Introduction

People with severe mental illness experience significantly worse surgical outcomes including greater postoperative morbidity, longer stays in hospitals and higher risk of readmissions compared to the general population [1,2,3,4,5]. A few studies have focused on the adverse prognosis for patients with comorbid mental illness undergoing major abdominal emergency surgery [1,2,3,4]. An overall underuse of major surgery for patients with severe mental illness has been identified [6], as well of more severe disease presentations upon admission in emergency situations—likely due to patient- and system-related delays in symptom recognition [4].

Peptic ulcer disease is a common condition that both primary care and secondary care providers encounter [7, 8]. The detection of the correlation between the Helicobacter pylori bacteria and peptic ulcer disease has changed the understanding of the disease. Symptoms of peptic ulcer disease are variable and may include abdominal pain, nausea, vomiting and weight loss. Complications of peptic ulcer disease include perforation and bleeding. Recent evidence suggests that hemorrhage is the most common complication [8]. Although the global prevalence of peptic ulcer disease has dramatically decreased in the past decades, the incidence of its complications has remained constant [7]. People with severe mental illness have an increased risk of developing peptic ulcers [9]. Individual factors related to nutrition, socioeconomic status, hygiene and sanitation may explain some of this variation. [7]

Both system- and patient-related factors may contribute to the poorer outcomes in patients with comorbid mental illness. System-related factors may play an important role, since people with mental illness may experience a systematic failure of the healthcare system to prevent, identify and treat physical diseases [10,11,12]. Stigmatization of patients with mental illness by healthcare professionals could be a contributing factor [13,14,15,16]. Patient-related factors may include low socioeconomic status, poor social network, impaired bodily sensation, disturbed perception of pain, poor health literacy, adverse lifestyle factors and physical comorbidities. These factors may all act as mediating factors of mental illness contributing to higher prevalence, later hospital arrival, higher urgency of disease upon admission and poor outcomes. [4, 17,18,19,20]

The timing of surgery is crucial in patients with perforated ulcers [21,22,23]. Investigating initial management, including the prehospital assessment and care, is important to improve quality of care. Also, including patients with different severities of mental illness may improve the understanding of the potential impact of comorbid mental illness. The aim of this study was to compare patients with and without a history of mental illness on process and outcome measures in relation to prehospital and emergency surgical care for patients with perforated ulcer.

Methods

Study design and setting

This was a nationwide cohort study combing patient-level data from Danish Clinical Register of Emergency Surgery [24] and the Danish Quality Database for Prehospital Emergency Medical Services (Prehospital Database) [25] with other nationwide Danish health registries.

The study was reported according to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines for observational studies.

A unique ten-digit Civil Personal Register (CPR) number is assigned to all citizens in Denmark. The CPR allows for individual-level record linkage of Danish registers [26]. Both the Register of Emergency Surgery and the Prehospital Database are organized under the publicly funded organization Danish Clinical Registries [27]. Reporting to the Danish Clinical Registries is mandatory for all hospitals including the emergency medical services (EMS).

The study was conducted in Denmark, which has a population of 5.8 million. The country has five healthcare regions responsible for healthcare services, and each region has an organization responsible for the entire EMS, including the regional Emergency Medical Communication Centre and the ambulance services. The healthcare system is primarily financed through taxation and has free access to healthcare services including access to EMS, general practitioners (GP) and hospital care. Referral from a GP or out-of-hour GP is mandatory prior to all hospital contacts, excluding psychiatric emergencies and patients arriving by ambulance after calling EMS. Emergency surgical care is provided only at public hospitals in Denmark [28]. During the study period, the number of hospitals receiving emergency patients—including emergency surgical patients—was reduced from 44 to 21 hospitals based on recommendations from the Danish Health Authorities in 2007 [29]. Since 2010, hospitals and GPs have access to an online Shared Medication Record [30] though the electronic medical record, providing information on the citizen’s medication including most recent prescriptions and dates of prescription and redemption. Hospitals and GPs also have online access to the hospital-based health records across hospitals providing information on, e.g., diagnosis and discharge letters.

Police or fire brigade personnel answer all EMS calls. Calls identified as health-related are forwarded to one of five public EMCCs [31]. At the EMCC, healthcare professionals handle these calls. The Emergency Medical Communication Centre personnel assess the individual calls using “Danish Index for Emergency Care,” a criteria-based dispatch decision support tool. Each call is assigned a main symptom, which can be selected among 37 standardized mechanisms or symptoms (e.g., symptom card 24 “Abdominal pain or back pain”). Depending on the type and urgency of the symptoms (e.g., sudden severe pain or vomiting red blood), the call is assigned a priority level from “A” to “E.” The highest priority level is urgency level A, corresponding to an immediate response, i.e., dispatching ambulance as category/priority 1.

