Operating Room Noise Environment and Behavior in Children Undergoing General Anesthesia: A Randomized Controlled Trial

Background Excessive operating room noise impairs communication, distracts from monitoring equipment, and may increase patient and provider stress. Aim This study investigates the effects of reduced noise on perioperative behavior in children undergoing general anesthesia and on anesthesia provider response time. Methods Healthy children (the American Society of Anesthesiologists class I-II), 2–8 years of age, and their anesthesia providers were randomized into a control or treatment group exposed to reduced stimulation during induction and emergence. Primary outcomes were patient behavior and provider response time. Secondary outcomes were postoperative pain scores, provider responses exceeding 30 seconds, and median and maximum noise exposure. Results 64 children (27 females and 37 males) were randomized into a control or treatment group, of whom 32 (50%) underwent tonsillectomy/adenoidectomy and 32 (50%) underwent dental procedures. The average age was 4.6 (SD 1.43) years. Children exposed to reduced noise were less likely to be “fussy about eating” (p=0.042), more “interested in what goes on around them” (p=0.008), and had fewer temper tantrums (p=0.004) on postoperative day one or two and on postoperative day five, six, or seven. No other differences were found between groups in behavioral assessment scores or provider response times. Conclusions Our study is the first to show that a low-stimulus environment improves postdischarge behavior. Provider response time was unaffected by reduced noise, and the average and peak noise exposure levels did not exceed national safety guidelines. This trial is registered with NCT03507855 and NCT03504553.


Introduction
Intraoperative adverse events may represent up to 65% of all adverse events in the hospital setting, with the majority of adverse surgical events being attributed to preventable human factors [1].One of the major human factors contributing to adverse events is miscommunication, often as the result of excessive noise in the operating room (OR) [2].Te World Health Organization recommends that ambient noise in the hospital should not exceed 35 dB during the day [3].However, a review of noise levels in the OR found an average of 51 to 75 dB [4].At these volumes, medical personnel would have to speak at 95 dB (the same level produced by an MRI machine) to be understood [5].
Healthcare providers are also adversely afected by noise in the OR.More than 95% of surveyed OR staf reported negative physiological and psychological efects (i.e., agitation, inattentiveness, anxiety, fatigue, and headaches) due to excess OR noise [6].Furthermore, stress due to acute noise exposure has been demonstrated to impair cognitive function and short-term memory among medical professionals [7,8].Noise-related fatigue and worsened state of mind may further amplify mechanical or communication errors, compounding preventable adverse events.Repeated requests in the OR were fve times more likely to occur during cases playing loud music than in those not playing music [9], and high noise level in the OR has even been linked to surgical site infection [10].
A loud OR environment can distress patients, distract providers, and impair communication [2,7].Previous studies by Kain et al. have shown decreased preoperative anxiety in patients exposed to low-stimulation on induction but did not reveal any posthospitalization changes [11].Tis group also found midazolam to be superior to parental presence in reducing preoperative anxiolysis and superior (in combination with acetaminophen) at reducing negative postoperative behaviors, as compared to acetaminophen alone, while Kain et al. and Sola et al. failed to distinguish between midazolam and tablet devices [12][13][14].What has not yet been shown in the literature is whether a combination of reduced noise and chemical anxiolysis can have an impact on the posthospitalization course of pediatric patients.
Induction and emergence, the most critical periods of an anesthetic, are often associated with the highest noise levels during a surgical case [4].Our primary hypotheses were that a reduced noise environment during both induction and emergence, combined with midazolam premedication, would improve pre-and postoperative patient behavior and decrease the response time of anesthesia providers to auditory alarms.

