High Number of Door Openings Increases the Bacterial Load of the Operating Room

Abstract

Background: Operating room (OR) traffic and door openings have emerged as potential modifiable risk factors for the development of surgical site infections.

Methods: This study compared the microbial load of a Control OR without traffic versus a Simulated OR with the traffic in a typical orthopedic surgery case. Air particle counts and colony forming units (CFUs) were measured. A novel iOS app was developed to provide real-time door counts.

Results: There were 1,862 particles >5.0 mcm in the Simulated OR compared with 56 in the Control OR. The CFUs from plates in the Simulated OR ranged from 4–22 (on brain heart infusion [BHI] agar), 2-266 (on mannitol salt agar [MSA]), and 1-19 (on Pseudomonas isolation agar [PIA]), while all plates in the Control OR grew 0–1 CFUs.

Conclusions: High number of door openings leads to more airborne bacteria in the OR and viable bacterial on OR surfaces. The increased bacterial load throughout the OR was independent of distance from the door.

Surgical site infections (SSIs) account for 14%–20% of hospital-acquired infections, and represent a major cause of peri-operative morbidity and mortality [1–3]. As many as 5% of patients undergoing surgery develop an SSI, with more than 780,000 occurring in the United States annually [4,5]. These infections increase length of stay by 5.7–13.7 days and increase costs by $6,700 to $37,500 per patient [6]. More than 90,000 re-admissions because of SSI occur in the United States every year, accounting for nearly $700 million. Overall, SSIs are believed to account for up to $10 billion in annual healthcare costs [7].

In recognition of the costly impacts of SSIs, there has been a growing effort to identify peri-operative risk factors. These risk factors fall into three main categories: patient characteristics (e.g., age and comorbidities), characteristics of the surgical procedure (e.g., operative time, antibiotic prophylaxis, contamination class), and the operating room (OR) environment (e.g., room traffic, type of air flow, prepping and draping of the patient) [8]. Recently, specific emphasis has been placed on operating room traffic as hospitals noticed an increasing number of personnel in the OR as well as an increase in movement into and out of the OR during surgical procedures [9]. Although some amount of room traffic is essential to the surgical case, high OR traffic has gained particular interest as a modifiable risk factor for SSIs.

Recent reviews have found that the rate of door openings in orthopedic operations ranges from 0.21–0.69 per minute [10–12]. Studies from the United States and Sweden have found that between 18% and 47% of OR door openings in orthopedic operations have no clear etiology [7,12]. Although efforts have been made to limit OR traffic to necessary openings, these have achieved limited long-term success. Simply telling OR personnel that door openings are being monitored (as well as having an observer in the OR as a silent reminder that they are being monitored) has not been shown to produce a long-term reduction in room traffic [13,14]. However, it is suspected that an automatic door counter—especially with real-time feedback—may discourage non-essential door openings [12,13]. Beyond serving as a physical reminder to dissuade non-essential room traffic, automatic door counters tell OR personnel exactly how many times the OR doors have been opened during a case and can change screen color to show when critical threshold of door openings is approaching.

The purpose of this study was to implement an automatic OR door counter with real-time feedback and measure the colony forming units (CFUs) and air particle counts (APCs) introduced from the number of door openings in a typical orthopedic case. A recent study found that colonies of Staphylococcus aureus on agar plates placed throughout the OR were substantially higher with increased door openings and staff members present [15]. Analysis of 79 orthopedic operations complicated by SSI found that among the most common bacterial causes were Staphylococcus species (present in 21% of SSIs studied) and Pseudomonas species (present in 19% of SSIs studied) [16]. For this reason, we will use three agars to measure CFUs: non-selective brain heart infusion (BHI) agar, mannitol salt agar (MSA), and Pseudomonas isolation agar (PIA). We hypothesize that the Simulated OR will have substantially more APCs than the Control OR, which would indicate an increased bacterial load caused by room traffic. Additionally, we expect that the agars placed in the Simulated OR will grow more CFUs than those in the Control OR, with the agar plates that are farthest from the OR doors growing the fewest CFUs.

