Amnesia of the Operating Room in the B-Unaware and BAG-RECALL Clinical Trials

Abstract

BACKGROUND: 

Patient memories of the operating room (OR) may serve as the informational basis for assessing satisfaction with individual anesthesiologists. Furthermore, the provision of clinically important information may assume that perioperative memories are retained. Studies assessing the extent of perioperative amnesia and factors associated with perioperative amnesia are sparse. Therefore, we assessed patient amnesia of the OR and of the preoperative holding area in hospitals where midazolam is typically administered in the preoperative holding area and evaluated whether midazolam dose administered in the preoperative holding area and patient age were associated with amnesia of the OR before induction of anesthesia.

METHODS: 

This was a retrospective study among 7750 adult patients who had general anesthesia and participated in the B-Unaware and Bispectral Index or Anesthetic Gas to Reduce Explicit Recall (BAG-RECALL) clinical trials. The last location the patient remembered before induction of anesthesia and the first location they remembered after induction of anesthesia were determined through a modified Brice questionnaire administered over the phone 30 days postoperatively. Regarding the preoperative period, patients were excluded if their last memory was unclear with respect to location before induction of anesthesia or if they were recruited at Winnipeg, where midazolam was typically first administered in the OR. Midazolam dose (mg/kg) administered in the preoperative holding area was divided into quartiles. Poisson regression models were used to calculate age- and multivariable-adjusted odds ratios (95% confidence intervals [CIs]) for the association between midazolam dose and amnesia of the OR before induction of anesthesia.

RESULTS: 

Of the 5339 patients included, 59.5% (95% CI, 58.2–60.9) of patients had amnesia of the OR before induction of anesthesia. In addition, 44.1% (95% CI, 42.8–45.7) last remembered the preoperative holding area, and 15.4% (95% CI, 14.4–16.4) only had preoperative memories before the holding area. The percentages of patients with amnesia of the OR before induction of anesthesia differed according to age groups: 50.7% (95% CI, 47.7%–53.7%) in patients aged 18 to 47 years versus 70.0% (95% CI, 67.0%–72.9%) in patients aged 73 to 99 years. Patients in the highest midazolam quartile had an adjusted prevalence ratio of 1.31 (95% CI, 1.22–1.42) for amnesia of the OR compared with those who did not receive midazolam.

CONCLUSIONS: 

In hospitals where patients typically receive midazolam in the preoperative holding area, the majority of patients do not remember the OR, and a clinically relevant number of patients does not remember the preoperative holding area. If additional studies produce results indicating that a substantial proportion of patients has amnesia of the anesthesiologist, these findings would argue against the validity of assessing patient satisfaction with individual anesthesiologists providing exclusively OR care in such hospitals. Furthermore, if additional studies yield findings suggesting patient amnesia of the preoperative holding area, these results would suggest reconsideration of providing clinically important information only in the preoperative holding area. Older age and midazolam-induced anterograde amnesia are probably associated with impaired perioperative memories.

In the United States, health care policy changes have led the American Board of Anesthesiology, medical centers, and health care funding agencies to recommend consideration of individual practitioner assessment into patient satisfaction surveys.^1–4 These assessments have potential implications for continuing accreditation and reimbursement of practitioners.^1–4 However, patient satisfaction with individual anesthesiologists is linked to memory, and many anesthesiologists spend most of their clinical time in the operating room (OR).⁵ Thus, valid assessment of patient satisfaction with an individual anesthesiologist assumes that the patient remembers interaction with the anesthesiologist in perioperative locations. Furthermore, patients’ ability to remember events in the preoperative holding area may have implications for the provision and retention of medical information.