Methods
Institutional Review Board (IRB) approval was obtained for the study at Nationwide Children's Hospital (IRB18-00203 and IRB18-00227), informed consent was obtained from patient guardians and providers prior to enrolling all participants, and the study was registered with clinicaltrials.gov(NCT03507855 and NCT03504553).Te study involved two arms: one exploring the behavioral efects of a reduced-noise environment on pediatric patients undergoing general anesthesia and the second examining the reaction times of anesthesia providers working in this environment.Te two arms initially began as separate studies but were then combined due to the interrelated nature of the subject matter.Te primary outcome of the study was the anxiety level at induction, as measured using the modifed Yale Preoperative Anxiety Scale (m-YPAS).A prior study looking at low stimulation in the OR had shown a decrease from 60 to 42 points on this metric [11].Assuming a standard deviation of 8 points, we determined that an independent t-test would have 90% power to appreciate this diference with 95% confdence, provided that a minimum of 12 patients were randomized into each group.Other outcomes were compared across the two groups using unpaired t-tests, rank-sum tests, Chi-square tests, or Fisher's exact tests, as indicated by the event rate and distribution of the endpoints.Analysis of ordinal measures of anxiety, compliance, and posthospitalization behavior was performed via Spearman correlation coefcients, and two-tailed p values less than 0.05 were considered statistically signifcant.We therefore determined that a sample size of at least 29 pairs was needed.

Study Population.
Patients 2-8 years of age, evaluated to be the American Society of Anesthesiologists (ASA) class I-II, undergoing tonsillectomy with or without adenoidectomy, tympanomastoidectomy, or single general abdominal laparoscopic or dental restoration procedure lasting at least 30 minutes were eligible for participation.Exclusion criteria included previous enrollment in the study, use of any antidepressant, anxiolytic, or pain medication other than acetaminophen or nonsteroidal antiinfammatory drugs (NSAIDs), allergy to or parent refusal of midazolam, history of emergence delirium, cardiac disease other than functional heart murmur, developmental delay, and hearing loss.All children included in this study received premedication with midazolam (0.25-0.55 mg/kg to a maximum of 15 mg).Eligible staf included anesthesia providers (certifed registered nurse anesthetists, pediatric anesthesiology fellows, or attending anesthesiologists) providing general anesthesia for patients who met inclusion criteria, unless they had been previously enrolled in the study.

Interventions.
Te children were randomized into one of two groups: the control group under typical OR conditions and the experimental group with reduced OR personnel, low ambient light, muted communication devices, and soft background music during the induction of and emergence from anesthesia (Bach's Air on the G String).A blocked randomization list was generated online at https:// www.sealedenvelope.comusing block sizes of two and four.Te providers were not specifcally randomized or blinded but instead followed whichever group to which their patient had been assigned.No losses or exclusions occurred following randomization.Research staf generated the random allocation sequence, enrolled participants, and assigned the participants to interventions.Patient enrollment began on October 26th, 2018, and follow-up concluded on November 29th, 2021.Te trial ended when an adequate number of participants had completed the study in order to appropriately power the analysis.Only the anesthesia providers interacted with the patient during induction.No parents or caregivers were present during induction.A microphone secured near the patient's head or torso on the operating room table or to an IV pole at the head of the bed measured the decibel level continuously from the time the patient entered the OR until they were taken to the recovery room.Potential diferences in results based on microphone location were not specifcally evaluated.Local anesthetic was utilized by the surgeon or dentist for fve of the patients from each group.
A blinded observer, using the m-YPAS instrument, evaluated each child before receiving premedication with oral midazolam, and again during induction 15-45 minutes later.Behavior during induction was also assessed using the Induction Compliance Checklist (ICC) instrument.Anesthetic management other than premedication was determined by the attending anesthesiologist, and specifc data regarding management choices for induction and maintenance of or emergence from anesthesia and depth of anesthesia throughout each procedure were not recorded.Postoperatively, patient behavior was assessed by a blinded observer for pain and delirium using the Face, Legs, Activity, Cry, and Consolability (FLACC) and Pediatric Anesthesia 2 Anesthesiology Research and Practice Emergence Delirium (PAED) scales, respectively, every 10 minutes while the patient was in phase I recovery.Frequency and quantity of analgesic medications administered during phase I recovery were also recorded.All patients in the study recovered in one of two postanesthesia care rooms, where noise levels were not recorded or controlled for.In order to capture potential behavioral changes previously demonstrated to occur on postoperative day two, blinded patient guardians were asked to assess patient behavior using the modifed Posthospitalization Behavior Questionnaire (PHBQ) instrument on postoperative day one or two, and again on postoperative day fve, six, or seven (a total of two times for each patient).We did not record which parent or guardian made these separate assessments.
During each case, at a single randomized time coinciding with induction, an audio stimulus meant to simulate a patient monitor alarm would emit from a device secured next to the real monitors.Te device emitted a sound 80-90 dB at a range of 0.5 meters.Te response time of anesthesia providers to silence this alarm was recorded in each instance.

