Mortality and Morbidity Associated With Out-of-Hospital Deep Sedation and General Anesthesia for Dental Treatment: A 36-Year Retrospective Study in British Columbia, Canada (1984-2019)

Morbidity and Mortality Event Search

This retrospective study identified mortality and morbidity events through searches of the CCBC and CDSBC databases as well as a gray literature search. Methods to identify mortality and morbidity events during the study period paralleled those of El-Mowafy et al.⁹ Ethics approval was obtained from the University of Toronto (REB Protocol 38915).

Anonymized coroner reports from the CCBC and case summaries from the CDSBC were screened for meeting study inclusion criteria (P.A. and C.Y.). Inclusion criteria included the following: events that occurred in British Columbia, involved out-of-hospital dental treatment under DS/GA by a qualified provider, and included mortality or serious morbidity that occurred within 30 days and that was caused by anesthesia or factors under the control of the anesthesiologist (Table 1). Out-of-hospital DS/GA included treatment provided with procedural team anesthesia and the separate provider model (see online-only Appendix for anesthesia models and team composition). Serious morbidity was assessed using the Severity of Injury Score (SIS) grades, as described by Brunner.¹² This study included morbidity that was classified as either major temporary injuries (SIS 4) or permanent injuries classified as minor/significant/major (SIS 5, 6, or 7) or grave (SIS 8). Events with no obvious, minor, insignificant, or emotional injury (SIS 0, 1, 2, or 3) were excluded from this study.

Table 1.Inclusion Criteria for Mortality and Serious Morbidity Events

View Fullscreen

Unless otherwise specified, in the context of this study, the term morbidity refers specifically to serious morbidity. Events were also evaluated to determine whether mortality or morbidity could be attributed to the provision of anesthesia, as defined by the Australian and New Zealand College of Anaesthetists (ANZCA) report on safety of anaesthesia.¹³ This study included events with mortality or morbidity falling into 3 categories: where it is reasonably certain that death was caused by anesthesia or factors under the anesthesiologist’s control (category 1), where there is some doubt whether death was entirely attributable to anesthesia or factors under the anesthesiologist’s control (category 2), and where it is reasonably certain death was caused by both medical/surgical and anesthesia factors (category 3). Events meeting the criteria of these 3 categories were termed anesthesia related.

The CCBC and the CDSBC were contacted to request database searches for potentially relevant events occurring from January 1, 1984, to December 31, 2019. The CCBC conducted a search of its digitized archive for coroner reports labeled with the tag dental, which is assigned whenever dental treatment is considered a potential contributing factor to death. The CCBC search spanned the years 1987 to 2019.

The CDSBC conducted both physical and electronic database searches for all case files related to “sedation” resulting in mortality or morbidity. This included Critical Incident Reports, the mandatory submissions to the CDSBC detailing an emergency after a patient’s transportation or admission to a hospital. The CDSBC search spanned the years 1984 to 2019.

To further our ability to identify mortality or morbidity cases, additional sources of information were utilized. A gray literature search of digital and newsprint media was conducted to identify mortality and morbidity events occurring between January 1, 1984, and December 31, 2019. Eligibility criteria for newsprint archives and web-based search engines comprised articles written in English, accessible through the link provided or available through the University of Toronto library, published by a provincial/national newsprint or digital media publisher, and including a description of an event related to mortality or morbidity under DS/GA for dental treatment in British Columbia. The search strategy was adapted from a methodology case study.¹⁴ A keyword search was executed across multiple newsprint archives and various web-based search engines and included ProQuest Canadian Newsstream, ProQuest Historical Newspapers: Vancouver Sun, the Vancouver Sun website (vancouversun.com), Google, Yahoo!, and Bing. The search terms, including dental, outcome, and location keywords, were used in all database and web-based searches (Table 2).

