour. This difference between the 2 methods may be caused by sampling method/sample size, the Hawthorne effect, or an artifact caused by the automated counter miscounting traffic. The sampling method (convenience vs all procedures during the study period) and sample size (7–100) difference between the direct observation method and automated monitoring were the most probable cause of this difference. Direct observation and staff interviews provided more information for quality improvement than automated monitoring.

Multiple studies showed that a high rate of traffic during lengthy operations represents a challenge to infection prevention and patient safety. Understanding the reason for OR traffic is essential for implementing quality improvement strategies to reduce the number of people entering or exiting the OR. Additional details regarding reasons for OR traffic were gathered by direct observation and uncovered 39% of the OR traffic occurred due to the need to gather supplies, instruments, equipment, or other tasks. Our findings align with similar findings in other studies.^12,17,18 Unfortunately, the 61% remaining reasons for OR traffic are unclear and require further investigation. In similar studies, the remaining reasons for OR traffic were due to communication/status updates (9%−15%), staff breaks (8%−10%), and the rest was unclassifiable.^12,17,18 Staff interviews identified teaching, assisting in the OR, and patient assessments as other reasons for OR traffic. Knowledge and awareness gaps were also discovered during staff interviews. Staff perceptions of OR traffic were lower than what was measured in direct observation and automated monitoring.

Staff education and increasing awareness of OR traffic can be an effective intervention in reducing OR traffic.^19,20 For example, an intercom system for staff communication can decrease traffic.²¹ Supply management, clear and advanced communication, a shift change schedule prepared in advance, a sign on the door advising caution, proper education of OR team and visitors or vendors, and a robust audit process³ are all necessary to decrease OR traffic. Notably, all suggestions shared during the staff interviews were relatively simple to implement and would likely reduce the amount of OR traffic based on previously published studies.

From our experience, automated monitoring of OR traffic is inexpensive and relatively easy to use. There are some challenges with cross-referencing procedure times to the time-stamped traffic data especially when looking at larger numbers of procedures. Providing regular feedback to staff about traffic rates would align staffs’ perception of OR traffic with what is being monitored. Staff knowing that the traffic is being monitored (Hawthorne effect) could lead to a reduction OR traffic.¹⁵ Coupling an automated monitor with regular feedback to staff and implementing staff-suggested interventions would reduce OR traffic.

CONCLUSIONS

Direct observation can gather the most information on OR traffic per procedure, although it is time-consuming to collect a representative sample. Automated monitoring through sensors allows for a larger data sample but also has limitations in the information it can provide. Staff interviews provide insight into perceived reasons behind OR traffic. Measurement of OR traffic is important to evaluate the risk of SSIs and environmental interruptions. OR traffic monitoring is necessary for the OR team to focus on decreasing OR traffic in SSI reduction efforts.