Methods

Operating room location and description

This work is part of a larger quality improvement initiative at a high-volume academic center. The study was Institutional Review Board-exempt because no human subjects were evaluated. Two OR suites were used for all aspects of the experiment. Both ORs are approximately 600 square feet and can be accessed through three different, one-way doors. One door allows access (A) to the core hallway for supplies, another (B) to the scrub room, and a third (C) to the main external hallway used to navigate between other ORs, pre-operative holding areas, waiting rooms, and other hospital facilities (Fig. 1). The OR contains standard features that were present during this study such as a patient bed, case cart, anesthesia machine, and computer monitors. The room temperature was monitored and held constant at 69°F throughout the experiment. This experiment was conducted in Ohio during the summer months.

FIG. 1.

Floor plan of an operating room (OR) showing location of doors (A), (B), and (C) as well as patient bed (D), case cart (E), far corner (F), particle counter, microbial collection plates, and anesthesia machine. Each grouping of microbial collection plates included one containing brain heart infusion (BHI), one with mannitol salt agar (MSA), and one with Pseudomonas isolation agar (PIA).

Particle counters and microbial collection plates

To detect air particles, a Handilaz Mini Particle Counter (Aimil Ltd, Delhi, India) was placed on a small table adjacent to the patient bed. It was set to continuously measure particles at three different sizes (0.3 mcm, 0.5 mcm, and 5.0 mcm) for the duration of the 90-minute experiment. To study airborne microbial counts and type, groups of three microbial collection plates (Corning, Inc., Corning, NY) were placed throughout the OR. In each group, one plate contained 1.5% BHI (Sigma Aldrich, St. Louis, MO) agar, one contained Staphylococcus-selective 1.5% MSA (Sigma Aldrich, USA), and one contained Pseudomonas-selective 1.5% PIA (Sigma Aldrich). Brain heart infusion is a general purpose, non-selective, nutrient agar, whereas MSA and PIA are selective agars for growth of Staphylococcus species and Pseudomonas species, respectively. One group was placed within two feet of each OR door, one on the patient bed, one on the case cart, and one in the farthest corner of the OR as shown in Figure 1.

Simulation experiment and control

A simulated, mock operation was completed following the protocol and script previously established by Gormley et al. [17] with the following modifications. Medical student and undergraduate volunteers were used to simulate the various roles in the OR. All students had observed orthopedic surgery cases prior to this mock surgery and were assigned roles for this mock surgery to regulate their movement. No surgical instruments such as electrocautery, suction, or drills were used. The simulation was carried out over 90 minutes with 100 door openings. The openings were divided among the three doors according to ratios determined by preliminary observation data from OR cases: 30 were through the core door (A), 60 were through the scrub door (B), and 10 were through the hall door (C). Of the openings through the hall door, four were held for approximately 60 seconds to represent patient entering, C-arm entry, C-arm exit, and patient exiting.

The Control OR was tested on a separate day when no other operations were scheduled in the OR. All equipment was present in the room, but no research personnel or volunteers were in the Control OR for 90 minutes of data collection.

Pulse of the OR application

Door openings were monitored with the Pulse of the OR application, a custom-developed electronic monitoring system that provides door counts and color-coded feedback to OR staff. This iOS application receives real-time feedback from door sensors with embedded accelerometers to measure the number and speed of door openings. An individual accelerometer is adhered to each OR door, and a baseline position is set by the user in with the door closed. When the accelerometers detect a change from baseline, this information is sent to the iOS app via Bluetooth technology and a door opening is recorded for the door corresponding to the accelerometer. Depending on how long the door is held open (and the accelerometer is away from baseline) either a quick or a held opening is recorded. The time threshold for a held opening is entered into the app by the user. The app also includes the ability for users to log timed events of interest such as case cart opening, first incision, patient entering the room, and other events with the option to include free text.

