Delays in presentation, severity of illness, and adverse surgical outcomes among patients transferred from Nunavik for acute care surgery

Abstract

Background:

Providing acute care surgery (ACS) to Nunavik, Quebec, is highly challenging because of health care infrastructure and access issues, with geographic distances and delays in transport translating to suboptimal care, particularly for acute surgical conditions requiring timely intervention. The main objective of this study was to identify clinical predictors of poor surgical outcomes in patients transferred from Nunavik to a tertiary care centre for operative ACS conditions.

Methods:

We conducted a retrospective cohort study at the Montreal General Hospital (MGH), the main tertiary care centre providing ACS care to Nunavik. We included all adult patients who were transferred from Nunavik to the MGH and underwent a surgical intervention for an ACS condition between 2015 and 2020. The outcome of interest was the occurrence of a postoperative complication, as defined by the Clavien–Dindo classification.

Results:

In total, 177 patients were transferred from Nunavik to the MGH and underwent a surgical intervention for an ACS pathology during the study period. The mean age was 40.0 (standard deviation 16.7) years, and half the patients (n = 90, 50.8%) were female. The most common ACS pathology was acute cholecystitis (n = 65, 36.7%), followed by acute appendicitis (n = 51, 28.8%) and biliary pancreatitis (n = 13, 7.3%). Forty-six (26.0%) patients had a postoperative complication of Clavien–Dindo grade 1 or higher. On univariate analysis comparing patients who experienced complications to those who did not, we observed significant differences in age, sex, Emergency Surgical Score (ESS), delays in operative intervention, proportion of open surgical interventions, proportion of intraoperative perforations, admissions to the intensive care unit, and hospital length of stay. On multivariable logistic regression, male sex (odds ratio [OR] 2.75, 95% confidence interval [CI] 1.22 to 6.17), higher ESS (OR 1.28, 95% CI 1.03 to 1.59), higher American Society of Anesthesiologists (ASA) score (OR 2.28, 95% CI 1.23 to 4.24), and longer delays in presentation (OR 1.07, 95% CI 1.01 to 1.14) were predictive of postoperative complications.

Conclusion:

Delays in presentation and severity of illness on arrival are associated with worse postoperative outcomes for patients with ACS conditions transferred from Nunavik to an urban tertiary care centre. Identifying these predictors of poor postoperative course will help shape future quality improvement initiatives, which should focus on point-of-care resuscitation, improved transport logistics and resources, and ultimately improved access to on-site surgical care.

Emergency general surgery constitutes a tremendous burden of disease, representing 7% of all hospital admissions in the United States.^1,2 In Canada, the acute care surgery (ACS) model is used to care for patients presenting with nontraumatic general surgical conditions requiring emergency treatment.³ Patients with ACS conditions requiring emergency operative procedures are at a higher risk of postoperative complications and death than patients undergoing elective procedures.^4–6 Given that acute surgical conditions are associated with unfavourable outcomes when resuscitation and interventions are not initiated in a timely manner, such conditions require prompt management.^7,8

Delivery of timely, quality medical care in rural and remote settings is a challenge globally. The World Health Organization estimates that more than half of the world’s rural population have inadequate access to essential health services and further characterizes rural populations as having lower incomes and worse health outcomes.⁹ In Canada, access to surgical care is not equitable, with remote Indigenous communities having lower utilization rates for both elective and life-saving surgical procedures.¹⁰ Similar issues have been noted in Australia, where the rural population has a higher disease and injury burden and lower life expectancy, with especially increased disparities noted in Aboriginal and Torres Strait Islander populations. With fewer surgeons available than in urban areas, this results in inequity in access and timeliness of care, as well as poorer health outcomes for people living in rural Australia.^11–13

Nunavik is the Inuit homeland in Quebec, Canada, and the majority of the region’s residents are Inuit.^16,17 As with other Indigenous Peoples in Canada, Inuit populations experience health disparities compared with the non-Indigenous population, given the detrimental impacts of racist and colonialist policies on the social determinants of health in these communities.^18–20 Indigenous patients also have inequitable access to care compared with the rest of the population and face well-documented discrimination in the health care setting.^20–22

The vast geographic distance between Nunavik and Montréal presents substantial logistic challenges to the delivery of ACS services to Nunavik’s population (Figure 1 and Figure 2).²³ Despite availability of supplies, equipment, and infrastructure, surgical capacity is limited in Nunavik, mainly owing to the lack of personnel to perform procedures.²⁴ Emergency surgical services for the region are provided in Montréal, more than 1400 km from Nunavik’s hospital facilities, with transport delays arising because of limited aircraft availability and unpredictable weather conditions.²³ Moon and colleagues identified transport time as a predictor of clinical deterioration during transport among patients from Nunavik with surgical trauma.²⁵

Fig. 1.