Table 2.Gray Literature Search Terms

View Fullscreen

Number of Out-of-Hospital Deep Sedation/General Anesthetic Procedures Administered

To estimate the number of out-of-hospital DS/GA procedures for dental treatment provided by each clinician, a survey questionnaire was administered to DS/GA providers who were dental anesthesiologists (DAs), oral and maxillofacial surgeons (OMSs), and physician anesthesiologists (PAs) registered with the CDSBC to provide DS/GA in dental offices (see online-only Appendix for anesthesia provider questionnaire).^15,16 Methods to estimate the number of DS/GA cases provided during the study time period paralleled those of El-Mowafy et al.⁹

A letter of request for study participation was mailed to each facility and included a survey questionnaire for the facility anesthesia provider that requested the following: the facility and anesthesia provider’s address, contact information for follow-up, the anesthesia provider group type (DA, OMS, or PA), the number of DS/GA procedures provided at the out-of-hospital facility in 2019, the method they used for this calculation, the percentage of treatment provided to pediatric vs adult patients, and the number of years they provided treatment at this facility.

For each provider group, the data provided by survey responders for the year 2019 were used to estimate the number of OHAPs administered by responders and nonresponders (Equations 1 and 2). $$\text{No}\text{. of OHAPs\hspace{0.17em}Performed by Survey Responders}=\left(\text{Mean No}\text{. of Years Practicing by Responders}\right)\times \left(\text{Mean No}\text{.\hspace{0.17em}of\hspace{0.17em}OHAPs by Responders}\right) \times \left(\text{No}\text{. of Survey Responders}\right)$$(1) $$\text{No}\text{. of OHAPs\hspace{0.17em}Performed by Survey Nonresponders}=\left(\text{Mean No}\text{. of Years Practicing by Responders}\right)\times \left(\text{Mean No}\text{.\hspace{0.17em}of\hspace{0.17em}OHAPs by Responders}\right)\times \left(\text{No}\text{. of Survey Nonresponders}\right)$$(2)

Period Prevalence of Serious Morbidity and Mortality

The number of events per total number of anesthetic procedures represents the period prevalence. This is defined as the number of people with a particular characteristic or outcome measured over a specified time interval.¹⁷ The total number of unique anesthesia-related mortality and morbidity events meeting the inclusion criteria was summed and used as the numerator for the calculation of period prevalence. The estimated number of out-of-hospital DS/GA cases for dental procedures served as the denominator (Equation 3). $$\text{Period Prevalence for Mortality and Serious Morbidity} =\frac{\text{Mortality Events}+\text{Serious Morbidity Events}}{\text{Estimated No}\text{. of OHAPs Administered}}$$

For ease of discussion, in this study the term rate is used interchangeably with period prevalence in reference to previous reports of mortality and morbidity in the anesthesia literature. However, it should be understood to reflect period prevalence.

CCBC Search

The CCBC search revealed 4 deaths related to dental treatment; however, all failed to meet inclusion criteria (Figure 1). Of those, 3 events (2010a, 2010b, and 2011a) involved treatment with local anesthesia and did not include any form of sedation. Additionally, one mortality event (1992) resulted from asphyxia 3 days after treatment because of anaphylaxis to a medication prescribed for the postoperative period. DS/GA was not mentioned in the coroner’s report, so there was no indication whether this patient received DS/GA.

View Figure

• Download as Powerpoint
• Download Figure

CDSBC Search

The CDSBC search revealed 6 mortality events and 4 morbidity events (Figure 2). Among the 6 mortality events, 3 met inclusion criteria and were labeled as ANZCA category 3 events, as both surgery- and anesthesia-related factors potentially contributed to death in all 3 events (1986b, 1987b, and 1987c). The remaining 3 mortality events did not meet inclusion criteria. One resulted from natural causes (2011b) and 2 were associated with in-hospital GA (1984b and 1984c).

View Figure

• Download as Powerpoint
• Download Figure

All 4 identified morbidity events did not meet inclusion criteria. One involved treatment with local anesthesia (1986a), 1 involved a moderate sedation provider unintentionally providing DS (2019), 1 involved a moderate sedation provider intentionally providing DS that was beyond their registered scope of practice in a facility that was not registered to provide DS/GA treatment (2012), and 1 involved out-of-hospital DS/GA resulting in surgery-related minor/nonpermanent morbidity (subcutaneous emphysema) from which the patient fully recovered (1987a).