Data Analysis.
Continuous data were presented as means with standard deviations or medians with interquartile ranges.Categorical variables were presented as frequencies and percentages.Te Mann-Whitney test was used to compare the reaction time, noise level, ICC score, and m-YPAS score.Te Wilcoxon test was used to compare the reaction time of providers who had participated in both the control and reduced noise groups.Te repeated measures' analysis of variance (RM-ANOVA) test was used to compare the PAED, FLACC, and PHBQ scores.All statistical analysis was performed using GraphPad Prism 9.0.0 (GraphPad Software, San Diego, CA).

Results
Characteristics of study participants: the unblinded anesthesia provider cohort included 36 anesthesia providers (34 nurse anesthetists, one pediatric anesthesiology fellow, and one anesthesiologist), administering general anesthesia for patients undergoing tonsillectomy/adenoidectomy (n � 32 and duration 22-60 min) and dental procedures (n � 32 and duration 16-105 min).Although the study protocol allowed a greater variety of procedures, these case types were ultimately the most accessible due to the availability of the investigative staf.Twelve of the providers participated in both the control and reduced noise groups, the remaining 24 participated in only one group or the other.Te patient study cohort included 64 patients, 10 (32.3%) females and 21 (67.7%)males in the control group and 17 (51.5%)females and 16 (48.5%)males in the reduced noise group.Diferences in patient sex were not statistically signifcant (p � 0.119 using Chi-square test).Patients in the control group were 2.4-7.8years, and those in the reduced noise group were 2.2-7.9 years (Table 1).All patients, but one in the control group and all patients in the reduced-noise group, spoke English as their frst language.Tere were no statistically signifcant diferences between participants in the two groups.
Noise level and patient behavior: the median noise level in the OR was signifcantly higher in the control group, as compared to the reduced noise group (59 versus 56 dB, respectively, p � 0.008; Supplemental Figure 1).However, the maximum noise level measured was not signifcantly diferent between the two groups (90 versus 89 dB, respectively, p � 0.556).Te provider response time to simulated audio monitor alarms was shown to be nonsignifcant (p � 0.421 for the unpaired test including all providers, p � 0.092 for the paired test including only those providers who participated in both the control and reduced noise groups; Supplemental Figure 2).Twelve of the 36 providers were included in the paired analysis (Table 2).
Te ICC, m-YPAS, FLACC, and PAED instruments showed no statistically signifcant diference in behaviors between children in the reduced noise and control groups during the preinduction, induction, and emergence periods (p � 0.087, 0.272, 0.856, and 0.366, respectively; Figures 1, 2,  and 3).Assessments for FLACC and PAED were discontinued once patients left phase I recovery.Te discharge criteria at our institution are uniform across all locations where anesthesia is provided and include physiologic metrics, as well as a minimum time duration for all patients receiving intravenous opiates, nebulized epinephrine, or medication reversal (such as naloxone), as well as for patients for whom the endotracheal tube is removed in the recovery room.Shorter recovery times were incidentally observed for patients in the reduced noise group, although this was not a variable included in our ofcial study protocol, and it was not analyzed statistically.Te PHBQ assessment, however, did demonstrate signifcant improvements for patients regarding fussiness about eating, interest in things happening around them, and frequency of temper tantrums, as assessed on postoperative day one or two, and again on postoperative day fve, six, or seven (p � 0.042, 0.008, and 0.004, respectively; Figure 4).Patients in the reduced noise group also required, on average, fewer doses of pain medication and lower total quantity of pain medication in phase I recovery, as compared to those in the control group (Table 1).

