Introduction

Chronic pain is a debilitating health condition estimated to affect 1 in 5 adults worldwide [1] with profound emotional and cognitive effects [2] and economic impacts, including loss of productivity, absenteeism due to illness, presenteeism, and loss of employment [3]. In the United States alone, 100 million adults — more than the number affected by heart disease, diabetes, and cancer combined — suffer from common chronic pain conditions, with an estimated loss in productivity due to chronic pain being $61 billion per year [4]. Several epidemiologic studies have shown associations between chronic pain and mood disorders, including generalized anxiety and depression [5,6,7,8]. More than half of people treated for depression also report problems with pain [9]. A growing body of evidence shows patients suffering from chronic physical pain are more likely to report suicidal behavior, ideation, plans, and/or attempts [10,11,12,13,14]. Of note, patients with chronic pain are at an increased risk of dying from suicide compared to those without chronic pain [13, 14]. Few previous studies have examined the association between chronic pain and depression and generalized anxiety in South America, including in Colombia [8, 15], Brazil [7, 16, 17], Paraguay [7] and Ecuador [7]. Only two of these previous studies have examined the association between pain and suicidal behaviors [18, 19], with one study reporting null findings [19].

Given that (1) few studies have examined pain with depression, generalized anxiety, and suicidal behaviors in South American populations; (2) one previous study found no association between pain and suicide risk in Colombia [19], and (3) no study has been conducted in Chile, the most stable and prosperous country in South America, we conducted a larger-scale cross-sectional epidemiological study to investigate the association of chronic pain with symptoms of depression, generalized anxiety, and suicidal behavior among injured working adults in Chile.

Materials and methods

Study population

The sample for this study included participants of the Stress, Pain, Sleep, and Neuropsychiatric Disorders (SPLENDID) study in a workers’ compensation hospital system in Chile. The SPLENDID study was designed to examine pain, work-related stress, and neuropsychiatric outcomes in the context of injured working adults in Chile with the intention of developing workplace intervention programs. The study was conducted between September 2015 and February 2018 among patients attending the Hospital del Trabajador in Santiago, Chile. The Hospital del Trabajador is the largest workers’ compensation hospital and the referral center for trauma and professional diseases of Asociación Chilena de Seguridad, the biggest workers’ compensation system in Chile, with almost 2.5 million affiliated workers. Working adults were eligible for the study if they attended the workers’ compensation hospital for the following injuries: spinal cord, mild brain injury, bone fractures, burns, and soft tissue injuries of various etiologies. Eligible participants also must be able to read and write Spanish. All participants provided written informed consent. The institutional review boards of the Hospital del Trabajador, Santiago, Chile, and the Office of Human Research Administration, Harvard T.H. Chan School of Public Health, Boston, MA, approved all procedures used in this study.

Analytic population

A total of 2000 participants were interviewed. Of these, 21 participants were excluded due to missing information on the Short Form McGill Pain Questionnaire, 4 were excluded for missing information on the Patient Health Questionnaire-9, 8 were excluded for missing information on the Generalized Anxiety Disorder-7, 9 were excluded for missing data on CSSRS suicidal ideation, 22 were excluded for missing data on CSSRS suicidal behavior, and 1 participant was excluded for being outside the age range (> 85). A total of 1946 participants were included in the final analysis.

Study procedures

Pain assessment

The Short-Form McGill Pain Questionnaire (SF-MPG) was used to measure current perceived pain among adults with chronic pain [20]. The SF-MPG comprises two subscales: a sensory subscale with 11 words or items and an affective subscale with 4 words or items. The sensory subscale includes items that describe the properties of the pain experienced (i.e., continuous, intermittent or neuropathic pain), whereas the affective subscale includes items that describe the emotional perception of pain (i.e., tiring-exhausting, sickening, fearful, and punishing-cruel) [21]. The items are rated on a Likert Scale for pain intensity with 0 = none, 1 = mild, 2 = moderate, and 3 = severe. The total score is obtained by summing the item scores. The sensory subscale total score ranges from 0 to 33, and affective subscale total score ranges from 0 to 12. The SF-MPG total overall score ranges from 0 to 45. We used the median score of the two subscales to define high and low pain scores (SF-MPG median: 11.0, SF-MPG sensory subscale median: 9.0, SF-MPG affective subscale median: 2.0). This approach is consistent with prior studies that summarized pain scores based on their distribution in the absence of a clinical diagnostic cutoff [22, 23]. In addition, the median split is a common method used to dichotomize variables in psychological research [24, 25], and in other fields [26], and oftentimes it is viewed as a more parsimonious way to split continuous data [24]. In addition, the cut-off we used to group between low and high pain in our study was consistent with mean scores [27, 28] and median scores [29] reported in the literature for the SF-MPG.