Although patients are known to have superior recall of the hospital ward compared with the postanesthesia care unit,⁶ there have been only a few small studies with findings pertaining to the extent to which patients do not remember being in the OR before induction of anesthesia or the preoperative holding area as well as factors associated with not remembering these locations.^7,8 Furthermore, many anesthesiologists routinely administer anxiolytic medications such as midazolam preoperatively, which unpredictably produce anterograde amnesia.⁹ It is probable that several factors such as age, gender, body habitus, pharmacological interactions, and time from administration are associated with substantial patient-to-patient variability in midazolam concentration within the brain, which may in turn affect the onset and duration of amnesia.^10–12 Consequently, it is unclear for individual patients whether a particular dose of midazolam, administered in the preoperative holding area, will exert amnesic actions or how long amnesia will persist.^10–12

Thus, patient amnesia of the perioperative period may have implications for both clinical practice and health care policy. In terms of clinical practice, if a substantial proportion of patients do not remember perioperative events, it could be helpful for clinicians to set expectations for patients and their families about possible amnesia and to be cognizant of the implications for communicating important medical information. In terms of health policy, if a substantial portion of patients indeed has amnesia of the perioperative period, this finding would argue against the validity of conducting patient satisfaction surveys of individual OR-based anesthesiologists (when compared with the anesthesia group) at hospitals where midazolam is typically administered in the preoperative holding area. If, owing to amnesia, a substantial portion of patients was unable to participate meaningfully in a satisfaction assessment process, such a process would be rendered uninterpretable.^a13

Thus, the aims of our retrospective study, conducted in hospitals where midazolam is typically administered in the preoperative area, were 3-fold. First, we assessed whether most (i.e., >50%) patients within a large surgical cohort had amnesia of the OR before induction of anesthesia. Second, we wanted to ascertain whether a substantial portion (i.e., >10%) of patients within a large surgical cohort had amnesia of the preoperative holding area. Third, we explored whether old age and midazolam administered in the preoperative holding area were associated with amnesia of the OR before induction of anesthesia.

Our hypothesis 1 was that most patients would not remember the OR before induction of anesthesia. Our hypothesis 2 was that a substantial portion of patients would not remember the preoperative holding area. Our hypothesis 3 was that both older age and higher midazolam doses would be associated with increased amnesia of the OR before induction of anesthesia.

METHODS

Study Population

This was a retrospective study among participants from the B-Unaware¹⁴ and the Bispectral Index (BIS) or Anesthetic Gas to Reduce Explicit Recall (BAG-RECALL) trial,¹⁵ which were 2 randomized clinical trials designed to test different anesthetic protocols for the prevention of intraoperative awareness with explicit recall (AWR). The design and methods of both studies have been described in detail elsewhere.^14–18 In brief, the B-Unaware trial was a single-center, randomized clinical trial designed to test whether a BIS-guided anesthesia protocol was superior to a protocol based on end-tidal anesthetic concentration alerts for decreasing AWR.¹⁴ Between September 2005 and October 2006, 2000 eligible patients aged ≥18 years who were at high risk for AWR and undergoing elective surgery with general anesthesia were recruited at Washington University in St. Louis.¹⁴ Patients with dementia, a history of stroke with residual neurological deficits, or who were unable to provide written informed consent were excluded from the study.¹⁴ Furthermore, patients were excluded if BIS monitoring was not possible as a result of the surgical procedure or positioning of the patient or if the surgery required a wakeup test.¹⁴ Applying similar eligibility criteria, the BAG-RECALL trial was a randomized multicenter clinical trial designed to evaluate whether BIS protocol was superior to a structured end-tidal anesthetic concentration protocol for reducing the incidence of AWR among patients at high risk for this complication.¹⁵ Between May 2008 and May 2010, 6100 patients aged ≥18 years were recruited from 3 medical centers, Washington University in St. Louis (n = 4700), the University of Chicago (n = 600), and the University of Manitoba in Winnipeg (n = 800).¹⁵ In both clinical trials, the patients’ medical and comorbidity status was assessed at baseline and patients were followed up to 30 days postoperatively for primary and secondary outcome ascertainment.^14,15 For the purpose of our analyses, we pooled the data from 1941 B-Unaware and 5809 BAG-RECALL participants who underwent randomization, for 7750 participants. The B-Unaware and BAG-RECALL clinical trials were approved by the institutional review boards of all participating institutions. No further institutional review board approval was obtained for this retrospective study, which was conducted under the umbrella approval of the parent trials.