Discussion
Young children undergoing adenotonsillectomy or dental restoration had improved postdischarge behaviors when exposed to a reduced noise environment in addition to preoperative oral midazolam anxiolysis.Te reduced noise environment did not infuence the response time to a simulated monitor alarm of the anesthesia providers caring for these patients during the induction of anesthesia.
Reducing preoperative anxiety and improving the quality of induction for children are large components of care for a pediatric anesthesiologist.Kain et al. showed a decrease in preoperative anxiety for patients exposed to low-stimulation conditions during induction without other preinduction anxiolytic measures (e.g., midazolam or parental presence at induction) [11], after previously showing premedication with midazolam to be superior to parental presence in alleviating preoperative anxiety [12].Sola  to show a diference in anxiolysis between premedication with midazolam and tablet devices [13,14].Our results indicate that there is no further beneft in reducing anxiety during the induction of anesthesia by adding a low-stimulus environment provided that preoperative midazolam is administered.While we did not test the benefts of distraction with a tablet device or the efects of other anxiolytic medications, given their comparable benefts to oral midazolam at reducing preoperative anxiety in previous studies, we would not anticipate any further beneft from a low-stimulation environment as well.
Emergence agitation (EA) is a behavioral disturbance during the recovery from anesthesia, consisting of hallucinations, delusions, and involuntary physical activity, which is commonly seen in young children.Te magnitude and severity of EA can be measured by a validated scoring system,  4 Anesthesiology Research and Practice such as the PAED [15].Factors linked to an increased risk of EA include age, preoperative anxiety, pain, patient personality, surgical procedure, and anesthetic type.Perioperative administration of midazolam, alpha-2 agonists, opioids, ketamine, and nonsteroidal anti-infammatory medications has been shown to decrease the incidence of EA [16].In our study, there was no beneft to a low-stimulation environment on pain scores or the incidence of EA.Te original PHBQ, developed in 1966, was modifed and revalidated, showing that children are most likely to  6 Anesthesiology Research and Practice demonstrate behavioral changes on postoperative day two [17].In our study, postoperative behavior in children exposed to a reduced stimulation operating room environment was better on postoperative day one or two, and on postoperative day fve, six, or seven, despite no diferences in preoperative anxiety, postoperative pain, or EA.Since the caregivers for our patient participants were contacted on postoperative day one or two, and again on postoperative day fve, six, or seven, it follows that our screening is likely to have captured behavioral changes exhibited by the children.
Our fndings difered from previous works that had shown a correlation between posthospitalization behavioral changes and preoperative anxiety [17], but no beneft for a low-stimulation environment when midazolam was not administered [11].Given that we did not see a diference between the groups for either pre-operative anxiety or EA, the reason why the addition of a low-stimulation environment resulted in a reduction in maladaptive behaviors is unclear and warrants a larger study with more tightly controlled environments.Te hypothesis that the response time to a simulated audio alarm in the OR would be signifcantly decreased by reducing the ambient noise level was not confrmed.Despite a 2 dB reduction in the median noise level, which represents a 77% decrease in sound intensity, provider response time was not altered.Although adverse patient events and communication errors were not specifcally tracked in our study, the previous work by Crockett et al. suggests that reductions in those occurrences might be achieved through signifcant noise reduction during critical moments of the anesthetic [2].Survey results from that initiative determined that, at a minimum, there was a subjective belief held by anesthesia providers that noise played a signifcant role in OR distractions.
Median and peak noise levels did not exceed the National Institute for Occupational Safety and Health (NIOSH) guidelines, which call for no more than eight hours of noise exposure at 85 dB or four hours of exposure at 88 dB [18].Neither of these levels was achieved during our study when total exposure time was calculated.Katz described that an upper limit of 55 dB would be needed in order to provide "universally safe conditions" in the workplace [19].Our reduced noise group achieved a median level of 56 dB, which unfortunately still exceeds that threshold.Terefore, while our anesthesia providers were not at signifcant risk for longterm hearing loss due to noise exposure in the OR, there was still a potential for other adverse outcomes.