Gray Literature Searches

Web-based searches required multiple executions of the search string, resulting in internal duplicate results that were removed prior to combining the results from different search streams (Figure 3). The 6 search strategies identified 82 potentially relevant articles, of which 32 did not meet inclusion criteria, resulting in 50 articles reporting on 7 unique events (5 mortality and 2 morbidity) involving treatment with DS/GA.

View Figure

• Download as Powerpoint
• Download Figure

Of the 5 mortality events, 3 (1986b, 1987b, and 1987c) met inclusion criteria, and both surgery- and anesthesia-related factors were reasonably certain to have contributed to these deaths. The other 2 events (1984b and 1984c) did not meet inclusion criteria as they involved in-hospital GA for dental treatment. The 2 morbidity events did not meet inclusion criteria as 1 involved surgery-related complications (1987a) and 1 involved a moderate sedation provider intentionally providing DS (2012).

Summary of Events Meeting Inclusion Criteria From All Searches

All searches collectively identified 3 mortality events that met inclusion criteria (1986b, 1987b, 1987c). Coroner’s reports were obtained for all mortality events, and each mortality event meeting inclusion criteria was identified by more than one search method (Table 3). The searches did not identify any serious morbidity events that met inclusion criteria.

Table 3.Unique Events of Morbidity and Mortality 1984-2019

View Fullscreen

Number of Out-of-Hospital Deep Sedation/General Anesthetic Procedures Administered

The provider survey achieved a response rate of 32.4% from DS/GA providers (Table 4). Among the respondents, DA providers had the highest response rate, and OMS providers had the lowest. The survey responders provided a combined total of 18,281 anesthetic procedures in 2019 (Table 5). Deep sedations accounted for 76% of all OHAPs for dental treatment.

Table 4.Survey Responders Compared With Registered Providers by Provider Group

View Fullscreen

Table 5.Total Number of Out-of-Hospital Anesthetic Procedures

View Fullscreen

The calculated estimate of OHAPs in 2019 based on actual reported cases by responders and estimated number of cases by nonresponders was 59,737. The mean number of anesthetic procedures and mean years in practice by survey responders in each provider group were extrapolated to include survey nonresponders. This yielded an estimate of 1,019,853 OHAPs provided for dental treatment over the period studied for those still in practice in 2019 (Table 5).

Period Prevalence

A total of 3 mortality events meeting inclusion criteria for this study were identified, and no morbidity events meeting inclusion criteria were found. The estimated number of anesthetic procedures provided by DS/GA providers was 1,019,853. This study determined a period prevalence of 2.94 mortality events per 1 million OHAPs for dental treatment in British Columbia from 1984 to 2019.

Mortality and Morbidity Events

There were 3 mortality events (1986b, 1987b, and 1987c) that met inclusion criteria for this study, and all 3 involved the same OMS provider but had 3 different PAs providing the DS/GA. In all 3 cases of death, venous air embolism led to cardiovascular collapse due to the use of a dental drill not designated for implant surgery by the manufacturer. Specifically, the surgeon used a drill with an air inlet that allowed air to enter the internally irrigated implant osteotomy bur, injecting pressurized air and water into the mandible. The correct type of drill for internally irrigated osteotomy burs—standard for implant placement at the time of these events—has only a water inlet port with no air inlet.