Identification of study population

The study population consisted of consecutive patients admitted from January 1, 2004, to December 31, 2018, identified in the Danish Clinical Register of Emergency Surgery. Only patients registered with a diagnosis of perforated ulcer (ICD-10 DK251, − 252, − 255, − 256, − 261, − 262, − 265, − 266, − 271, − 272, − 275, − 276) in combination with a surgical procedural code (any “KJ”-code in the administrative system covering all surgical gastrointestinal interventions) who had the initial surgery performed within 48 h from hospital arrival were included.

Patients were excluded if the patient had immigrated to Denmark less than 10 years prior to the admission or if the patient was not registered as living in Denmark at the time of the admission (Fig. 1). The 10-year cutoff was set due the definition of the exposure, requiring patient’s mental health record for at least 10 years.

Fig. 1
figure 1

Flowchart of included admissions from Danish Clinical Register of Emergency Surgery 2004–2108 who has surgery performed within 48 h from hospital arrival

Full size image

The cohort was combined with data from the Prehospital Database covering the years 2016–2017 using the CPR number. If an EMS call was made 72 h prior to hospital arrival, the patient was defined as “EMS patient.” If no EMS call was identified 72 h prior to hospital arrival, the patient was defined as “non-EMS patient.” We chose 72 h, since patients with a history of mental illness are more likely to receive telephone advise or be released at scene, when they call the EMS [32], and may have called more than once or contacted different health professionals prior to admission.

Ethics

The study was approved by the Danish Data Protection Agency. According to Danish law, approval from an ethical committee is not required for register-based studies.

Variables

Definition of exposure: mental health history

Patients were categorized according to the severity of their mental health history: Major, moderate, minor or none based on information from The Danish National Patient Registry, The Danish National Patient Registry-Psychiatry, The Danish Psychiatric Central Research Register, The National Register of Medicinal Product Statistics and The Danish National Health Service Register. If the patient fulfilled criteria in several categories, the most severe category defined the patient’s exposure.

  1. 1.

    Major mental illness: An inpatient or outpatient contact with schizophrenia or bipolar disorder (ever), or inpatient contacts with severe depression or emotionally unstable personality disorder (within the last 10 years).

  2. 2.

    Moderate mental illness: Other inpatient or outpatient psychiatric contacts with diagnoses of mental illness (other than “major”) or consultations at a private practicing psychiatrist (within the last 5 years).

  3. 3.

    Minor mental illness or vulnerability: None of the above. Reimbursement of at least two prescription of selected drugs (antidepressants or benzodiazepines) or two or more sessions of talk therapy or psychometric testing in a primary care setting or referral to a private psychologist (within the past 12 months).

  4. 4.

    No mental illness: None of the above.

While there is no gold standard for categorizing mental illnesses, it is generally accepted that major mental illness includes schizophrenia and bipolar disease [10]. Hospital-based diagnoses are often used to identify mental illness in registry-based studies [33] and is considered to be more severe than mental health conditions requiring community-based services only (e.g., GP or psychologist services). Please see Supplementary Table 1, for details on included ICD-10 diagnoses, other codes (e.g., ATC codes) and data sources. Please see Supplementary Table 2 for information about timing of most recent mental health disease-related activity.

Definition of outcomes: Prehospital care and emergency surgical care

We defined prehospital care and emergency surgical care using process performance measures and outcome measures from the Danish Prehospital Database and the Danish Clinical Register of Emergency Surgery. See Table 1 for details on each measure. The Danish guidelines for emergency surgical care are in accordance with the guidelines from World Society of Emergency Surgery (WSES). [8]

Table 1 Performance measures from the Prehospital Database and the Danish Clinical Register of Emergency Surgery
Full size table

While we used the measures from the Emergency Surgery Registry 2004–2018, some measures have been updated since 2018 to reflect new evidence and best clinical practice within the field. Since 2019, the registry includes other high risk surgical patients than patients with perforated ulcers. Also, measures have been added regarding, e.g., timing of CT abdomen, measurement of serum lactate and timing of preoperative optimization. These measures were not available when we received our dataset in 2019. However, the original measures from 2004 to 2018 still reflect the provision of the core processes of early surgical care and emergency care for patients with perforated ulcers.