Midazolam Administration

IV midazolam administration was part of routine clinical care in both clinical trials. At Washington University in St. Louis and the University of Chicago, midazolam was typically administered in the preoperative holding area, whereas at the University of Manitoba, midazolam was typically administered in the OR. Because our study primarily focused on hospitals where midazolam was administered in the preoperative holding area, participants from the University of Manitoba were excluded from our primary analyses. For patients from Washington University in St. Louis and the University of Chicago who were included in our formal statistical analysis, we divided midazolam dosage (mg/kg) into quartiles: quartile 1 (≤0.020 mg/kg), quartile 2 (>0.020–0.028 mg/kg), quartile 3 (>0.028–0.047 mg/kg), and quartile 4 (>0.047 mg/kg). In our analyses, we established the following midazolam categories: no midazolam, quartile 1, quartile 2, quartile 3, and quartile 4. Actual body weight was used for dosage computations.

Perioperative Memory Assessment

In both clinical trials, a structured telephone interview using the modified Brice questionnaire was performed at approximately 30 days postoperatively to assess AWR.¹⁹ The modified Brice questionnaire is a 5-item instrument assessing patients’ perioperative memories and experiences, but it is not specifically validated for ascertaining the locations of these recollections (Appendix 1). Within the modified Brice questionnaire, we used patient responses to the question “What was the last thing you remembered before going to sleep?” to assign the location of the patient’s last memory before induction of anesthesia. Two researchers independently abstracted the information from the questionnaires and assigned patient responses to the following predefined locational categories: OR, preoperative holding area, and location before the preoperative holding area. Because patient responses to some of the questionnaires (especially from the B-Unaware trial) were handwritten notes within paper charts, 2 investigators independently performed the abstraction process for these patients. For most of the patients in the BAG-RECALL trial, discrete responses were coded in a database, limiting ambiguity. Disagreements were resolved by the principal investigator when possible. Ambiguities were coded as inconclusive, and these patients were excluded from the analysis. For our main analyses, we collapsed the 3-level categorical variable into a binary outcome variable and distinguished between patients “remembering the OR before induction of anesthesia” and patients “not remembering the OR before induction of anesthesia.” Patient responses that were inconclusive with respect to location were excluded from our statistical analyses. A more in-depth description of this coding process is included in Appendix 2.

Similarly, within the modified Brice questionnaire, we used patient responses to the question “What was the first thing you remembered after waking up?” to assign the location of the patient’s first memory after induction of anesthesia. An analogous approach was used to abstract, categorize, and dichotomize the patient responses as well as exclude those that were inconclusive with respect to location.

In addition to assessing for AWR at 30 days postoperatively, the modified Brice questionnaire was also used to assess for AWR at 1 to 3 days postoperatively. Our investigators used an analogous approach to assign the location of patients’ last memories before induction of anesthesia and first memories after induction of anesthesia. These results are included in Appendix 3.

Statistical Analyses

We calculated the percentages of patients who had amnesia of the OR before induction of anesthesia and holding area along with the corresponding 95% confidence intervals (CIs). We also stratified the percentages of patients who had amnesia of the OR before induction of anesthesia according to age quintiles.

Analysis of variance for continuous and Mantel-Haenszel χ² tests for categorical variables were used to compare patients’ baseline characteristics with respect to the various midazolam dosage categories. We used Poisson regression models with robust variance estimation to calculate age- and multivariable-adjusted prevalence ratios for amnesia of the OR before induction of anesthesia according to midazolam quartiles with no midazolam serving as the reference category. We adjusted the multivariable models for the following variables chosen a priori: age (continuous), sex, study (B-Unaware or BAG-RECALL), American Society of Anesthesiologists physical status (I–II or III–IV), chronic benzodiazepine use (yes or no), chronic opiate use (yes or no), and daily alcohol use (yes or no). We also created a second multivariable model that adjusted for additional variables that could serve as confounders between midazolam dose administered in the preoperative holding area and amnesia of the OR before induction of anesthesia. These variables were body mass index (continuous), smoking status (current smoker or not current smoker), diabetes (yes or no), hypertension (yes or no), cardiovascular disease (yes or no), obstructive sleep apnea (yes or no), cerebrovascular disease (yes or no), and type of surgery (cardiothoracic or other).