Venous air embolism occurred because the implant osteotomy site directly communicates with mandibular venous sinuses and provides entry for air into the venous system. The severity of air embolism and its effects depend on the volume of gas, accumulation rate, and patient position. Although small volumes of air can be asymptomatic and absorbed without sequelae other than occasional coughing, larger volumes can cause increased pulmonary artery pressures due to obstruction of the pulmonary vessels by air, leading to elevated resistance to right ventricular outflow and causing right ventricular strain.¹⁸ The adult lethal dose is estimated at 200 mL to 300 mL (3-5 mg/kg), with increasing proximity of air injection to the right heart requiring progressively less gas volume for fatal consequences. Air infusion rates exceeding 1.5 mL/kg/min are associated with bradycardia and cardiovascular decompensation, corresponding to approximately 100 mL/min for a 70-kg adult.¹⁹ Investigation by the coroner revealed that the dental drill and air supply lines used by the surgeon could produce air flow rates of 1 L/min to 2 L/min at pressures between 10 cm and 80 cm of water. Although the air flow rate of the air supply lines would have been reduced by the small orifices of the implant osteotomy bur, the extent of this reduction was not tested. The surgical equipment used, being capable of injecting air into the mandible, makes it reasonably certain that surgical error was the direct causal factor for death in these 3 cases.

All 3 patients who died arrested suddenly at the same stage of surgery, just after beginning the second implant osteotomy. Symptoms of air embolism are nonspecific and include seizure, loss of consciousness, confusion, altered mental status, cyanosis, hypoxia, hypercapnia, hypotension, tachypnea, wheezing, bronchospasm, tachycardia, and bradycardia.²⁰ Clinical signs and symptoms depend on the location of the air embolus: for instance, cerebral air embolism leads to neurological signs such as weakness and seizures.²⁰ Large volumes can fill the entire right atrium and ventricle, leading to arrhythmias, decreased cardiac output, myocardial ischemia, and cardiac arrest, which is associated with higher mortality rates.¹⁸ Signs of air emboli in the right ventricle or pulmonary arteries, as suspected in these 3 mortality events, include a sudden decrease in end-tidal carbon dioxide, followed by hypoxia, hypotension, and cardiovascular collapse.²⁰ The 3 patients were challenging to resuscitate with professional cardiopulmonary resuscitation (CPR). Two showed signs of right ventricular strain and myocardial ischemia on electrocardiogram (ECG), and all 3 had pulmonary edema, all potential complications of air embolism. The 1986b and 1987b patients also exhibited upper body cyanosis, pathognomonic for superior vena cava obstruction, as would occur with massive air embolism originating above the level of the heart.

This cluster of events began in 1986 with a coroner investigation into the 1986b incident. The patient was resuscitated, but cardiac output was low and could not be improved. This patient died after several days. Seven months later, in the 1987b event, a 16-year-old boy underwent the same procedure by the same OMS, leading to another fatality. As the coroner probed into this and the preceding fatality, awareness increased regarding the heightened rate of serious complications linked to implant surgery by this surgeon. Air embolism emerged as a suspected factor and was added to the list of potential causes of death. The coroner noted that the existing literature reports on dental implant surgery were devoid of air embolism as a complication. Investigation into the cause of death in these 2 instances was underway for 3 months when the third mortality event (1987c) occurred. By the time of the third mortality event in 1987, the surgeon had performed a total of 11 implant surgeries. Of those 11 patients, 5 had serious complications during treatment. Four received DS/GA during treatment, with 3 experiencing cardiovascular collapse resulting in death and 1 suffering massive subcutaneous emphysema but recovering without sequelae. The one patient treated with local anesthesia lost consciousness and had an undetectable pulse but continued breathing and spontaneously regained consciousness after approximately 40 minutes while paramedics were attempting to start an intravenous line.

Although the anesthesiologist did not cause the air emboli, their monitoring of patients was found to be below the standard of care at the time. When these mortality events occurred, the anesthetic monitoring standard included intermittent noninvasive blood pressure (NIBP) and continuous ECG monitoring, both of which were omitted during these anesthetic procedures. Had these monitors been utilized, the anesthesiologist might have detected and treated the air emboli before the sudden cardiovascular collapse. Due to the below-standard monitoring practices during these events, the conduct of the anesthesia can be considered an indirect causal factor in these deaths, aligning with ANZCA category 3.

