The operating room (OR) can be an exceedingly loud environment due to three principal sources of noise: equipment (surgical and anesthetic), staff conversations, and music.1,2,3,4 Noise is the result of air pressure fluctuation a source creates and is expressed on a logarithmic scale in decibels (dB).5 To best capture human ear frequency, the dB(A) scale was created.4,5,6,7,8 Ambient noise in the OR may or may not be modifiable. For example, cardiac monitors produce noise at 50–55 dB(A), vacuum aspiration systems at 50–60 dB(A), ventilation fans (baseline speed) at 60–65 dB(A), and surgical drills upwards of 75 dB(A).7 Surgical specialties also vary in terms of noise levels; otolaryngologic, orthopedic, and neurosurgical procedures can be associated with noise levels exceeding 100 dB(A) for 40% of the procedural time.7, 8

Induction of anesthesia is one of the loudest perioperative periods and reflects a critical period where the concept of sterile cockpit is paramount.3, 8, 9 An inability of anesthesiologists to hear subtle changes in physiologic monitors may have a wide range of significant negative consequences, including adverse effects on communication, cognition, reaction time, stress, and, most importantly, patient safety.3, 8,9,10,11 Similarly, the surgical team is exposed to untoward effects of noise primarily during the maintenance phase of anesthesia, and reducing noise has been shown to lower cortisol levels.12, 13 Lastly, there are associations between excessive noise and adverse patient outcomes, such as surgical site infection and postoperative complications in pediatric patients.13,14,15,16

Although induction of general anesthesia is a potentially distressing time for patients, a paucity of data exists regarding the impact of extraneous noise during this phase on patient satisfaction, the latter of which has been advocated as a strong indicator of healthcare quality.17,18,19,20,21

As such, we conducted a prospective non-randomized interventional quality improvement (QI) project with the goals to (1) objectively measure and document noise levels at our centre during induction of general anesthesia, (2) collect data on associated anesthesiologist and patient satisfaction, and (3) determine the impact of a program to educate perioperative healthcare professionals in noise reduction strategies on noise levels and anesthesiologist and patient satisfaction.9

Methods

We conducted a prospective non-randomized interventional QI project at St. Paul’s Hospital, a tertiary/quaternary academic health sciences centre in Vancouver, Canada, between October 2019 and February 2020. Institutional ethics approval was received on 6 August 2019 from the University of British Columbia Research Ethics Board, and the project’s protocol was registered with ClinicalTrials.gov (NCT04204785). The project took place in three phases (preintervention, intervention, and postintervention) with two parts. Part A was dedicated to measuring noise levels and assessing associated anesthesiologist satisfaction. Part B focused on gathering data on patient satisfaction. Staff engagement was achieved by multiple feature presentations of the QI project prior to commencement at anesthesia, surgery, and nursing rounds, which were followed up by email correspondence. To report the results, we used the Revised Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0)Footnote 1 for guidance.

Part A: Noise levels during induction and anesthesiologist satisfaction

Pre-intervention phase

An initial “pre-intervention” phase involved collection of data during care of patients who were aged 18 yr or older, underwent elective noncardiac surgery, and received a general anesthetic as the primary mode of anesthesia. Patients who were younger than 18 yr, were scheduled to receive regional/neuraxial anesthesia without a general anesthetic, underwent emergent surgery, were medically unstable postoperatively, or had cognitive dysfunction or hearing impairment as determined by their anesthesiologist were excluded. Surgical bookings and expected anesthetic modality were screened to identify potential cases. All anesthesiologists working at St. Paul’s Hospital were educated about the project, and 29 of 46 provided consent and participated in all three phases (see below). On the day of surgery, the attending anesthesiologist was informed of their patient’s eligibility.