Map of Nunavik, Quebec, Canada. Adapted from Natural Resources Canada Reference Maps (available at https://natural-resources.canada.ca/maps-tools-publications/maps/atlas-canada/reference-maps#qc). Contains information licensed under the Open Government Licence, Canada. Please review the full licence terms, including attribution requirements and restrictions, at: https://open.canada.ca/en/open-government-licence-canada).

Fig. 2.

Example of patient trajectory from Nunavik to Montréal with approximate distances travelled. Adapted from Natural Resources Canada Reference Maps (https://natural-resources.canada.ca/maps-tools-publications/maps/atlas-canada/reference-maps#qc). Contains information licensed under the Open Government Licence, Canada. Please review the full licence terms, including attribution requirements and restrictions, at: https://open.canada.ca/en/open-government-licence-canada).

Predictors of complications for patients with ACS conditions from remote Canadian regions have not been well characterized. As 18% of the country’s population lives in rural or remote areas across the vast Canadian landscape, this is an important research question.²⁶ The objective of this study was to elucidate predictors of postoperative morbidity in patients with ACS conditions from Nunavik.

Methods

Study design and setting

We conducted a single-centre retrospective cohort study at the Montreal General Hospital (MGH), which serves as the main tertiary care centre for patients with ACS conditions from the Nunavik region. The population of Nunavik was 14 050 people as of 2021, of whom 90% are Inuit.¹⁷ There are 14 villages in Nunavik, which range from a population of 233 to 2668 people.²⁷ Each village has a medical clinic staffed by nurses. Villages such as Inukjuak, Salluit, and Kuujjuarapik are also staffed with full-time family physicians, while other villages receive monthly visits from family physicans.²⁴ Nunavik has 2 regional hospitals, Ungava Tulattavik Health Center in Kuujjuaq and Inuulitsivik Health Centre in Puvirnituq, which offer secondary care services. Patients from the other villages access these hospitals through air travel. Although these medical facilities have health care professionals available, resources are very limited. During the study interval (2015 to 2020), there were no functioning operating facilities available for emergency surgical care in Nunavik. Furthermore, computed tomography (CT) imaging was not available in Nunavik during the study period.^23,24

Patients who require a transfer from Nunavik to the MGH must first be transported to 1 of the 2 hospitals in Kuujjuaq or Puvirnituq. Patient transfer is arranged following consultation and approval by an on-call acute care surgeon at the MGH. The patient is then transported to the MGH via fixed-wing aircraft.

Patient population

All adult patients (aged ≥ 18 yr) transferred from Nunavik to the MGH for surgical intervention related to an ACS condition during the period from 2015 to 2020 were eligible for inclusion. We categorized ACS conditions according to the American Association for the Surgery of Trauma’s published list of emergency general surgery conditions.²⁸ They included a large range of procedures for the treatment of intra-abdominal infections, bleeding, intestinal compromise, intestinal obstruction, hernia-related conditions, and peptic ulcer disease.

Variables and outcomes

We collected data from electronic medical records using assessment and consultation notes by the emergency medicine team and the ACS team upon transfer to the MGH emergency department (ED). We collected patient- and disease-related characteristics, including sociodemographic data and the patient’s initial location of transfer. Severity of illness was assessed using the American Society of Anesthesiologists (ASA) score and the vital signs upon presentation at the MGH ED. We calculated the Emergency Surgery Score (ESS) by summing scores assigned to 22 variables, including demographic data, comorbidities, and preoperative laboratory values, for a maximum score of 29. A full list of variables included in the ESS is provided in Appendix 1, available at www.canjsurg.ca/lookup/doi/10.1503/cjs.005425/tab-related-content).²⁹ Given that chronic comorbidity status was inconsistently documented in our data source, we used ASA score and ESS as proxies for patient comorbidities and physiologic reserve. In the multivariable logistic regression analysis, we included ASA score as a continuous ordinal variable (1 to 5), and ESS as a continuous variable (0 to 29) without any cut-off.