Although pulse oximetry and capnography monitors were recommended at this time, they were not required monitors and had not yet been widely implemented for in-hospital anesthesia in Canada.²¹ In 1987, the Canadian Anaesthetists’ Society published its second edition of guidelines for the practice of anesthesia in Canada.²¹ It stated that appropriate monitoring of patients undergoing anesthesia include monitoring inspired oxygen tension, ECG, and ventilator pressure and regular observation of pulse, NIBP, peripheral circulation, and respiratory adequacy. Unfortunately, in the mortality events discussed, intermittent pulse checks and ensuring respiratory adequacy appear to have been the extent of monitoring. The initial signs for the 3 mortality events involving air emboli identified by this study included seizure-like activity (1986b), lip pallor quickly turning to cyanosis (1987b), and coughing/jerky movements (1987c). Because pulse oximetry, capnography, NIBP, and ECG monitoring were not performed, it is unknown whether the use of these monitors could have facilitated earlier identification, treatment, and prevention of mortality.

The 3 mortality events were reported in the Canadian Journal of Anesthesia to alert anesthesia providers of the potential for air embolism during implant surgery.²² Although the current study identified surgery and anesthesia factors contributing to death according to the ANZCA (2021) classification system, there are differing opinions regarding the anesthesiologists’ role in these deaths. The coroner investigation revealed gross surgical error in the use of incorrect surgical equipment as the causative factor of the mortality events. It is uncertain if standard anesthetic monitors at that time (ECG, intermittent NIBP, intermittent pulse checks) or the current standard monitors (addition of pulse oximetry and capnography) could have prevented the deaths had they been present and utilized properly, as large air emboli causing cardiac arrest are known to have poor outcomes.¹⁸ Although appropriate monitoring might have provided some warning, it is unclear what contribution this might have made to rescue given the sudden and catastrophic nature of the events and complexity of rescue from air embolism. In the 1987c event, despite a sedation goal of DS, the depth of anesthesia became light enough for the patient to turn their head at the surgeon’s request and vocally complain of pain from drilling just prior to losing consciousness, becoming profoundly cyanotic, and arresting. Conversely, one of the implant patients treated with local anesthesia and no sedation experienced a sudden loss of consciousness and had an undetectable pulse for a prolonged period, yet they recovered without intervention or sequelae. Although these mortality events had a direct surgical cause, it cannot be ruled out that anesthesia may have indirectly affected survivability due to the omission of standard anesthesia monitors, which could have provided early warning of physiologic compromise. Therefore, these were considered category 3 events in which both surgery and anesthesia are considered contributing factors in death.

Although they did not meet inclusion criteria for this study, 2 pediatric anesthesia-related in-hospital events identified (1984b, 1984c) are worthy of discussion. They occurred just over 2 weeks apart. This temporal proximity, coupled with both events happening under the care of the same PA, led to the suspension of surgical services under GA in the hospital until an investigation was completed. There were no suspected complicating factors resulting from the surgical procedure, as 1 of the 2 children died before receiving any dental treatment. In both events, the PA’s choice of anesthetic medications and doses, lack of appropriate monitoring of vital signs, and substandard resuscitation efforts were all deemed contributing factors to the deaths. The results of this investigation prompted sweeping changes to this rural hospital’s anesthesia staff, emergency preparedness, CPR/Advanced cardiovascular life support/Pediatric advanced life support training requirements, and anesthesia recovery nursing protocols before the resumption of GA for surgical procedures in the hospital.

Period Prevalence of Serious Morbidity and Mortality

This study found a mortality rate of 1 event per 339,951 out-of-hospital DS/GA provided for dental treatment. The ANZCA has been studying public and private hospital anesthesia-related mortality in triennial Safety of Anaesthesia reports since 1985.¹³ The anesthesia-related mortality rate varies with the American Society of Anesthesiologists (ASA) physical status classification of the patient, with higher ASA classifications having higher mortality rates. The 2021 ANZCA report indicated that the overall mortality rate for patients classified as ASA I to IV was 1 death per 59,125 anesthetic procedures. In contrast, the subset of patients classified as ASA I to III had a significantly lower mortality rate, with 1 death per 131,278 anesthetic procedures. Patients receiving dental treatment at out-of-hospital facilities typically fall within the ASA I to III categories, which generally have a lower risk of anesthesia-related mortality. The physiologic stress from the type of dental treatment performed in out-of-hospital settings is also lower than that with more invasive in-hospital surgical procedures and should further reduce the risk of anesthesia-related mortality for patients deemed suitable for out-of-hospital dental treatment. Thus, the findings of this study appear consistent with those of the latest ANZCA report.