Anesthesiologists administered premedication and general anesthesia according to usual practice. Once a patient was in the OR, participating anesthesiologists discretely placed a decibel meter (Decibel X software [SkyPaw Co. Ltd, Hanoi, Vietnam] running on a dedicated iPhone 7 [Apple Inc., Cupertino, CA, USA] and calibrated with a HT-80A Sound Meter [Hti, Dongguan City, China] to +/−1 dB[A]) by the patient’s head, unbeknownst to both the OR staff and the patient. Sound levels in dB(A) were measured during the phase of induction of general anesthesia, defined as the period between patient entry into the OR and establishment of a secure airway with either laryngeal mask or endotracheal tube. Anesthesiologists noted the average and maximum noise readings and subsequently completed a questionnaire (Electronic Supplementary Materials [ESM], eAppendix A). No conversations were recorded. Questions targeted anesthesiologists’ perceptions of patient’s anxiety and whether noise impacted their ability to perform; responses were recorded on a five-point Likert scale (ESM, eAppendix A).

Intervention phase

The second “intervention” phase occurred over a two-week period following completion of data collection from the pre-intervention cohort. To reduce noise during induction, we conceived and delivered a multidisciplinary educational program to the perioperative staff. Formal didactic teaching sessions were delivered to surgeons, anesthesiologists, anesthesia assistants, perioperative nurses, and room aides. The training team consisted of authors S.Y.M. (project leader), C.V.Y., and P.Y. Noise reduction techniques recommended for surgeons included answering calls/pages outside the OR and reducing traffic through the OR doors during induction.9 For anesthesiologists, this included checking the anesthetic machine and equipment prior to the patient entering the OR, turning off suction when not in use, reducing alarm volume settings, and maintaining and enforcing a sterile cockpit.9 For nurses, this included preparing equipment prior to patient arrival and distancing noisy equipment away from the head of the bed.9 Every staff member was encouraged to minimize non-patient-related conversations and asked to limit personnel in the OR during induction to only those directly involved.9 The importance of reducing noise was emphasized, specifically regarding potential effects on patient outcomes and occupational health. A detailed description of the multidisciplinary educational program is provided in the ESM, eAppendix B.

Post-intervention phase

In the third and final “post-intervention” phase, data were collected and patients were recruited in the same way as in the “pre-intervention” phase. Nevertheless, this phase included daily OR personnel education and reminders about noise reduction. Examples included signs on OR doors, discussions with the surgical team requesting quiet during induction, and reinforcing non-essential personnel not be present. Patients were unaware of these active noise reduction efforts.

Part B: Patient satisfaction pre- vs postintervention

Inclusion/exclusion criteria were as described for part A. To minimize bias and in accordance with the Research Ethics Board approval, eligible patients were considered for enrolment by their anesthesiologist if they met inclusion criteria, but consent was not given until after the procedure when the patient was awake. Following completion of the operative procedures, if deemed medically stable and of appropriate mental capacity by a physician, patients were approached by a member of the study team, either in the post-anesthetic recovery unit or on the surgical ward (postoperative day 0/1).

Patients who consented to participate were asked to complete an anonymous questionnaire that explored areas of anxiety and discomfort on a five-point rating scale and assessed satisfaction with their care using a five-point Likert-type scale (see ESM, eAppendix C; and Fig. 3). In the absence of a validated instrument to measure patients’ experience of noise, we designed the questionnaire on the basis of previously published interview questions on patient perception of sound levels in the surgical suite.4 A secondary objective of our QI initiative was to obtain information on patients’ OR experience during the time of induction to avoid clear delineation of the project’s focus on exposure to noise during the induction period thereby minimizing bias.

Data analysis

For this QI project, we studied a sample size of convenience based on a prespecified plan to allocate six weeks for each of the pre and post-intervention phases, with two weeks for the educational intervention in between.

We analyzed categorical variables using Chi squared test and, in instances where sample sizes were < 5 in at least one of the cells, Fisher’s exact test. We compared the noise level data with the Mann–Whitney U test and calculated the Hodges–Lehmann estimator for the difference in medians with their 95% confidence intervals (CIs) to report effect sizes. We used the ROUT method to detect outliers (false discovery rate, Q = 1%). Patient and anesthesiologist satisfaction questionnaire responses were transferred to a Microsoft Excel (Redmond, WA, USA) spreadsheet and coded as discrete numeric rating scale (NRS) scores from 1 to 5. Questionnaires were stored in a locked cabinet in our research office. The NRS data were analyzed with the Mann–Whitney U test. We also compared the percentages of respondents who answered “quite a bit” or “extremely” vs “not at all”, “a little bit”, or “moderately” and calculated the differences with their 95% CIs. We defined statistical significance as P < 0.05.