To explore whether the excluded patient population differed from those patients who were included in our statistical analyses, we compared baseline characteristics between the 2 populations by using χ² tests for categorical and t tests for continuous variables.

Secondary Analyses

Descriptive statistics for amnesia of the OR, along with the corresponding 95% CIs, are presented separately for patients from the University of Manitoba in Winnipeg. We also calculated the percentages of patients who had amnesia of the OR after induction of anesthesia along with the corresponding 95% CIs.

Sensitivity Analysis

A sensitivity analysis of preoperative memories was performed by calculating the percentage of patients with amnesia of the OR before induction of anesthesia after assuming that all excluded patients remembered the OR before induction of anesthesia.

All statistical analyses were performed by using SAS 9.4 (SAS Institute Inc., Cary, NC). All P values are 2-tailed. Each P value <0.05 was considered statistically significant. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology statement.²⁰

RESULTS

Of the 7750 participants, we excluded 24 patients with missing responses to the modified Brice questionnaire, 10 patients with missing midazolam dosage information, 1758 patients whose responses were inconclusive with respect to preoperative location, and 619 patients from the University of Manitoba, leaving us with 5339 participants for our statistical analyses (Fig. 1).

Figure 1: Exclusion criteria. BAG-RECALL = Bispectral Index or Anesthetic Gas to Reduce Explicit Recall.

Table 1: Percentage of Patients with Amnesia of the OR Before Induction of Anesthesia According to Age Quintiles (n = 5339)a

Table 2: Associations Between Midazolam Categories and Amnesia of the OR Before Induction of Anesthesia (n = 5339)a

Table 3: Baseline Characteristics According to Midazolam Quartiles (n = 5339)

Table 4: Baseline Characteristics Comparison of the Included and Excluded Study Populations (n = 7750)

Among the 5339 patients, 59.5% (95% CI, 58.2–60.9) had amnesia of the OR before induction of anesthesia. In addition, 40.5% (95% CI, 39.2%–41.8%) last remembered the OR; 44.1% (95% CI, 42.8%–45.5%) last remembered the preoperative holding area; and 15.4% (95% CI, 14.5%–16.4%) only had memories before the preoperative holding area. The percentages (95% CI) of patients with amnesia of the OR before induction of anesthesia varied among age groups and midazolam doses (Tables 1 and 2). For example, 50.7% (47.7%–53.7%) of patients in the lowest age quintile (18–47 years) had amnesia of the OR before induction of anesthesia versus 70.0% (67.0%–72.9%) in the highest age quintile (73–99 years) (Table 1). Similarly, increasing midazolam quartiles were associated with increasing prevalence ratios of amnesia of the OR before induction of anesthesia. After adjusting for age, sex, study, American Society of Anesthesiologists physical status, benzodiazepine use, opiate use, and alcohol consumption, patients in the highest midazolam quartile had a prevalence ratio of 1.31 (95% CI, 1.22–1.42) for amnesia of the OR compared with those who did not receive midazolam. The effect estimates did not substantially change after adjusting for additional possible confounding factors (model 2). The baseline characteristics with respect to midazolam quartiles are presented in Table 3. Compared with patients who did not receive midazolam, patients in the highest midazolam quartile in general were younger with fewer comorbidities. They were also more likely to have a history of benzodiazepine, opiate, alcohol, and cigarette use (Table 3). Compared with the patients included in our statistical analyses, the excluded patient population tended to be older and less likely to have received midazolam (Table 4).