Using closed claims data from 2000 to 2013 from the Oral and Maxillofacial Surgery National Insurance Company (OMSNIC) database, the equivalent of 2.87 events involving brain damage or death occurred per million out-of-hospital dental DS/GA was reported for the United States.²³ Although the OMSNIC database exclusively comprises event data for OMS providers, its findings align with the 2.94 events per million anesthetic procedures reported in this study.

The 2 previous studies in Canada included the same practice models as the current study: the procedural team anesthesia model and the separate provider model. The study by El-Mowafy et al⁹ included DA, OMS, and PA providers, whereas the study by Nkansah et al⁸ included only DA and OMS providers, excluding PA providers. The OHAP-related mortality and serious morbidity combined rate reported by El-Mowafy et al⁹ was 1.05 events per 1 million cases in Ontario from 1996 to 2015, and the mortality rate reported by Nkansah et al⁸ was 1.4 events per million anesthetic procedures from 1973 to 1995. The mortality rate in this study, 2.94 events per million anesthetic procedures, is 2 to 3 times higher than that in these previous Canadian studies. The primary distinction between the previous studies and the current study is that the 3 mortality events identified in this study all involved one OMS repeating the same surgical error.

It is worthwhile to note that if these 3 surgically precipitated events had been excluded from our study, the number of out-of-hospital dental DS/GA-related mortality events in British Columbia over the 36-year study period would be zero. This study spans a time during which all aspects of anesthesia improved, including a culture of patient safety that has become central to all aspects of anesthetic care as well as the availability of general anesthetic agents with improved safety profiles. Additional patient monitors, such as pulse oximetry and capnography, became standard after 1987 and significantly reduced major intraoperative incidents approximately 5-fold after their use became routine.²⁴ The mortality events identified in this study occurred before these monitors became standard in Canadian anesthetic practice.²¹

Patients seeking dental treatment under DS/GA often face barriers like fear or phobia and/or behavioral or developmental conditions. For these patients or their guardians, objective risk values can help them understand treatment risks and form an important component of the informed consent process. Communicating rare event risks is challenging, as most people have a poor understanding of probability.²⁵ Comparing risks with real-world equivalents can aid comprehension. For instance, the mortality and morbidity rate of dental treatment under DS/GA (1:339,951) can be compared with the annual risk in Canada of serious injury due to a motor vehicle collision (1:3500), of drowning (1:100,000), and of being injured or dying due to lightning (1:200,000).^26-28 Driving 600 kilometers in an automobile, the average number of kilometers driven every 2 weeks by Canadian drivers, has a 1:333,000 odds of death. Thus, the risk of death or serious injury during dental treatment under DS/GA found in this study is equivalent to or lower than everyday risks that people typically do not avoid due to perceived risk of death.

Although this and the quoted studies provide an estimate of the risk of death or injury, the actual risk for each patient is affected by a multitude of variables, including the patient’s ASA classification, comorbidities, age, and surgical procedure.²⁵ Risk should, therefore, be individualized based on these factors, and it would be erroneous to conclude that dental treatment under DS/GA is low risk for all patients.

Study Strengths and Limitations

Mortality and morbidity events were retrospectively obtained from the CDSBC and CCBC, a common approach for studies of low-probability anesthetic events. The CCBC investigates all child deaths and all unnatural or sudden deaths in British Columbia and would be expected to capture all mortality events. Since 2008, the CDSBC has mandated reporting of critical sedation events, and thus was prospectively collecting relevant events for approximately one-third of the study period.^29-31 Prior to this, events were reported to the CDSBC voluntarily by clinicians or via patient complaints.