The statistical analyses were performed using IBM® SPSS® Statistics 26 (IBM Corp., Armonk, NY, USA), Prism version 8 (GraphPad, San Diego, CA, USA) software, and, for computation of standardized mean differences (expressed as Cohen’s d), the online calculator by Davis B. Wilson, PhD (George Mason University, Fairfax, VA, USA).Footnote 2

Results

Part A: Noise levels during induction and anesthesiologist satisfaction

We collected data during the care of 209 patients (pre-intervention cohort, N = 100; post-intervention cohort, N = 109) with baseline characteristics shown in Table 1.

Table 1 Project part A (noise levels during induction and anesthesiologist satisfaction) patient characteristics at baseline
Full size table

Median [interquartile range (IQR)] noise levels throughout induction were 66.0 [62.5–68.6] dB(A) preintervention vs 63.5 [60.1–65.4] dB(A) post-intervention (Hodges–Lehmann estimator of the difference, − 2.7 dB[A]; 95% CI, − 4.0 to − 1.5; P < 0.001). No outliers were detected. Given the logarithmic dB(A) scale, the magnitude of noise level reduction implied that the associated sound intensity was nearly halved.5, 6, 22 These results with individual data points are graphically represented in Fig. 1.

Fig. 1
figure 1

Average noise levels during induction of anesthesia preintervention (N = 100) vs postintervention (N = 109). Horizontal bars denote medians and interquartile ranges. Mann–Whitney U test, P < 0.001 (***). Application of the ROUT test yielded no outliers (false discovery rate, Q = 1%).

Full size image

Regarding the median [IQR] peak noise levels during induction, analysis revealed 87.3 [83.9–90.4] dB(A) preintervention and 86.2 [81.8–89.3] dB(A) postintervention (P = 0.10) (Fig. 2A). The ROUT test detected four outliers in the pre-intervention cohort and one in the post-intervention cohort; after removal, median [IQR] peak noise levels during induction were 87.3 [84.0–90.5] dB(A) preintervention and 86.2 [81.8–89.3] dB(A) post-intervention (Hodges–Lehmann estimator of the difference, − 1.8 dB(A); 95% CI, − 3.3 to − 0.3; P = 0.02) (Fig. 2B). The single highest peak noise level recorded was 101.7 dB(A) preintervention and 107.9 dB(A) postintervention.

Fig. 2
figure 2

Peak noise levels during induction of anesthesia. A) Shown are all data points preintervention (N = 100) vs postintervention (N = 109). Horizontal bars denote medians and interquartile ranges. Mann–Whitney U test, P = 0.10. B). Shown are the data following removal of four outliers the in the pre-intervention cohort and one in the post-intervention cohort, detected by ROUT test (false discovery rate, Q = 1%). Mann–Whitney U test, P = 0.02 (*).

Full size image

The highest average and peak noise levels were recorded during otorhinolaryngology procedures (data not shown). Similar to the complete cohort, an additional exploratory analysis of patients within the domain of otorhinolaryngology only showed a reduction in average noise from a median [IQR] of 66.3 [62.3–69.6] dB(A) preintervention (41/100; 41%) to 63.8 [60.4–65.4] dB(A) postintervention (27/109; 25%; Hodges–Lehmann estimator of the difference, − 2.3 dB[A]; 95% CI, − 4.7 to 0.0; P = 0.05).

In thirty-four out of one hundred cases (34%) in the pre-intervention cohort vs 18/109 (17%) cases in the post-intervention cohort, anesthesiologists responded that there were more than ten minutes from the patient arriving in the room to time of induction (difference, − 17%; 95% CI, − 29 to − 5; Cohen’s d = 0.53; P = 0.004). There were no significant differences between the two cohorts in percentage of patients noted by the anesthesiologist as appearing anxious prior to induction, having difficulty hearing, or having significant cognitive impairment (data not shown). The overall anesthesiologist satisfaction questionnaire results are shown in Fig. 3. Of the seven domains that focused on anesthesiologist satisfaction, all except “care-related noise” saw statistically significant improvements post education. Table 2 shows the number and percentages of anesthesiologists who responded to the seven questions with “quite a bit” or “extremely” (vs “not at all”, “a little bit”, or “moderately”) pre- vs postintervention. There were significant improvements postintervention in the domains, “staff were noisy”, “noise could be reduced”, “noise distracted me”, and “noise bothered me”; there were no changes in care-related noise.