For the combined outcome “Days alive and out-of-hospital,” we applied the same methodology as described by colleagues [34]: At 90-day follow-up post-surgery, we subtracted the length of stay of the initial admission and all readmissions within this time period to calculate the number of days alive and out of hospital. If a patient died within the 90 days, number of days alive and out of hospital was defined as 0 days.

Covariates

Covariates available for all patients included: Age, sex, comorbidity (Charlson comorbidity index), American Society of Anaesthesiologists Classification (ASA) score, type of surgery, smoking status, social position, cohabitation status and drug abuse. Detailed definitions of covariates and data sources are presented in Supplementary Table 3.

Supplementary Fig. 1 provides a directed acyclic graph (DAG) presenting covariates as confounding factors or mediators.

Statistical methods

We compared all measures according to the severity of mental illness using the group with no history of mental illness as a reference.

For the time delays (time to surgery and time to antibiotics), we compared the exposure groups applying robust linear regression based on Huber and biweight iterations, as implemented in the Stata routine rreg. The purpose of this regression was to limit the effect of serious outliers. Please see boxplot Supplementary Fig. 2 for the distributions of time to surgery and time to antibiotics including outliers. We performed an additional analysis comparing medians and 75 percentile to further explore the differences between groups. Moreover, we estimated a concordance index (c-index), which is rank-based statistics ranging from zero to one. Thus, the c-index is a measure of how well the exposure groups discriminate the outcome, with c-index equal 0.5 means zero discrimination, while a c-index equal zero or one means full discrimination. Adjusted c-indices were estimated using inverse-probability-of-treatment-weights, i.e., first, the probability of being exposed delayed surgery is calculated, given an individual’s characteristics. Second, weights are calculated (the inverse of the propensity score). The application of weights creates a pseudopopulation in which confounders are equally distributed across exposed and unexposed groups.

For the analysis of differences in days-alive-and-out-of-hospital, we used linear regression. For the binary outcomes, we calculated a Risk Ratio (RR) using Poisson regression. To relax the assumption of independence for recurring subjects, we applied the clustered Huber variance estimator to these regressions.

All estimates were calculated as unadjusted and adjusted for age and sex (model 1) and for age, sex and Charlson comorbidity index (model 2). We also performed additional analyses adjusting for age, sex, Charlson comorbidity index and ASA score (model 3). Please see Supplementary Fig. 1: Directed acyclic graph for consideration of covariates as confounders or mediators. Also, since both mortality and the proportion of patients with a history of mental illness decreases in the study period (Supplementary Table 5), analyses regarding days-alive-and-out-of-hospital and 30-day mortality were adjusted for calendar year.

We implemented restricted cubic splines for age with three knots at quantile 0.1, 0.5 and 0.9.

Missing ASA scores and time to surgery were imputed implementing multiple imputation chained equations using 20 imputation sets. All outcomes, exposures and covariates were included as predictors. For the c-index, we used singular value imputation by only using the first imputation set of the multiple value imputation.

Statistical analyses were performed using Stata 16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC.).

Results

We identified 4.767 admissions with patients undergoing emergency surgery for perforated ulcer. Of these 25% had a history of minor mental illness or vulnerability, while 4% and 4% had a history of moderate or major mental illness, respectively. Characteristics of the population are shown in Table 2. A total of 84 patients (2%) were missing timestamp (hours and minutes) (supplementary Table 4). Please see Table 2 regarding missing values for each covariate.

Table 2 Characteristics of the patients undergoing surgery for perforated ulcers within 48 h from hospital arrival 2004–2018 in Denmark according to severity of mental health history
Full size table

Considering patients admitted 2016–2017 (n = 640), 153 (24%) called the EMS prior to admission. While 48% of patients with no history of mental illness were assigned an immediate response (Level A) when calling the EMS, only 38% and 25% of patient with a history of moderate or major history of mental illness, respectively, were assigned this level of urgency (Table 3). Among patients with no history of mental illness, 51% were identified with abdominal or back pain when calling the EMS (Table 3). This number was 40%, 31% and 25% for patients with a history of minor, moderate and major mental illness, respectively (Table 3).