A screenshot of the application interface is included for reference in Supplementary Figure S1. This application was developed and tested during five orthopedic cases. Accurate door counting ability of the accelerometers was previously verified against manual counts of door openings, included in Supplementary Table S1. During the simulation, a tablet running the app was displayed in the OR to track door openings. The number of desired openings per door for the Simulated OR was chosen by averaging the counts of previously measured cases and applying the timed averages to fit the simulation time frame of 90 minutes.

Data collection and analysis

To compare air particle movement and size between the experimental and control simulations, APCs of 0.3 mcm, 0.5 mcm, and 5.0 mcm were measured as described previously. To test microbial content at various locations within the OR, CFUs were counted for each microbial collection plate at 12, 24, 48, and 72 hours incubation. The microbial collection plates were incubated at 37°C 5% carbon dioxide in a humidified incubator (Fisher Scientific, Pittsburgh, PA). Data analysis for determining bacterial load in the Simulated OR versus the Control OR included a direct comparison of APCs and CFUs.

Results and Discussion

The purposes of this study were to compare the APCs and CFUs observed with the OR traffic of a typical orthopedic case to a control without door openings, as well as to monitor the number of door openings with an automatic counter that would display the progress on an iPad (Apple Inc., Cupertino, CA).

Particle counts

The particle counts were measured at sizes of 0.3 mcm, 0.5 mcm, and 5.0 mcm for both the Simulated OR and Control room. For the 0.3 mcm particles, counts of 1.1 × 10⁶ (Simulation) and 1.6 × 10⁶ (Control) were observed. For the 0.5 mcm particles, counts of 2.7 × 10⁵ (Simulation) and 7.9 × 10⁴ (Control) were observed. For the 5.0 mcm particles, counts of 1,862 (Simulation) and 56 (Control) were observed. Airborne biologic particle size depends on a number of factors, including temperature and class of organism; in general, airborne bacteria are between 1–7.5 mcm, sometimes existing in clumps [18]. With this is mind, the 5 mcm particle count being approximately 29 times higher in the Simulated OR than in the Control OR suggests that increasing the number of door openings increases the amount of airborne bacteria in the OR.

Microbial collection plates

Table 1 shows the CFU counts for all microbial plates in both the Simulated OR and Control OR. Higher counts were observed at all time points in all but three plates in the Simulated OR compared with the Control OR. The highest count in any plate in the Control OR was 1 CFU. The Control OR had no door openings to determine what portion of the airborne particles and bacteria on surfaces was caused by the OR environment, such as bacteria left in the OR from the previous patient or particles introduced to the OR from heating, ventilation, and air conditioning. Given that in the Control OR the highest CFU on any agar plate was 1, we can assume that any increase in CFUs observed in the Simulated OR were caused by personnel entering, as opposed to the innate microbes that may be present because of environmental factors.

Table 1.

Colony Formig Unit Counts for Control and Experimental Rooms at Progressive Incubation Time Points in Groups of Three Agars

		Control OR				Simulated OR
		12 h	24 h	48 h	72 h	12 h	24 h	48 h	72 h
Location A	BHI	1	1	1	1	4	3	4	4
	MSA	0	0	0	0	3	5	6	6
	PIA	0	0	0	0	0	0	1	2
Location B	BHI	0	0	0	0	15	16	16	19
	MSA	0	0	0	0	3	3	116	266
	PIA	0	0	0	0	0	1	7	7
Location C	BHI	0	0	0	0	12	15	17	22
	MSA	0	0	0	0	5	11	16	17
	PIA	0	0	0	0	10	11	17	19
Location D	BHI	0	0	1	1	11	11	12	12
	MSA	0	0	0	0	9	9	9	10
	PIA	0	0	0	0	0	1	1	1
Location E	BHI	1	1	1	1	3	3	3	4
	MSA	0	0	0	0	2	4	5	7
	PIA	0	0	0	0	0	1	1	1
Location F	BHI	1	1	1	1	6	6	6	7
	MSA	0	0	0	0	3	4	4	4
	PIA	0	0	0	0	3	4	4	5