We also collected perioperative characteristics, including recurrent presentation, ACS condition, operative diagnosis, surgical approach, presence of intestinal perforation, and the need for blood product transfusion. Two distinct time intervals were measured, as described in the patient’s chart: the time from the onset of symptoms until arrival at the MGH (time to presentation), and the time from symptom onset to operative intervention (time to operating room [OR]). For the final multivariable logistic regression analysis, we included the time to presentation as a variable of interest, given that our main interest was in delays occurring before assessment in MGH ED that were related to the barriers in reaching timely care (e.g., acceptance and arrangement of transfer, transportation, restrictions related to weather, available personnel) rather than barriers in accessing the OR at the receiving institution. Given the wide confidence interval (CI) and right skew observed in the time to presentation, we performed a sensitivity analysis where the top 5% of outliers in time to presentation (≥ 14 d) were excluded on the final multiple logistic regression analysis.

We documented postoperative hospital trajectory, including necessity of intensive care unit (ICU) admission, mechanical ventilation, and length of stay (LOS). A proportion of patients from Nunavik occasionally use the Northern Quebec Module, a Montréal-based lodging facility that allows patients to remain in the city following their hospital discharge and helps mitigate the necessity for repeated travel for outpatient clinic visits early in the postoperative period. To capture this extended LOS in the city, we measured both hospital LOS and total Montréal LOS.

The outcome of interest was the occurrence of postoperative complications, as defined by the Clavien–Dindo classification.³⁰

Statistical analysis

We presented descriptive variables as means with standard deviations (SDs) for continuous variables and frequency percentage for categorical variables. Statistical analyses included Student t tests or Kruskal–Wallis tests for continuous variables and univariate analysis or the Fisher exact test for categorical variables. We conducted a multivariable logistic regression analysis, controlling for clinically relevant factors that were determined a priori, to identify predictors of adverse postoperative outcomes. Data analysis was performed using IBM SPSS version 27.0, and statistical significance was defined as p < 0.05.

Nunavik community and institutional involvement

This study was approved by the Director of Professional Services of Nunavik (N.B.). Community input and approval was obtained from Inuit leadership (L.W.) before submission for publication. The findings were also discussed with local health care workers (A.V.), and feedback was incorporated in the interpretation of the data.

Ethics approval

This study was approved by the research ethics board of the McGill University Health Centre (no. 2021-7555).

Results

A total of 177 patients were transferred to the MGH and subsequently underwent a surgical intervention for an ACS condition during the study period. Among the cohort, 69 patients (39.0%) were from Kuujjuaq or Puvirnituq, while the remaining 108 patients (61.0%) were from surrounding villages. The mean age was 40.0 (SD 16.7) years, and approximately half the patients were female (n = 90, 50.8%). The predominant ACS condition observed was acute cholecystitis (n = 65, 36.7%), followed by acute appendicitis (n = 51, 28.8%), and biliary pancreatitis (n = 13, 7.3%). Of the 172 patients discharged from the MGH, most were transferred to the Northern Quebec Module (n = 106, 61.6%), while the remainder were discharged directly home (n = 66, 38.4%). Of the remaining 5 patients in the cohort, 2 patients were transferred to another acute care hospital, 1 patient was transferred to a chronic care hospital, and 2 patients died in hospital. The average LOS in hospital was 8.4 (SD 15.6) days, while the average LOS in Montréal was 15.5 (SD 17.0) days.

Of the total cohort of 177 patients, 46 (26.0%) had postoperative complications of Clavien–Dindo grade 1 or higher. Of these 46 patients, 25 (54.3%) had Clavien–Dindo grade 1 or 2 complications, 19 (41.3%) had grade 3 or 4 complications, and 2 (4.3%) had grade 5 complications. The most common type of postoperative complication was infection (n = 18, 39.1%), followed by bleeding (n = 6, 13.0%).