Secondary Analyses

Regarding the postoperative period, 94.4% (95% CI, 93.7–95.2) of 3601 patients had amnesia of the OR after induction of anesthesia. At the University of Manitoba in Winnipeg, 24.9% (95% CI, 21.6–28.4) of 619 patients had amnesia of the OR before induction of anesthesia. With regard to the postoperative period at Winnipeg, 96.9% (95% CI, 94.8–98.1) of 445 subjects had amnesia of the OR after induction of anesthesia.

Sensitivity Analysis

For patients’ last memories before induction of anesthesia, we excluded 2411 participants from an original study population of 7750 subjects. If all of these excluded patients remembered the OR before induction of anesthesia, then 41.0% (95% CI, 39.9%–42.1%) of patients would have amnesia of the OR before induction of anesthesia.

DISCUSSION

Summary of Main Findings

In this large retrospective study of patients at high risk for AWR undergoing procedures at hospitals where midazolam is typically administered in the preoperative holding area, most patients did not remember being in the OR. Based on the sensitivity analysis, it may be likely that a substantial portion of patients, at the very least, do not remember the OR. A clinically relevant proportion of patients did not remember the preoperative holding area. The proportions of patients with amnesia of the OR before induction of anesthesia increased with increasing age and midazolam dosage.

Potential Implications

Our study findings are consistent with the results of 2 earlier studies assessing patient recognition of anesthesiologists.^21,22 Although further investigation into patient perioperative memories is needed, mostly as a result of the sparsity of literature in the field, our results may provide preliminary evidence, suggesting that a substantial portion of patients have amnesia of the OR. Because the modified Brice questionnaire does not specifically ask about patient memories of the anesthesiologist, our study cannot determine with objective certainty whether patients remembered the anesthesiologists who handled their surgical cases. If additional studies using instruments with specific questions about patient memories of the anesthesiologist reinforce our preliminary evidence for patient amnesia of the OR with results indicating patient amnesia of the anesthesiologist, then our findings would argue against the validity of assessing patient satisfaction with individual anesthesiologists. If future studies yield results consistent with amnesia of the anesthesiologist among substantive proportions of patients receiving midazolam, then our findings, in combination with pre-existing evidence suggesting that patient satisfaction may depend more on nonintraoperative events (e.g., delays in surgical start times), would support the prudence of assessing patient satisfaction with anesthesia groups on a managerial level rather than with individual anesthesiologists.^b23–27 For instance, in the case of delays in surgical start times, patient satisfaction with anesthesia groups could be assessed within the context of evaluating anesthesia groups’ use of statistical methods that reduce patient waiting times.^28–30

From a clinical standpoint, if future studies similarly show that a substantial portion of patients only remembers events before the preoperative holding area, then our findings would suggest reconsideration of providing any clinically relevant information to patients only in the preoperative holding area. Instead, it may be prudent to reinforce information provided in the preoperative holding area with written material and communication with family members if appropriate.

Even if clinicians wanted to ensure that their patients did remember the OR before induction of anesthesia, our results do not suggest that simply omitting midazolam before the OR would invariably accomplish this. Even among young patients, those who first received midazolam in the OR (at Winnipeg), and those who did not receive midazolam at all, there was a large proportion of patients with amnesia of the OR (Table 2).

Comparison with Other Studies

Although other studies have not specifically focused on patients’ last preoperative memories, there has been investigation in small studies of memories of preoperative education, consent, information regarding potential risks and complications, and memory of the OR. For example, investigators found that approximately half of patients who demonstrated preoperative understanding of key educational facts regarding bariatric surgery did not remember these educational facts postoperatively.³¹ An Australian study found that even when information on anesthetic-related risks was provided to patients 2 weeks before surgery, one-fourth of them could not remember any of the risks, one-third of them could not recall any of the major risks, and >84% of patients could not remember any of the minor risks on the day of surgery.³² Similar results have been found in relation to patients’ limited retention of memories regarding surgical risks.^33,34 Interestingly, some studies have found that there might have been a bias in information retention in favor of potential benefits of surgery as opposed to adverse outcomes.^34,35 Some of the memory limitations might have been attributable to poor comprehension; both comprehension and memory could have been affected by older age and lower education level.^36,37 Similarly, recall of preoperative information has been found to be worse with emergency than with elective orthopedic surgeries.³⁸