Depressive symptoms assessment

The Patient Health Questionnaire-9 (PHQ-9) was used to screen participants for depression. The PHQ-9 contains 9 questions about a participants’ depression symptoms over the 14 days prior to assessment. The items were scored from 0 to 3 according to the response categories: never, several days, more than half the days, or nearly every day. The total PHQ-9 score ranged from 0 to 27, and presence of depressive symptoms was defined as a PHQ-9 score ≥ 10 based on the accepted diagnostic properties for depression at this cut-off [30]. The PHQ-9 was previously validated in Spanish-speaking populations [31, 32] and specifically in the Chilean population [33].

Generalized anxiety symptoms assessment

Generalized anxiety symptoms was assessed using the Generalized Anxiety Disorder Scale (GAD-7). The GAD-7 contains seven questions about generalized anxiety symptoms over the 14 days prior to assessment. Items were scored from 0 to 3 according to response categories: never, several days, more than half the days, or nearly every day. The total score ranges from 0 to 21. The presence of anxiety symptoms was defined as a GAD-7 score ≥ 10, being this a reasonable cut-point to identify moderate to severe cases of GAD [34]. The GAD-7 has previously been used in Spanish-speaking populations [35, 36].

Suicidal assessment

Suicidal ideation and suicidal behaviors during the past month were assessed using the Columbia Suicidal Rating Scale (C-SSRS) Screen version [37, 38]. The C-SSRS Screen is a structured interview based on the more comprehensive full-length version. The scale measures suicidal ideation severity and suicidal behavior using two subsets of items. The first subset captures suicidal ideation severity during the past month. The questions ask: wish to be dead, non-specific active suicidal thoughts, suicidal thoughts with methods, suicidal intent, and suicidal intent with plan. The second subset measures suicidal behavior during the past three months. The questions asked about the presence of actual and aborted suicide attempts. For the purposes of the present study, suicidal ideation was defined by a “yes” answer to any one of the five suicidal ideation severity items. Suicidal behavior was defined as a “yes” answer to the suicidal behavior question. The C-SSRS has previously been validated in Spanish-speaking populations [39, 40].

Covariates

Structured questionnaires were used to determine participants’ sociodemographic and occupational characteristics. Participants age was categorized as 18–24, 25–34, 35–44, 45–54, 55–64, 65–74, and 75–84 years. Other sociodemographic characteristics examined were sex (male vs. female), country of birth (Chile vs. others), belonging to indigenous or native groups (no vs. yes), highest degree of education attained (elementary school, high school, college or technical training), marital status (married/living with a partner, single, previously married), body mass index (BMI; < 18.5, 18.5–24.9, 25–29.9, > 30 kg/m2), lifetime smoking (no vs. yes), and lifetime alcohol consumption (no vs. yes). Participants’ occupational characteristics included work sector (construction, finance, commercial, manufacturing, public services, transportation, others), type of occupation (administrative, manual worker, professional, salesperson, technician, teacher, others), time since the accident in days (1–42 days: Acute; 43–84 days: Subacute; > 84 days: Chronic) [41, 42], type of accident (commute or work injury), and type of injury (burn, fall, cut, attrition, firearm, blunt trauma, repetitive use, others) [43].

Statistical analysis

Frequency distributions of sociodemographic and occupational characteristics were examined using mean ± standard deviation (SD) for continuous variables and numbers and percentages (%) for categorical variables. Associations between pain status and sociodemographic and occupational characteristics were examined using Chi-squared tests for categorical variables and analysis of variance (ANOVA) for continuous variables. The median value [25th and 75th percentiles] was calculated for time since the accident, as the distribution of these values was right-skewed. Wilcoxon-Mann Whitney test was used to evaluate differences in medians. PHQ-9 and GAD-7 scores were used to create categorical variables: depression or anxiety symptoms (no vs. yes); and mood disorder status (no depression or anxiety, depression only, anxiety only, or both depression and anxiety). Multivariable logistic regression procedures were used to calculate odds ratios (ORs) and 95% confidence intervals (95% CI) for the association of chronic pain with depressive symptoms, generalized anxiety symptoms, and suicidal behaviors. We fit separate models with symptoms of depression, generalized anxiety and suicidal behaviors as outcome and chronic pain as exposure. Assessment of confounders was performed by entering a priori selected putative confounders (e.g., age, sex, marital status, type of workplace and type of injury) [44,45,46] into a logistic regression model one at a time and by comparing the adjusted OR (aOR) and unadjusted ORs. Final logistic regression models included covariates that altered unadjusted ORs by at least 10%. Additionally, the suicidal ideation and behavior models included further adjustment for depression status as measured by the first eight questions of the PHQ-9 [46, 47]. Question 9 was not included since it asks about suicidal ideation (‘Thoughts that you would be better off dead, or of hurting yourself in some way’). The PHQ-8 depression questionnaire has been demonstrated to minimally influence scale performance, mean scores, or diagnostic cut points compared with the PHQ-9 [48, 49]. Statistical analyses were performed using SPSS Statistics, Version 23.0 (IBM SPSS v23.0, Armonk, NY, USA).