Location A represents the plates near the core hallway door, location B the scrub room door, location C the external hallway door, location D the patient bed, location E the case cart, and location F the farthest corner of the room. There were 100 door openings over the course of the two-hour mock surgery in the Simulated OR. Of these door openings, 30 were through the core door, 60 were through the scrub door, and 10 were through the hall door. Of the opening through the hall door, 3 were held for approximately 60 seconds to represent patient entering, C-arm entry, and patient exiting.

OR = operating room; BHI = brain heart infusion agar; MSA = mannitol salt agar; PIA = Pseudomonas isolation agar.

As seen in Table 1, in the Simulated OR the highest counts were recorded outside the scrub door, with the hallway door and patient bed showing the next highest counts. The fewest CFU counts were recorded outside the core door. There were 100 door openings over the course of the 90-minute mock operation in the Simulated OR. Of these door openings, 30 were through the core door, 60 were through the scrub door, and 10 were through the hall door. Of the opening through the hall door, four were held for approximately 60 seconds to represent patient entering, C-arm entry, C-arm exit, and patient exiting.

When considering agar types, in three of the six locations BHI plates showed the highest CFU counts, followed by MSA and then PIA. We believe these CFUs represent bacteria, although because we did not use a fungal growth suppressor on the BHI plates, the CFUs could represent fungal colonies. There was one plate (MSA at location B, outside the scrub room) that showed exceptionally high counts. This measurement stands out as a possible contaminate as the CFUs measured at 266 (compared with less than 25 CFUs on all other agars in all locations). Given that this drastic increase in CFUs was observed between the 24- to 48-hour incubation period, it is likely that the contamination arose during sample handling. In all cases, CFUs after 72 hours from dishes in the Simulated OR were greater than the corresponding plates in the Control OR (Fig. 2). The Control OR that had no door openings had little to no growth after 72 hours (three locations grew 1 CFU on the nonselective BHI agar); this suggests that viable bacteria in the OR can be reduced by minimizing room traffic.

FIG. 2.

Growth after 72 hours exposure for the Simulated operating room (OR) and the Control OR on brain heart infusion (BHI) agar, mannitol salt agar (MSA), and Pseudomonas isolation agar (PIA). (A) represents the plates near the core hallway door, (B) the scrub room door, (C) the external hallway door, (D) the patient bed, (E) the case cart, and (F) the farthest corner of the room.

Microbial load of the OR

These results lend themselves to the growing body of evidence that high OR room traffic may be an independent risk factor for the development of an SSI. When included in a bundle of interventions (including door openings, peri-operative antibiotic agents, hair removal before surgery, and peri-operative normothermia), decreasing door openings was included in the factors that were able to decrease risk of SSI development [19,20]. Taaffe et al. [21] videotaped surgical procedures to determine areas of high traffic and plotted them against microbial load measured by the air samplers and by settle plates. They concluded that microbial load was correlated with the physical movement of people in the same area but not with the number of door openings [21]. Given these findings, we had hypothesized that the Location F dishes, those farthest from the majority of the room traffic, would have the least growth. Figure 3 shows the bacterial growth on BHI after 72 hours in relation to the location of the dish within the operating room. There was substantial growth in location F, even though there was close to no foot traffic in this area and it was away from OR doors and the patient bed. These results indicate that increased room traffic had an effect on bacterial load throughout the OR, regardless of distance from the door, likely because of OR airflow.

FIG. 3.

Bacterial growth in relation to the location of the plates within the operating room. Colony forming units (CFUs) after 72 hours on brain heart infusion (BHI) agar. Numeric values presented in Table 2.