Fig. 3
figure 3

Anesthesiologist satisfaction questionnaire responses by category. Shown are all data (represented as % of total respondents per given response) comparing pre-intervention (N = 100) vs post-intervention (N =109) cohorts. P values are from the Mann–Whitney U test for comparison of numeric rating scale scores (cf. body text, Data analysis). Categories A to G (abbreviated) correspond to eAppendix A in the Electronic Supplementary Materials (questionnaire items 8 to 14).

Full size image
Table 2 Percentage of anesthesiologists responding with “quite a bit” or “extremely”
Full size table

Part B: Patient satisfaction pre vs post intervention

Two hundred eligible patients (N = 100 in the pre-intervention phase and N = 100 in the post-intervention phase) consented to complete an anonymous satisfaction questionnaire postoperatively; eTable 3 (ESM) shows their characteristics at baseline and eTable 4 (ESM) shows the questionnaire results. Overall, patients rated their levels of anxiety and discomfort low and their satisfaction high. Among the sixteen questions, we observed no differences between the cohorts or further improvements postintervention, except for four domains: fear of pain due to the anesthetic, confidence in the OR staff, assessment of OR staff as having been professional, and assessment of OR staff as having paid attention to complaints. These were statistically significant in favour of the pre-intervention phase but these magnitudes were unlikely to be clinically important. Notably, patients’ perceived levels of surgical noise, general noise, and noise from staff conversations were not different pre- vs postintervention.

Discussion

In this interventional QI project, we observed a reduction of average noise throughout induction of anesthesia by 2.7 dB(A), i.e., nearly halving sound intensity. Comparatively, this approximates the difference between the perceived noise standing 20 metres from a passenger car travelling at 60 km/hr versus a normal conversation between two people 1 to 2 metres away.5, 6, 22, 23

While average noise levels were notably reduced, peak noise levels saw a smaller reduction of 1.8 dB(A) after removal of five outliers; this is not surprising as any single loud event, such as a surgical tool dropping on the ground, would have resulted in a large peak value. What is concerning is that the single highest peak noise levels recorded during our project were 101.7 dB(A) preintervention and 107.9 dB(A) postintervention, which is equivalent to the noise levels produced by a jackhammer and a rock band concert, respectively.22, 23 Procedure types also played a role in determining noise, with otolaryngology procedures producing the highest average and peak levels, which is in keeping with the literature.8, 10

The National Institute of Occupational Safety and Health and Occupational Safety and Health Administration (OSHA) recommends a maximum exposure of 85–90 dB(A) during an eight-hour workday.1 When accounting for interrupted periods of less noise, a total average of < 45 dB(A) for continuous noise levels while in hospital is recommended.1 This chronic exposure to noise has deleterious effects on healthcare providers, with occupational noise-induced hearing loss (NIHL) being a particular concern when levels exceed 80 dB(A).1, 4,5,6,7, 22,23,24,25 NIHL is a preventable occupational health risk for anesthesiologists; however, it is clear that despite an overall reduction in noise levels brought about by educational intervention, the safe levels recommended by established guidelines were not achieved.25 In addition, other common and often underappreciated harmful effects of extraneous noise on anesthesiologists warrant consideration—especially in the context of chronic exposure—such as those on cognition, communication, and performance.1 Our centre does not have a specific protocol with respect to noise in the OR. We observed that “safe” levels would be very difficult to achieve given the baseline noise produced by air circulation systems alone, surgical equipment, and some outdated anesthetic monitors.1, 5,6,7, 23, 25

With respect to the anesthesiologist questionnaire, the pre-education cohort results indicated pre-existing concerns regarding OR staff-related noise. While some of it was perceived as care-related (e.g., preparation of surgical equipment, instructions for the patient pre-induction), the majority was non-care-related noise (e.g., superfluous conversations among staff). By implementing noise reduction strategies, anesthesiologists perceived overall staff-related and reducible OR noise as decreasing significantly. There was also less of a delay from patient OR entry to induction, which may be a function of less distracted OR staff. Care-related noise did not change significantly, probably because this is less avoidable. Ultimately, our anesthesiologists were able to perform their duties despite the surrounding noise; however, post education, they perceived that noise had significantly reduced adverse effects in terms of distracting or bothering them. Thus, noise reduction seems to be associated with improved anesthesiologist satisfaction and preparedness during the critical phase of induction.