A comparison of demographic and clinical outcome data for patients with or without complications is shown in Table 1. Patients with complications were significantly older (53.3 [SD 17.9] v. 37.9 [SD 14.2] yr, p < 0.001) and had a higher proportion of males (65.2% v. 43.5%, p = 0.01) than those without complications. Additionally, patients with complications had a higher ESS (4.5 [SD 3.6] v. 2.3 [SD 1.6], p < 0.001), a longer time to presentation (6.2 [SD 8.5] v. 3.5 [SD 3.9] d, p = 0.005), and a longer time to OR (9.1 [SD 10.2] v. 5.8 [SD 5.0] d, p = 0.004). The distribution of time to presentation, stratified by outcome, is illustrated in Figure 3. Open surgical cases were more prevalent among patients with complications (58.7% v. 11.5%, p < 0.001), in contrast to patients without complications, among whom laparoscopic procedures were predominant. Moreover, intraoperative perforations, ICU admissions, and extended hospital and Montréal stays were more frequent among patients with complications (Table 1).

Table 1.

Demographics and outcomes for patients with or without postoperative complications^*

Variable	No (%)*of patients		p value
	With complications n = 46	Without complications n = 131
Age, yr, mean ± SD	53.3 ± 17.9	37.9 ± 14.2	< 0.001
Sex, female	16 (34.8)	74 (56.5)	0.01
ESS, mean ± SD	4.5 ± 3.6	2.3 ± 1.6	< 0.001
ACS diagnosis			< 0.001
Acute cholecystitis	11 (23.9)	54 (41.2)
Acute appendicitis	7 (15.2)	44 (33.6)
Acute diverticulitis	9 (19.6)	5 (3.8)
Acute pancreatitis	2 (4.3)	11 (8.4)
Hernia or bowel obstruction	12 (26.1)	10 (7.6)
Other	5 (10.9)	7 (5.3)
Hemodynamic instability upon MGH arrival	9 (19.6)	5 (3.8)	< 0.001
Time to presentation, d, mean ± SD	6.2 ± 8.5	3.5 ± 3.9	0.005
Time to OR, d, mean ± SD	9.1 ± 10.2	5.8 ± 5.0	0.004
ASA score, mean ± SD	2.7 ± 0.8	2.0 ± 0.7	0.6
Open procedure	27 (58.7)	15 (11.5)	< 0.001
Recurrent disease	17 (37.0)	23 (17.6)	< 0.001
Perforation	27 (58.7)	39 (29.8)	< 0.001
ICU admission	11 (23.9)	1 (0.8)	< 0.001

ACS = acute care surgery; ASA = American Society of Anesthesiologists; ESS = Emergency Surgery Score; ICU = intensive care unit; MGH = Montreal General Hospital; OR = operating room; SD = standard deviation.

^*Unless indicated otherwise.

^†Clavien–Dindo grade 1 or higher.

Fig. 3.

Box and whisker plot of time to presentation, classified by outcome.

In multivariable logistic regression, predictors of postoperative complications included male sex (odds ratio [OR] 2.75, 95% CI 1.22 to 6.17), higher ESS (OR 1.28, 95% CI 1.03 to 1.59), higher ASA score (OR 2.28, 95% CI 1.23 to 4.24), and longer time to presentation (OR 1.07, 95% CI 1.01 to 1.14) (Table 2). A sensitivity analysis was performed, which resulted in loss of significance in the association between time to presentation and complications. The other variables (male sex, higher ESS, higher ASA score) remained significant in this sensitivity analysis.

Table 2.

Predictors of postoperative complications^*

Predictors	OR (95% CI)
Main analysis
Male sex	2.75 (1.22–6.17)
ESS	1.28 (1.03–1.59)
ASA	2.28 (1.23–4.24)
Time to presentation	1.07 (1.01–1.14)
Sensitivity analysis
Male sex	1.19 (1.05–1.34)
ESS	1.04 (1.01–1.07)
ASA	1.15 (1.05–1.25)
Time to presentation	1.00 (0.98–1.02)

ASA = American Society of Anesthesiologists; CI = confidence interval; ESS = Emergency Surgery Score; OR = odds ratio.

^*Clavien–Dindo grade 1 or higher.

Discussion

More than a quarter of the patients with ACS conditions who were transferred from the remote region of Nunavik to an urban tertiary care centre had a postoperative complication. Male sex, higher ESS, higher ASA, and longer time to presentation from symptom onset were predictors of postoperative morbidity.