Our study had a number of limitations.All patients in our study underwent surgery of the head (i.e., in close proximity to the auditory system), which could have introduced confounding efects into our results.It would be interesting in future studies to include a greater variety of procedure types to further investigate those efects.Similarly, we included in our protocol children undergoing tympanomastoidectomy, which is a procedure often performed on individuals with underlying auditory pathology.Even though hearing loss was one of our specifc exclusion criteria, and no enrolled study participants actually underwent this procedure, it may have been more appropriate to exclude this procedure from our protocol entirely.Although we noted a statistically signifcant diference in the noise levels between the two study cohorts, the causes of the increased noise (e.g., conversations away from the operating room table or activities such as opening equipment and OR preparation) were not investigated and identifed.In addition, the noise levels in the recovery room were not measured or controlled for.It is possible that statistically signifcant diferences existed between the two groups, adding a confounding variable.Te caregivers for our patients were contacted for PHBQ assessments on postoperative day one or two, and again on postoperative day fve, six, or seven, but it is possible that behavioral changes in the days between may have been notable.In addition, we did not record which caregiver was contacted on each occasion, thereby potentially introducing another confounding variable if a diferent observer reported on each day.For the purposes of this study, we assumed that the provider reaction time to a simulated patient monitor alarm correlated to the reaction time to a genuine monitor alarm, indicative of a precursor or in-process patient critical event.Since it was not possible to blind the providers, this may refect, in part, the Hawthorne efect, whereby individuals increase their performance when they are aware that they are subjects in an experiment.In our study, while not signifcant, we saw an inversely proportional trend in reaction speed to silencing a simulated alarm with years of experience.Tis may in fact refect the enhanced ability of a more experienced provider to tune out unimportant (albeit intrusive) stimuli, such as a fake alarm, while focusing on their patient during critical points of the anesthetic.For this reason, if we had succeeded in establishing statistically signifcant diferences in response times to this alarm, it still might not have elucidated the answer to the question we ultimately sought: does a quieter OR provide a safer anesthetic environment for our patients?Future studies should focus on whether noise reduction afects more complex interactions, such as communication between members of the anesthesia team or between the anesthesia and surgical teams.
In conclusion, we have shown a beneft in postoperative maladaptive behavior in young children exposed to a lowstimulation environment, in combination with oral midazolam.Additional studies are needed to further delineate the various potential efects of a reduced noise environment for both pediatric patients and their anesthesia providers during the critical moments of induction and emergence.
Midazolam is superior to parental presence (but not superior to tablet devices) in reducing preoperative anxiety.What does this article add? Patients given midazolam premedication followed by low stimulation on induction experience posthospitalization behavioral improvement, in addition to decreased preoperative anxiety.Tis trial is registered with Clinicaltrials.gov(NCT03507855 and NCT03504553).

Figure 1 :
Figure1: ICC and MYPAS results.ICC score represents the number of negative behaviors displayed by patients during induction (up to a maximum of 10).m-YPAS scores represent a weighted amalgam of fve categories of behavior: activity, vocalizations, expressivity, arousal, and involvement of parents, with higher numbers associated with greater anxiety.

Figure 2 :
Figure2: FLACC results.Each of the fve metrics of the scale is scored from zero to two, for an overall rating of zero representing no expected pain to ten representing likely excruciating pain.Data for the reduced noise group were not recorded at the 60-and 75-minute marks due to patient discharge from phase I recovery.

Figure 3 :Figure 4 :
Figure3: PAED scale results.Individual scores are graded subjectively from zero to four, with a higher score being associated with greater degree of delirium.Data for the reduced noise group were not recorded at the 60-and 75-minute marks due to patient discharge from phase I recovery.

Table 2 :
OR noise levels and provider reaction times.
* Only the 12 providers who participated in both control and reduced noise groups were included.

Table 1 :
Provider and patient demographics and premedication dosing.