Pulse of the OR app and real-time feedback

Given the growing evidence that high OR room traffic increases bacterial load, there have been efforts to determine the etiology of these door openings to direct interventions. Surprisingly, multiple studies have identified that a substantial portion of door openings during operations had no clear reason. Andersson et al. [11] determined that one-third of the door openings in the procedures observed were unnecessary. Table 2 summarizes findings from three recent articles that sought to determine the primary reason for OR door openings. In each case, a substantial portion of the room traffic could be avoided by reducing the number of door openings that occur for reasons that are not obvious and using telephones/intercoms to collect information. Furthermore, the rate of door openings ranged from 0.21 per minute to 0.69 per minute. One study found that it took the door approximately 20 seconds to fully close, resulting in as much as 15 to 20 minutes of every hour of the sterile procedure occurring with the door open [22].

Table 2.

Results from a Literature Review on Etiology of Operating Room Door Openings

Study	n	Setting	Rate	Information	Breaks	Supplies	Unknown/no reason
Lynch et al. [22] (2009)	28	United States (cardiac, orthopedic, plastic, and general surgery)	0.59 per min	27%$-$54%	20%$-$26%	11%$-$22%	6%$-$10%
Panahi et al. [12] (2012)	116	United States (orthopedic surgeries)	0.69 per min	14.5%	1.5%	23.3%	47.3%
Andersson et al. [11] (2012)	30	Sweden (orthopedic surgeries)	0.21 per min	14.2%	20.4%	25.9%	17.6%

Efforts have been made to educate OR staff about the effects of high room traffic, however, there has limited long-term success to reducing door openings by education of OR personnel and passive monitoring [13,14]. Future directions of this project aim to use the automatic monitoring system used in this study to count the number of door openings during surgical cases and provide real-time feedback to the OR staff. This technology can also be expanded to include real-time monitoring of multiple ORs simultaneously from a central control board, allowing for institution-wide installation and implementation. The Pulse of the OR app screen is able to change colors from green to orange to red as the number of door openings gets closer to the set limit. With this visual feedback, OR personnel may be more inclined to adjust their patterns of behavior to avoid unnecessary door openings. In the future, the orange or red screen of the Pulse of the OR app may prompt OR personnel to ask themselves if their next door opening is necessary. This effort must be balanced against growing concerns of alarm fatigue in hospitals and ORs, dictating the need for further discussion and observation of the technology in use [23]. The Pulse of the OR technology has potential to provide more granular data such as length of door openings and door opening acceleration (which is also likely correlated with the velocity of a person entering the OR), which would be expected to have an effect on air turbulence and thus also the spread of airborne bacteria and fomites.

Limitations

Our study is limited by the simulated design; although the setting of an OR was replicated, we lacked a patient that would introduce new bacteria especially in the case of a contaminated trauma. Additional limitations are introduced by the single experiment design and could be strengthened by validating in multiple ORs. Finally, our Control OR did not have any door openings; although the reduction of bacterial growth shows promise, it is not feasible to perform an operation without opening the door. The critical capacity of door openings that would pose unacceptable risk of SSI to the patient is not yet known. However, one study found that any door opening increased the expected number of CFUs by almost 70% [4]. Additional work is needed to determine the number of door openings per case that is both practicable and safe for patients.

Conclusions

High numbers of OR door openings increase the bacterial load within the operating room and could translate to an increased risk for development of an SSI. Targeted interventions at reducing OR room traffic should be used in an effort to decrease the number of new organisms being introduced into the OR. The Pulse of the OR iOS app presented here with real-time and color feedback is one such potential intervention to lessen OR traffic and SSI risk.

Funding Information

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Paul Stoodley's work is funded by the National Institutes of Health, NIH R01GM124436.

Author Disclosure Statement

The authors have no conflicts of interest or financial ties to disclose.

Supplementary Material

Supplementary Figure S1

Supplementary Table S1