Substantial advancements in surgical and anesthesia techniques and perioperative recovery pathways have led to improved patient outcomes overall in the past few decades.^31–33 However, postoperative complications remain a pressing issue.^34–36 Furthermore, surgical advancements have not uniformly benefitted all patient populations, with Indigenous patients continuing to experience poorer postoperative outcomes than their non-Indigenous counterparts.^10,37,38

In fact, the complication and mortality rates for our cohort of patients were comparable to the literature on patients with ACS conditions, despite the numerous geographic distances and challenges.^41–43 This highlights the tremendous care provided by primary health care teams within Nunavik and the resilience of the patient population. However, compared with existing studies, in which the average ages of patient populations were 50 to 60 years, our patient cohort was considerably younger at an average age of 40 years. Given that older age is associated with greater complications in emergency general surgery,^39,44 it is notable that our complication and mortality rates were similar to that of older cohorts of patients in the literature. This finding may be at least partially explained by the impact of systemic barriers in reaching definitive emergency surgical care.

We examined patient outcomes for a cohort of patients from remote communities, albeit without a comparative urban cohort. Prior literature has shown conflicting results on the effects of rurality and distance from home to hospital on the outcomes of emergency general surgery. Diaz and colleagues found a doubling of the odds of death for each 132 miles between their academic medical centre and the patient’s home zip code among patients who underwent nontrauma emergency general surgery in the US.⁴⁵ Conversely, in their study of patients who underwent emergency general surgery in Scotland, Wohlegemut and colleagues found decreased inpatient mortality among those who lived farthest from the hospitals, as well as those from the most rural category.⁴⁶ However, this study did not examine the time delays to admission, which may be the important factor affecting patient outcomes. The degree of remoteness and distance travelled in areas of Scotland may be different from the North American context. Salih and colleagues similarly found no effect of travel time to hospital on mortality for patients undergoing emergency laparotomy.⁴⁷ However, distances were short, with a median travel time of 16 minutes to the hospital.

The predictors of poor postoperative outcomes identified in our current study overlap with those described in the literature for patients with ACS conditions. In particular, ASA score, ESS, and sex have been previously linked to postoperative outcomes.^29,48–52 Importantly, we also identified time to presentation as a predictor of complications, although this association was attenuated and did not reach statistical significance in our sensitivity analysis excluding outliers. Results of the sensitivity analysis suggest that the relationship between delays and outcomes is largely driven by patients with the most prolonged presentations. However, it is important to highlight that the outliers excluded in the sensitivity analysis represent true clinical scenarios of prolonged delays that often occur in remote northern Quebec, reflecting real system-level challenges that are clinically relevant. This is an important modifiable factor for our cohort of patients, with previous literature also showing that delays in surgical interventions are linked to greater mortality among acute surgical patients.^7,8,53

Delays in care can be viewed through the Three Delays framework, which is often used to evaluate maternal health care, whereby the 3 delays reflect delay in seeking care, reaching care, and receiving care.⁵⁴ Although not specifically examined in this study, all 3 types of delays are relevant to our patient population. Delays in seeking care have been attributed to financial or geographic barriers, cultural beliefs, poor education, low awareness of or low confidence in available services, and a history of being disconnected from formal health systems, most of which have not been formally studied within the Inuit population.⁵⁴ Most importantly, Indigenous patients may have previous negative experiences of discrimination within the Canadian health care setting and fear of judgment based on race or social class, which can lead to delays in seeking care.²² Furthermore, historical trauma can result in distrust in government-provided health care. Delays in reaching care arise when hospitals with surgical capacity are scarce, requiring prolonged travels for patients.⁵⁴ Substantial delays in transferring critically ill patients have been previously documented in Nunavik.^23,24 Delays in accessing care within the hospital setting have been documented in Canadian health care systems. McIsaac and colleagues found that almost 20% of patients requiring urgent and emergency surgeries experienced delays exceeding the institutionally defined acceptable wait times at a Canadian tertiary care centre.⁷ Delay was associated with increased mortality, LOS, and total costs.