Table 3 Contact to emergency medical services (EMS), presenting symptoms and proportion receiving a response with the highest level of urgency (level A) for patients undergoing surgery for perforated ulcers 2016–2017 according to mental health history
Full size table

For all patients in the cohort (2004–2018), the median time from hospital arrival to surgery was 6.0 h (IQR: 3.6;10.7) (Supplementary Table 4). Adjusting for age, sex and comorbidity, patients with a history of major mental illness underwent surgery 46 min (95% CI: 4;88) later compared to patients with no history of mental illness (Fig. 2a). Differences in medians and 75 percentile and c-indices can be found in Supplementary Table 6. The c-indices revealed a statistically significant, but small discrimination in treatment delay for patients with a history of major illness compared to no history of mental illness. The regression analyses on 75% quartile showed that for a portion of patients with a history of moderate or major mental illness, surgery was delayed compared to the patients with no history of mental illness.

Fig. 2
figure 2

Time differences (minutes) regarding time to surgery (2A left) and time to antibiotics (2A right) and differences in days-alive-and-out-of hospital (2B) between patients with a history of mental illness. Patient with no history of mental illness as reference. CCI Charlson comorbidity index, ASA American Society of Anaesthesiologists Classification

Full size image

Considering patients admitted 2014–2018, the median time from hospital arrival to antibiotics was 207 min (IQR: 90;423). (Supplementary Table 4). Adjusting for age, sex and comorbidity, patients with a history of major mental illness received antibiotics 72 min (95% CI: − 10;148) later, compared to patients with no history of mental illness (Fig. 2a). While differences were not statistically significant, notable differences were observed for patients with moderate and major mental health history, with a large proportion of patients receiving antibiotics very late (e.g., 75 percentile above 500 min) (Box plot supplementary Fig. 2 and Supplementary Table 4). Differences in medians, 75 percentile and c-indices can be found in Supplementary Table 6.

We found no differences regarding preoperative risk assessment nor preoperative hemodynamical optimization between groups (Supplementary Table 4)30-day mortality for all patients was 23% and number of days-alive-and-out-of-hospital at Days 90 was 67 days (IQR: 0;83) (Supplementary Table 4). Adjusting for age, sex and comorbidity, patients with a history of major mental illness had a higher risk of death at 30-day risk ratio 1.43 (95% CI: 1.14;1.80) and 9 days (95% CI: 4;14) less alive and out-of-hospital at 90-day follow-up, compared to patients with no history of mental illness (Fig. 2b). Patients with a minor history of mental illness had a higher risk of death at 30-day Risk Ratio 1.16 (95% CI: 1.04;1.29) and 4 days (IQR: 2;6) less alive and out-of-hospital at 90-day follow-up (Fig. 2b). During the study period, 30-day mortality decreased and the median number of days-alive-and-out-of-hospital at 90-day follow-up increased (Supplementary Table 5). We found no significant differences in 30-day mortality of days-alive-and-out-of-hospital among patients with a history of moderate mental illness (Fig. 2b).

Discussion

In this nationwide study, we found that patients with a history of moderate or major mental illness were less likely to be assigned an immediate response and less likely to be recognized with abdominal pain, if calling the EMS prior to hospital arrival. Further, we found that patients with a history of major mental illness underwent surgery and had antibiotics later, compared to patients with no history of mental illness. Patients with a history of major mental illness or minor mental illness/vulnerability also had a higher risk of 30-day mortality and had significantly fewer days alive-and-out-of-hospital at 90-day after surgery.

Our finding of a median of six hours delay from hospital arrival to surgery—regardless of mental health history—is long. This delay is longer than the national recommendations, and both local and national initiatives have been developed to decrease surgical delays by introducing, i.e., multi-disciplinary protocols [35,36,37]. The even longer delays for patients history of major mental illness is concerning and may be partly explained by the fact, that these patients were more likely to have atypical or unclear presentation when calling the EMS. This is likely to delay both antibiotics and surgery. Atypical and unclear presentation may be due to an altered pain response observed in patient with major mental illness [2, 4, 20, 38]. This may cause the longer patient delay that we found from onset of symptoms to hospital arrival—especially in the major group—if the patient is not aware of the urgency of the symptoms.

The importance of a bystander has been described in other time-dependent diseases (i.e., stroke [39]) and may be true for abdominal emergencies as well. However, an important system-related factor may also be the role of stigma and diagnostic overshadowing (e.g., assuming that a symptom is due to coexisting mental health condition rather than exploring the cause of the patient's symptoms) [13,14,15,16, 20]. Another Danish study, investigating EMS responses among selected emergency diseases, found an overall tendency among patients with perforated ulcers, to receive a lower urgency of EMS response compared to other time-dependent diseases, i.e., only 64% of patients received an immediate EMS response [40]. Hence, recognizing the severity of symptoms in perforated peptic ulcers may be difficult for both patients and health professionals, regardless of mental health history. Also, considering the low proportion of patients even calling the EMS prior to admission, compared to other time-dependent diseases, i.e., stroke and myocardial infarction [41] identifying these vulnerable subgroups may be a particular challenge, especially as patients with severe mental illness also have been reported to have an increased risk of developing peptic ulcers [9]. Ensuring correct system response for these high-risk patients is crucial. Staff competencies, stigma within the healthcare system and organizational health literacy including decision support tools supporting the health professionals dealing with vulnerable and high-risk patients must be addressed.