Consistent with our findings as well as previous studies demonstrating the pharmacokinetic and pharmacodynamic influences on patient-to-patient variability in duration of midazolam-induced amnesia,^10–12 a French study found that 50% of patients had anterograde amnesia after receiving midazolam,³⁹ and studies that have examined memories of the OR specifically have found that preoperative midazolam is associated with approximately one-third to two-thirds of patients not remembering the OR before induction of anesthesia.^7,8 In contrast to many reports, one small qualitative study from Hong Kong did find that independent of age, sex, education level, occupation, and the modality of anesthesia, the majority of patients recalled information on risks provided by anesthesiologists during the preoperative visit.⁴⁰ It is possible that a combination of written and verbal information could enhance patients’ preoperative memories.⁴¹

In general, studies seem to suggest that when patients are about to undergo surgery, memory retention may be suboptimal. In particular, older age seems to be associated with impaired retention of preoperative memories,^36,37 and preoperative midazolam administration has been demonstrated to provide anterograde amnesia, which may often include memory impairment of being in the OR.^7,8,39 Our study enhances the body of literature in this field by corroborating many of these published findings with a large participant population containing thousands of patients.

Plausible Explanations/Mechanisms for Our Study Results

Our study suggests that amnesia of the OR before induction of anesthesia may increase in frequency with age. One plausible explanation is a decline in cognitive function associated with the aging process. For instance, increased age may be associated with decreased ability to form new memories and recall earlier events.^42,43 Furthermore, elderly patients are more likely to suffer from comorbidities such as dementia and cerebrovascular disease, which could adversely affect a patient’s memory. However, patients with a history of dementia and those who were not able to provide written informed consent were excluded from both clinical trial populations. In addition, because the majority of patients in the youngest age group did not remember the OR before induction of anesthesia, this finding suggests that age, although likely an important factor, may not be the only explanation for the large proportion of patients with amnesia of the OR before induction of anesthesia.

Although age is an unmodifiable factor, midazolam dose may be a modifiable factor that is associated with amnesia of the OR. Our data suggest that increased midazolam dose administered in the preoperative holding area is associated with amnesia of the OR before induction of anesthesia, which is plausible given that midazolam is typically associated with anterograde amnesia.⁹ Further studies are needed to better understand the unmodifiable and modifiable factors that contribute to amnesia of the OR before induction of anesthesia.

Interestingly, at Winnipeg, where midazolam was typically first administered in the OR, approximately one-fourth of the patients had amnesia of the OR before induction of anesthesia. This finding supports the hypothesis that timing of midazolam administration impacts preoperative recall. Another intriguing result of our study is that 15.4% of included participants only had memories before the preoperative holding area. These findings are not likely explained by midazolam administration because midazolam typically does not cause retrograde amnesia.⁹ Instead, because it is likely that many of these patients may have had strong emotions before undergoing surgery, one plausible mechanism for these two findings may be emotion-induced retrograde amnesia.^44,45 Although the mechanisms of emotion-induced retrograde amnesia have been studied within research-centered environments,^44,45 this phenomenon may be worthy of further investigation within a clinical setting.