Interestingly, even pre-education, patients did not report poor satisfaction with noise in general, including noise specifically related to conversations or surgical equipment. They also reported a pleasant time going to sleep, both pre- and postintervention. Our results contrast with those of a previous study, which measured an average noise level of 70.3 dB(A) at induction and reported that 52% of patients surveyed would have preferred a quieter environment, and 16% found the noise distressing.4 Patient satisfaction was overwhelmingly high across the board. Nevertheless, between the pre- and post-education cohorts, there was a slight statistically significant decrease in some satisfaction scores; regardless, the scores remained high at ≥ 4.4 on the five-point scale. We postulate that subtle baseline differences between the patient populations pre- and postintervention (Table 3) and/or chance alone may contribute to these observations.

There are several limitations to consider. First, this was a QI project based on a limited sample size of convenience and not a randomized controlled trial (RCT). Hence, confounding factors may be present. Second, because of logistical challenges following patients through their perioperative course, participants were recruited separately for parts A and B. As such, data collected for each part may not pertain to the same operative cases. Third, as there are no tools specifically validated to measuring patient experience of perioperative noise, we conceived questionnaire items with similar nomenclature and formatting as Likert scale-based questionnaires, which are most consistently utilized and validated for evaluating patient satisfaction.4, 20, 26 Fourth, while patients were recruited for part B postoperatively to reduce the risk of biasing their experiences, there was a risk of selective memory. In turn, anesthesiologists who completed questionnaires for part A were aware of whether they were in the pre-intervention or post-intervention phase, potentially biasing their results. Fifth, the nature of our exclusion criteria may have led to selection bias regarding the types of surgical procedures; for example, at our centre, orthopedic procedures tend to be conducted with regional anesthesia and sedation, which would have been excluded as general anesthesia was not the primary anesthetic. Further, while the present project of our QI initiative was dedicated to the domain of noncardiac surgery, we acknowledge that noise in the cardiac surgery OR is an important issue that warrants similar (and dedicated) attention. Sixth, despite formal education on proper placement and timing of the decibel meter, there may be inter-user variability among anesthesiologists. We did not take into account the number of peak noise events, which may have been relevant when observing maximum dB(A) levels. Similarly, we did not take into account trainees taking part in the induction period, which may contribute to more conversations, repeated attempts at intubation, or rescue equipment needing to be assembled in rapid fashion. Seventh, while we applied a two-week latency before collecting post-intervention phase data, we do not know how sustained the observed benefits are. Our project was not designed to evaluate the effectiveness of noise reduction, retention of education, or any particular educational technique, which would be useful points to elucidate in future research. The results of our project showed that noise during induction may impact anesthesiologist satisfaction and ability to perform essential tasks; however, given the above combined with the limited sample size, this warrants further investigation. Lastly, we did not assess trough noise levels, the implications of OR noise during extubation or other phases of the perioperative timeline, different anesthetic techniques such as procedural sedation, or specific ways to reduce the risk of perioperative NIHL, all of which warrant future research.

Conclusion

This QI project highlights that noise can no longer be ignored as a nonmodifiable certainty in the OR. Although patients were generally pleased with their care, anesthesiologists’ perceived that their ability to perform—and, by extension, the concept of the sterile cockpit—is negatively impacted by excessive noise in the OR. Noise is partially embedded in OR culture, but noise reduction is welcomed and achievable with education. With the potential health implications faced by those in the OR, an opportunity exists for anesthesiologists to lead the cultural shift of noise reduction, for the welfare of not only patients but also all members of the perioperative care team. This initiative will include future conduct of RCTs to establish evidence of the benefits of noise reduction in the perioperative setting.