Our finding of worse outcome among patients with a history of moderate or major mental illness are in accordance with studies investigating different types of urgent surgery [1,2,3,4,5]. These studies also find greater postoperative morbidity, longer stays in hospitals and higher risk of readmissions [1,2,3,4,5]. Our finding of a higher 30-day mortality among patients with a history of major mental illness is in accordance with a study investigating outcomes in patients undergoing emergency laparotomy (for other causes than perforated ulcers) according to socioeconomic status. They found poorer outcome and higher 30-day mortality for patients with the most disadvantaged socioeconomic position patients compared to patients with a more advantaged socioeconomic position [17]. In our study, patients with a history of major mental illness also had the most disadvantaged socioeconomic position. The complex interplay between patient-related factors [4, 17,18,19,20] and system-related factors [10,11,12,13,14,15,16] within the healthcare system is likely to mediate or exacerbate poor outcome within this subgroup. Patient-related factors such as social network, individual health literacy, adverse lifestyle factors including nutrition and hygiene may also contribute to surgery-related complications and ability to participate in follow-up.

Our finding of higher ASA scores among patients with a history of either minor or major mental illness may also explain some of this variation. The higher ASA scores may be due to both higher burden of comorbidity, substance abuse, functional status, level of independency, as well as acute derangement. In addition, the poorer outcome among patients with a history of minor mental illness is likely due to higher age burden of comorbidity, which may also explain why we do not find a dose–response relationship between the different severities of mental illness and our outcomes. That is, patients with a history of minor mental illness experienced the worst outcomes, while patients with a history of moderate mental illness experienced almost equal outcomes, compared to patients with no history of mental illness.

In a larger perspective, our finding of fewer days-alive-and-out-of-hospital is in accordance with existing studies describing lost life years and excess mortality in general among patients with mental illness [10, 42,43,44,45]. While other studies investigated lost life years over a lifetime combing the contributions from emergency and chronic diseases, our study contributes with knowledge regarding lost life years in relation to urgent care.

Strengths and limitations

The data sources used in this paper represent a key strength with an overall high completeness and accuracy of the Danish databases and registries. However, several limitations should be considered.

First, although the study was based on population-based registries and the Danish Emergency Surgery Registry has been shown to covers 89% of all admissions with perforated ulcers [46], patients with mental illness may be more likely to have undetected events of perforated ulcers [6, 20, 38]. Hence, our results can only be generalized to patients, who are admitted to the hospital, recognized and treated for a perforated ulcer.

Also, information bias may have been introduced. The definition of exposure is a proxy measure of mental illness. It is defined based on consensus between local experts since there is no gold standard for these definitions. In particular, the definition of the minor illness group may be prone to misclassification, since these measures may express a state of mental vulnerability rather than mental illness. Supplementary Table 2 provides information on available measures of recent mental health disease-related activity.

Our estimates may also be skewed due to confounding. While we adjusted for some factors (e.g., age and sex), other factors (CCI and ASA score) are most likely mediators. Adjusting our estimates for both CCI and ASA score may remove some of the effect of having a history of mental illness. Either ways, residual confounding is expected to persist in these very heterogenic groups. In particular, patients with a minor history of mental illness may have an inherent frailty (considering their higher age, burden of comorbidity combined with mental illness) that we were not able to adjust for using the available measures.

Also, due to the small number of patients, especially in the moderate and major groups, estimates regarding antibiotics based on data from 2014 to 2018 had a moderate to low precision.

Conclusion

One-third of patients admitted with a perforated ulcer had a history of mental illness or vulnerability. Patients with a history of major mental illness had longer delays from symptom onset to hospital arrival. Patients with a history of major or moderate mental illness were less likely to be recognized with abdominal pain and less likely to be assigned the highest levels of urgency, if calling the EMS prior to arrival. Patients with a history of major mental illness had longer delays from symptom and from hospital arrival to surgery and showed a tendency to receive antibiotics later, compared to patients with no history of mental illness. They also had a higher risk of 30-day mortality and has less days alive-and-out-of-hospital at 90-day follow-up.