Strengths and Limitations

Our study has several strengths including the large sample size and the availability of data on a variety of patient characteristics and clinical comorbidities, which allowed us to adjust for numerous potential confounders. There are several limitations of our study that should be considered when interpreting our findings. First, the design of our study was retrospective and did not allow us to draw any conclusions about a causal link between midazolam and preoperative memories. Second, perioperative memories were assessed from answers to the modified Brice questionnaire, which was not designed to assess such memories. Although our findings are consistent with the limited literature on this subject, it is possible that a targeted assessment tool may have yielded different percentages, but it is unlikely that these percentage differences would have changed the conclusion that a clinically significant proportion of patients have amnesia of the OR. Third, because the modified Brice questionnaire does not include specific questions about patient memories of the anesthesiologist, our study cannot definitively determine whether patients remembered the anesthesiologists who handled their respective surgical cases. Consequently, additional research on patient amnesia of the anesthesiologist would likely be needed to reinforce our preliminary evidence that a substantial portion of patients has amnesia of the OR. Fourth, we excluded 31% of patients from our analyses; this includes a total of >1700 patients whose responses were inconclusive with respect to preoperative location and >600 patients from the University of Manitoba. Although there are statistically significant differences in a number of patient characteristics between the included and excluded cohorts, the differences do not appear to be clinically meaningful. However, given our retrospective study design, there is the inherent possibility that the included and excluded cohorts may have had clinically significant differences in unmeasured confounding variables. Fifth, our study population of patients at high risk for AWR may limit the generalizability of our findings. Sixth, all the participants in the B-Unaware and BAG-RECALL trials had general anesthesia with potent volatile agents, which may limit generalizability to patients who have regional anesthesia, sedation, or total IV anesthesia. Seventh, because we only included findings from sites where midazolam was typically administered in the preoperative holding area, our results may not be generalizable to hospitals where midazolam is not administered at all or is given in the OR before induction of anesthesia. Eighth, our use of midazolam quartiles only reflected dosage, which does not comprehensively incorporate all of the various pharmacokinetic and pharmacodynamic factors that may affect midazolam concentration within the brain.^10–12 Because the midazolam quartiles 1 to 3 represented a narrow range of dosages, this discrepancy may explain the imperfect dose–response relationship between midazolam quartiles and proportion of patients with amnesia of the OR before induction of anesthesia. Lastly, although we did not record the time of midazolam administration, practitioners typically administered midazolam only when the patients were transported from the preoperative holding area to the OR. Thus, it is unlikely that there was a marked heterogeneity in the time between midazolam administration and OR entrance.

CONCLUSIONS

In this large retrospective study among clinical trial participants at hospitals where midazolam was typically administered in the preoperative holding area, the majority of patients did not remember the OR before induction of anesthesia, and a clinically relevant portion of patients also did not remember the preoperative holding area. These findings provide preliminary evidence suggesting that a substantial portion of patients has amnesia of the OR. In terms of policy implications, if additional studies using instruments with specific questions about patient memories of the anesthesiologist yield results suggesting that a substantial proportion of patients have amnesia of the anesthesiologist, then our findings support assessment of patient satisfaction with anesthesia groups on a managerial level rather than at the level of the individual anesthesiologist. In terms of clinical implications, if additional studies demonstrate that a substantial portion of patients do not remember the preoperative holding area, then our findings may also suggest reconsideration of providing important medical information only in the preoperative holding area. Although age, midazolam-induced anterograde amnesia, and emotion-induced retrograde amnesia may be plausible mechanisms driving our findings, additional studies are warranted to better understand perioperative memory retention and the factors associated with amnesia of the perioperative period.

APPENDIX 1

Modified Brice Questionnaire

1. What is the last thing you remember before going to sleep?
• -Being in the preoperative area
• -Seeing the operating room
• -Being with family
• -Hearing voices
• -Feeling a mask on my face
• -Smell of gas
• -Burning or stinging in the IV line
• -Other [free text]: _________________________________
2. What is the first thing you remember after waking up?
• -Hearing voices
• -Feeling the breathing tube
• -Feeling the mask on my face
• -Feeling pain
• -Seeing the operating room
• -Being in the recovery room
• -Being with family
• -Being in the ICU
• -Nothing
• -Other [free text]: ______________________________
3. Do you remember anything between going to sleep and waking up?
• -No
• -Yes—hearing voices
• -Hearing events of the surgery
• -Unable to move or breathe
• -Anxiety/stress
• -Feeling pain
• -Sensation of the breathing tube
• -Feeling surgery without pain
• -Other [free text]: ______________________________
4. Did you dream during your procedure?
• -No
• -Yes
• -What about? [free text]: _______________________
5. Were your dreams disturbing to you?
• -No
• -Yes
6. Did you experience any nausea or vomiting after your operation?
• -No
• -Yes
• If yes, how many times? _________
7. What was the worst thing about your operation?
• -Anxiety
• -Pain
• -Recovery process
• -Functional limitations
• -Awareness
• -Other [free text]: ______________________________