Results

Sociodemographic characteristics of the study population are shown in Table 1. The mean age of participants was 45.9 ± 13.7 years old. Participants were more likely to be men (72.9%) and born in Chile (94.1%). Only 3.2% of participants self-identified as belonging to an indigenous or native group. Most participants had at least a high school or some college education (83.9%), were married or living with a partner (62.1%) and had a lifetime history of smoking (58.4%) and alcohol consumption (61.8%). Approximately 46% of participants were classified as having high perceived chronic pain as measured by the SF-MPG. Participants with high pain were more likely to be older, men, and be more highly educated (p < 0.001; Table 1).

Table 1 Sociodemographic characteristics of a population of injured working adults in Santiago, Chile according to pain status as measured by the Short-Form McGill Pain Questionnaire a (N = 1946)
Full size table

Participants worked in the sectors of manufacturing (26.6%) or public services (21.4%). The majority of participants were most likely to classify their occupation as manual workers (58.1%). The median time since their accident was 188 days. The majority of accidents occurred during work (65.6%) and were attributed to falling (35.5%) or blunt trauma (32.6%). Participants with high pain were less likely to be manual workers, reported a longer time since the accident occurred, and were also more likely to have falling or blunt trauma injuries (p < 0.001; Table 2).

Table 2 Occupation and injury characteristics of a working adult population in Santiago, Chile according to pain status as measured by the Short-Form McGill Pain Questionnaire a (N = 1946)
Full size table

Psychiatric characteristics of the population are presented in Table 3. Overall, 38.2% of the population had depression symptoms, and 23.8% had generalized anxiety symptoms. 13.4% of the study population had suicidal ideation, and 2.4% had suicidal behavior. Participants with high pain were significantly more likely to have depression or generalized anxiety symptoms, suicidal ideation, and suicidal behavior (Table 3).

Table 3 Psychiatric characteristics of a population of injured working adults in Santiago, Chile according to pain status as measured by the Short-Form McGill Pain Questionnairea (N = 1946)
Full size table

Participants with high pain had 4.05-fold increased odds of experiencing symptoms of depression or generalized anxiety compared to those who reported low pain after adjusting for sociodemographic and occupational confounders (aOR = 4.05; 95%CI: 3.32–4.94). Compared to participants with low pain, those who reported high pain had higher odds of depression (aOR = 2.74; 95% CI: 2.13–3.52) and generalized anxiety symptoms alone (aOR = 2.16; 95% CI: 1.32–3.53), and co-occurring symptoms of depression and generalized anxiety (aOR = 7.19 95% CI: 5.45–9.51). After adjusting for putative confounders, participants with high pain had 3.48-fold increased odds of reporting suicidal ideation than those with low pain (aOR = 3.70; 95%CI: 2.72–5.04). Further adjustment for depressive symptoms attenuated this association but remained strong (aOR = 2.34; 95%CI: 1.68–3.27). Participants with high pain had 4.05-fold increased odds of reporting suicidal behavior compared to those with low pain (OR = 4.05; 95%CI: 1.92–8.54). After adjusting for symptoms of depression, the magnitude of the association between high pain and suicidal behavior was attenuated (OR = 2.35; 95%CI: 1.08–5.07; Table 4). We further conducted an additional analysis using the SF-MPG continuous score, and the interpretation remained the same, but the associations were largely attenuated (Supplementary Table 1). The associations between pain and symptoms of depression, generalized anxiety, and suicidal behaviors were similar when analyzed separately using the SF-MPG sensory pain and affective pain subscales (Supplementary Tables 2 and 3).

Table 4 Association of pain with depression, anxiety, and suicidal behavior among a population of injured working adults in Santiago, Chile (N = 1946)
Full size table

Discussion

Pain is highly prevalent and contributes greatly to morbidity and mortality. In a review of 18 adult population surveys from 10 high-income countries (HICs) and seven low- or middle-income countries LMICs (N = 42,249), the prevalence of chronic pain was 37.3% in HICs and 41.1% in LMICs [8]. In our study of injured Chilean working adults, 46% of participants were classified as having chronic pain, and they were more likely to be older, men, current smokers, and had higher levels of education compared to participants with low pain. Our study findings and available data collectively show that chronic pain is generalized. The positive association that we observed between chronic pain and age was largely consistent with findings reported from other epidemiologic studies [50, 51]; however, the prevalence of pain across demographic characteristics such as gender and education level in our study differed from previous evidence [51], including that from a study conducted in Chile and another in Colombia [52, 53]. These differences in the prevalence of chronic pain might be attributable to differences in populations included, measurement scales used, and study designs employed.