The existing body of research offers limited insights into optimizing transfer logistics for patients with ACS conditions originating from rural areas. Nevertheless, we propose several avenues for quality improvement initiatives based on the results of this study. Importantly, the association between prolonged transfer time and poor postoperative outcomes highlights a specific area that can be targeted for improvement. Increased aircraft availability and direct medevacs from each village could play a pivotal role in expediting transfers to tertiary care centres. Currently, patients must travel from their initial village of origin to the hospital centres in Kuujjuaq or Puvirnituq before traveling to Montréal. Allocating more resources to northern villages to facilitate transport from the patient’s home village directly to the urban tertiary care centre would reduce travel time substantially. We suggest optimization of point-of-care resuscitation and stabilization capacity. For this, valuable guidance can potentially be drawn from the trauma literature.⁵⁵ The Rural Trauma Team Development Course was formulated in response to poor outcomes among rural trauma patients.⁵⁶ This program is designed to impart essential knowledge and skills for the initial assessment and stabilization of trauma patients in resource-constrained environments.⁵⁷ The implementation of this program demonstrated a noteworthy decrease in the time required to initiate transfer to a tertiary centre and resulted in a significant reduction in mortality.⁵⁸ A similar initiative for ACS care could help improve patient outcomes. A critical mass of skilled health care workers must be maintained in Nunavik to avoid clinic closures and the associated delays in reaching stabilizing and definitive care. We propose community outreach initiatives in conjunction with local health care partners — notably, community leaders and Inuit-led organizations — to help address the complex factors associated with delays in seeking care. Interestingly, diverticulitis and bowel obstruction or hernia accounted for approximately half of the cases with complications while comprising only 20% of all patients who underwent ACS surgical intervention. Given that both diverticulitis and bowel obstruction are conditions known to carry inherently high rates of complications compared with other ACS conditions,^59,60 these findings further highlight the potential value of earlier surgical involvement. Earlier surgical referral could help prevent acute, severe presentations that may lead to higher morbidity. Focused education among health care providers on the management of these conditions, along with clear referral algorithms to prompt surgical evaluation, could further improve patient outcomes. Lastly, our analysis may prove useful in advocating for on-site surgical resources for emergency surgical care, including surgical team personnel and on-site OR facilities.

Limitations

Transferred patients from rural regions often present with more advanced illness and more comorbidities than patients directly admitted through the ED.^61–64 Moreover, Indigenous populations have higher prevalences of important comorbidities (e.g., diabetes, obesity, hypertension) that are important determinants to surgical outcomes.^18,19 Detailed comorbidity data were not consistently available in the clinical records from the ACS and emergency medicine services used for the retrospective study. As a result, we did not include detailed information on comorbidities, and we used ASA and EES scores as proxies to represent comorbidity burden and physiologic reserve. Given the single-centre, retrospective nature of our study, although our analysis controlled for potentially confounding variables, we could not establish with certainty whether the primary driver of postoperative outcomes was delayed surgical attention or whether outcomes were inherently linked to the initial severity of patient presentation. We could not obtain granular data on details of the transport trajectory (i.e., delays at each time point and reasons for delay, including weather, equipment, or personnel-related issues). We included Clavien–Dindo complications grades 1 to 5 in our analysis, without differentiating between the severity of the complications, given the limited sample. Combining more serious adverse events with minor ones makes it difficult to deduce the true clinical impact of the associated predictors. Moreover, we did not have a comparison group from other rural areas and were therefore unable to comment on whether the associations were attributable to the Nunavik region or rurality as a construct. Our next step is to compare outcomes between this study’s cohort and other patients with ACS conditions of different geographic origins. Lastly, the patients included in this study were specifically those transferred from Nunavik only, a region with a population that is 90% Inuit. Therefore, this group of patients is not representative of all rural regions, nor of all Indigenous patients nationally and globally, which affects the study’s generalizability. This, however, also represents a main strength of this study as it applies a distinctions-based approach and is more representative of the intricacies related to care of Nunavik’s Inuit population.⁶⁵

Conclusion

This study identified that delays in presentation from symptom onset and initial severity of illness upon arrival were predictors of adverse postoperative outcomes among patients with ACS conditions transferred from Inuit communities in Nunavik to an urban tertiary care centre. Prospective quality improvement endeavours should prioritize mitigation of transport delays, enhancement of point-of-care resuscitation, and ultimately, augmenting access to on-site surgical care to improve clinical outcomes for patients with ACS conditions from Nunavik.