APPENDIX 2

Coding Process

When conducting the structured telephone interview using the modified Brice questionnaire, trained interviewers recorded patient responses within a dropdown menu format, which also contained a free-text section to accommodate miscellaneous patient comments. For assessing patient memories of the operating room (OR) before induction of anesthesia, we have listed the question “What was the last thing you remembered before going to sleep?” along with its corresponding dropdown menu options below:

• What is the last thing you remember before going to sleep?
• -Being in the preoperative area
• -Seeing the OR
• -Being with family
• -Hearing voices
• -Feeling a mask on my face
• -Smell of gas
• -Burning or stinging in the IV line
• -Other [free text]: _________________________________

For each of the prelisted dropdown menu options above, we have listed our locational assignment, along with a brief rationale for our approach, in the table below:

Table: No title available.

Table: No title available.

In addition to the dropdown menu options above, we have listed below other commonly encountered patient responses, our locational assignment, and an explanation of our reasons for this approach:

APPENDIX 3

Results

Amnesia of the Operating Room (1–3 Days Postoperatively)

Among 4692 patients, 52.7% (95% confidence interval [CI], 51.3%–54.2%) of patients had amnesia of the operating room (OR) before induction of anesthesia. With regard to the postoperative period, 94.1% (95% CI, 93.2%–94.8%) of 3411 patients had amnesia of the OR after induction of anesthesia.

University of Manitoba Study Site

At the University of Manitoba (Winnipeg) study site, 23.1% (95% CI, 19.9%–26.7%) of 589 patients had amnesia of the OR before induction of anesthesia at 1 to 3 days postoperatively. As for the postoperative period, 96.0% (95% CI, 93.8%–97.5%) of 455 patients had amnesia of the OR after induction of anesthesia at 1 to 3 days postoperatively.

Differences in Timing of Midazolam Administration Between Study Sites

At Washington University in St. Louis and the University of Chicago, midazolam was typically administered in the preoperative holding area, whereas at the University of Manitoba, midazolam was typically administered in the OR.

RECUSE NOTE

Dr. Franklin Dexter is the Statistical Editor and the Section Editor for Economics, Education, and Policy for Anesthesia & Analgesia. This manuscript was handled by Dr. Gregory Crosby, and Dr. Dexter was not involved in any way with the editorial process or decision.

DISCLOSURES

Name: Yulong Chen, BA.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Yulong Chen has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Alice Cai, BS.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Alice Cai has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Bradley A. Fritz, MD.

Contribution: This author helped analyze the data and write the manuscript.

Attestation: Bradley A. Fritz reviewed the analysis of the data and approved the final manuscript.

Name: Franklin Dexter, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Franklin Dexter reviewed the analysis of the data and approved the final manuscript.

Name: Kane O. Pryor, MD.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Attestation: Kane O. Pryor reviewed the analysis of the data and approved the final manuscript.

Name: Eric Jacobsohn, MB, ChB.

Contribution: This author helped write the manuscript.

Attestation: Eric Jacobsohn reviewed the analysis of the data and approved the final manuscript.

Name: David B. Glick, MD, MBA.

Contribution: This author helped write the manuscript.

Attestation: David B. Glick reviewed the analysis of the data and approved the final manuscript.

Name: Mark D. Willingham, MD.

Contribution: This author helped analyze the data and write the manuscript.

Attestation: Mark D. Willingham has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Krisztina E. Escallier, MD.

Contribution: This author helped write the manuscript.

Attestation: Krisztina E. Escallier reviewed the analysis of the data and approved the final manuscript.

Name: Anke C. Winter, MD, MSc.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Anke C. Winter has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Michael S. Avidan, MBBCh.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Michael S. Avidan has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.