This study did not pursue self-reported event data from clinicians, as these data are subject to recall and might have reduced survey response rate given the sensitive nature of disclosing morbidity and mortality in out-of-hospital settings. However, missed events in the database searches could lead to underestimating reported rates. Mortality and morbidity event data were collected from as many sources as possible to increase the likelihood of identifying all relevant events during the study period. The CCBC search spanned from 1987 to 2019 but failed to identify the 2 mortality events occurring in 1987 that were associated with dental treatment under DS/GA. Upon personal communication with the Office of the CCBC, it was revealed that although the coroner reports from 1987 to present were digitized, not all reports from 1987 had undergone digitization. Coroner reports for events involving the 2 mortality events (1986b and 1987b) identified by the other search streams were requested and obtained from the CCBC. All events in which dental treatment was a contributor to death were found to be correctly tagged in the respective coroner reports and aligned with the communication from the CCBC regarding the two 1987 events not being initially identified in the database search due to lack of digitization. Labeling the cause of death or contributing factors is essential for relevant CCBC events to be identified by a keyword search. There is no indication that this type of error occurred for any of the mortality events identified in this study.

The CDSBC performed both electronic and manual searches for events from 1984 to 2019. Although it is possible that an event or events were missed during the search process or that the CDSBC was unaware of certain events, the likelihood of this has been reduced since 2008. In that year, the CDSBC implemented mandatory reporting of critical events.^29,30 The gray literature search identified all but one of the known events and is recommended for event identification in future studies of this nature. The overlap in event identification across sources suggests that robust multisource searches provide appropriate rigor for identification of events; this is a strength of this study. The number of OHAPs for dental treatment is not systematically collected in British Columbia, resulting in the absence of any official records for these data. This study aimed to fill this gap by requesting voluntary submission of DS/GA case counts by clinicians for 2019 via a paper-based survey. The response rate was 32%, similar to that of a 2015 survey targeting DS/GA providers in Ontario, which achieved a response rate of 39%.⁹ A higher response rate would have improved the accuracy of the estimate of the number of out-of-hospital dental DS/GA provided.

Higher rates of survey responses may be difficult to achieve, as the study’s topic is not universally welcomed by DS/GA providers. The discussion of mortality and morbidity as well as sedation-practice models can evoke strong emotions of either support or disapproval in clinicians and may have influenced participation. To avoid overestimating the number of DS/GA cases, this study estimated the number of anesthetic procedures provided by all DS/GA providers based on the mean years in practice of those who responded and did not attempt to extrapolate the number performed over the entire study period. However, this does not rule out underestimation by not surveying clinicians who had practiced during this period but had retired prior to 2019. Other factors affecting the accuracy of the estimate include variability in the annual number of anesthetic procedures performed over a provider’s career and variability in the number of clinicians per year.

Survey response bias is unlikely to be a major contributor of error, as this study’s survey questions sought descriptive data, such as the number of years the clinician had been providing OHAPs for dental treatment and the number of anesthetic procedures performed in 2019. Nonresponse bias typically affects voluntary surveys. The survey was sent multiple times to remind providers of the importance of their participation and to encourage them to complete the survey.

To provide reference for the estimated number of out-of-hospital DS/GAs administered, a comparison was made with similar studies conducted in Canada. The comparison considered the respective provincial populations at the time each study took place. This study estimated the number of anesthetic procedures provided by all registered DS/GA providers in British Columbia in 2019 to be 59,737. The percentage of the population receiving an OHAP for dental treatment can be calculated by dividing the estimated number of anesthetic procedures in 2019 by the total population of British Columbia in 2019, yielding 1.2%. Using data from the El-Mowafy et al⁹ and Nkansah et al⁸ studies, the percentages of the population in Ontario receiving an OHAP for dental care in 2015 and 1990 were approximately 1.4% and 1.3%, respectively. The percentage of the population receiving an OHAP for dental treatment, as estimated from clinician surveys in these 3 Canadian studies, is consistent and comparable.^29,30