Chronic pain was associated with depression and anxiety spectrum disorders in both LMICs and HICs [8]. In a recent review of 47 LMICs (N = 273,952), depression was significantly associated with severe pain in 44 of the 47 countries included (OR = 3.93; 95% CI 3.54–4.37) [7]. Neither of these prior reviews includes associations between pain and mood disorder symptoms in a Chilean population. This is important as chronic pain has been reported to be highly prevalent in 32% of the general population [52]. To the best of our knowledge, our study is the first to examine this association in a Chilean population. Taken together, our study and previous work demonstrate a robust association between chronic pain and mood disorders, including depression and anxiety.

Previous studies have examined the association between chronic pain and suicidal behaviors. Several studies conducted in the United States [54, 55], Australia [56], Canada [57], Japan [58], and India [59] have shown associations between chronic pain and suicidal ideation or behaviors. For example, among U.S. adults (N = 5692), the presence of a pain condition was significantly associated with lifetime suicidal ideation (OR = 1.4; 95%CI: 1.1–1.8) and plans (OR = 1.5; 95%CI: 1.1–2.1) [55]. Among Canadian adults over 66 years old, completed suicide was associated with moderate (OR = 1.91; 95% CI: 1.66–2.20) and severe pain (OR = 7.52; 95% CI: 4.93–11.46). Among patients admitted for attempted suicide and controls at a teaching hospital in India (N = 137 patients and N = 137 controls), those with idiopathic pain (OR = 6.78; 95% CI: 2.39–20.76) and physical illnesses (OR = 3.12; 95% CI: 1.37–7.24) had an increased odds of suicide attempts compared to controls [59]. To our knowledge, only one study has examined the association between pain and suicide risk in South America. In an underpowered study conducted at a hospital pain clinic in Colombia (N = 49), the authors found no association between the risk of suicide and perceived levels of pain [19]. In our study, participants with high pain were more likely to report suicidal ideation; however, the association was not statistically significant.

Chronic pain is highly prevalent and is associated with increased odds of depression, generalized anxiety, suicidal ideation, and suicidal behavior among working Chilean adults. The findings of this study and those from other populations [54, 55], reinforce the knowledge that chronic pain and associated psychiatric symptoms may not be independent entities but rather comorbid symptoms interacting synergistically [60, 61]. This is consistent with the five most prevalent co-occurring symptom clusters labeled as the “SPADE pentad,” which include chronic pain, anxiety, depression, sleep disturbance, and low energy/fatigue [60]. The co-occurrence of SPADE symptoms negatively affects treatment response and undermines patients’ general health, quality of life, and physical functioning [60]. The recent Institute of Medicine report indicates chronic pain is debilitating and requires immediate, appropriate treatment rather than being sidelined [2]. Given its comorbidity with psychiatric disorders, an interdisciplinary approach that addresses the burdensome co-occurring SPADE symptoms will be most promising for patients with chronic pain and optimize their quality of life and other health outcomes [60].

The results of our study are compatible with the current understanding of the biological mechanism of chronic pain and psychiatric outcomes [50, 62, 63]. Functional imaging studies suggest that this bidirectional relationship is partly attributed to shared neural mechanisms. For instance, participants experiencing lower back pain, fibromyalgia, or other types of chronic pain, show functional imaging alterations in regions responsible for the response to emotional stimuli, like the anterior cingulate cortex and the prefrontal cortex regions [64]. In comparison, participants reporting depressive symptoms showed a shift in the insula region responsible for processing emotional stimuli to a region associated with processing pain in pain-free individuals [64]. Similar evidence has been reported for anxiety and other mood disorders [64].

Some important limitations must be considered when interpreting the results of our study results. First, our study’s cross-sectional data collection design does not allow for determining the temporal relationship between chronic pain and psychiatric outcomes. Second, measures of chronic pain were mainly related to the type of work injury reported by participants at the time of enrollment in the study; thus, we were unable to account for the influence of chronic pain predating the work injury. Furthermore, our use of self-reported data in the exposure, outcome, and other covariates, may have introduced some degree of measurement error. Although we used multivariable logistic regression procedures to adjust for putative confounders, we cannot exclude the possibility of residual confounding. Finally, the participants in this study were injured working adults in Santiago, Chile; thus, the results may not be generalizable to other populations.