The Pain Suicidality Association: A Narrative Review

Abstract

Objectives

The objective of this narrative review was to examine recent evidence and, when necessary, past evidence on the association between pain and suicidality.

Design

Fifty-eight research reports were found on this topic, which had not previously been reviewed. These reports were divided into groups by whether they addressed suicide ideation (SI), suicide attempts (SAs), or suicide completion (SC), and what population they represented (acute pain patients [APPs], chronic pain patients [CPPs], other than APPs/CPPs) and whether they controlled for relevant confounders. Information as to whether the results of these studies supported/did not support the association of pain and suicidality was abstracted. For each group of studies (above), a vote counting method was utilized to determine the overall percentage of studies supporting/not supporting the association of pain and suicidality. According to this percentage, the consistency of the data supporting this association was rated according to Agency for Healthcare Research and Quality guidelines.

Results

The following groups of studies received an A consistency rating (consistent evidence from multiple studies): SI, SA, and SC for other than APPs/CPPs; and SI, SA, and SC for CPP prevalence greater than an appropriate control group. Also, a subgroup of the SI, SA, and SC studies for other than APPs/CPPs had controlled for behavioral issues (potential confounders). These three subgroups also received an A consistency rating. The 58 studies also identified a number of new predictor variables for SI, SA, and SC in CPPs.

Conclusions

These studies solidify the evidence for an association between pain and SI, SA, and SC in both CPPs and other than APPs/CPPs.

Introduction

At the present time, the relationship between chronic pain (CP) and suicidality is of significant concern to the pain clinician. Early non-CP literature [1–8] indicated that there may be a relationship between CP and suicidality. Perhaps, in response to this literature or as a general approach to the evaluation of depression, a major early textbook on pain contained a chapter on pain and depression. Here, a discussion on the importance of evaluating pain patients for suicidality was presented, although no evidence was referenced for this association (Ward NG, Pain and Depression, Chapter 18 in The Management of Pain, Bonica Editor, 1990). The first pain literature report on this association appeared in 1991. Here, in a long-term follow-up study (6 months to 1 year) of 2,146 chronic pain patients (CPPs) treated at a pain center, but not currently under that pain center's care, it was discovered that three patients had completed suicide [9]. The suicide rate for this CPP population was calculated and statistically compared with the US general population. The CPP suicide rate was found to be significantly greater than that of the general population (two to three times). Subsequent pain reports [10–19] also appeared to link CP and suicidality. This literature was subsequently reviewed [20]. The conclusions of this review were that suicide ideation (SI), suicide attempts (SAs), and suicide completions (SCs) appeared to be commonly found in CPPs and that CP may be a suicide risk factor [20]. As a result of this research in 2003, the American Psychiatric Association included CP as a risk factor for SC in their guidelines for the assessment and treatment of patients with suicidal behavior [21].

In 2006, Tang and Crane [22] again reviewed this literature as six new studies [23–28] had been added. They reported that the prevalence of SI within CPPs was approximately 20%, lifetime prevalence of SA was between 5% and 14% and that relative to controls, the risk of death by suicide appeared to be at least double in CPPs [22]. They also reported that the reviewed studies identified six risk factors or predictors of suicidality in CPPs: type/intensity/duration of pain, sleep onset insomnia, helplessness/hopelessness in reference to pain, desire to escape from pain, pain catastrophizing/avoidance, and problem-solving deficits [22].

Finally, in 2007, Spiegel et al. [29] reviewed eight studies on the association of chronic abdominal pain and suicidality. Chronic abdominal pain was found to be an independent predictor of suicidality after adjusting for comorbid psychiatric conditions [29].

The narrative structured review described below is an update on this literature and is necessary for a number of reasons. These are the following. Firstly, a significant amount of time has passed since the last review during which a significant number of studies on suicidality and pain have been added to this literature (specifically 58). Secondly, the previous literature is of itself was not convincing as to whether pain was actually associated with SCs in CPPs and non-CPPs, as SCs are extremely rare, resulting in few studies. Thirdly, the previous literature was not clear as to whether the association of pain and suicidality was relevant to different subgroups, such as veterans with pain, geriatric patients with pain, acute pain patients (APPs), etc. Fourthly, the previous literature had few studies where CPPs were compared with controls for suicidality. Fifthly, the previous literature had identified a limited number of risk factors (predictors) for suicidality in CPPs. More studies in this area are, however, needed as there is still limited information as to which patients are at greater risk for SC. Sixthly, the previous literature contained few studies where the information was gathered prospectively, and there were controls for psychiatric/behavioral issues that have a great impact on suicidality. Seventhly, the previous literature contained few studies on specific pains such as neuropathic pain. Eighthly, recent reports of opioid overdoses have had caused a reaction within the medical community, causing reduced access to pain treatment with opioids. This in turn may be causing an increase in suicidality in CPPs unable to get adequate care. Lastly, the study on the association between pain and suicidality has recently become a priority issue for the Veterans Administration.

The three previous reviews [20,22,29] of this literature were narrative reviews that did not utilize any systemic methods. The review described below, although a narrative nonsystemic review, used some systemic methods to gather data. In addition, it utilized a vote counting method to determine whether the reviewed studies supported/did not support the association of pain and suicidality. It then utilizes the Agency for Healthcare Research and Quality (AHRQ) guidelines [30] to rate the overall consistency of this evidence based on the percentage of studies supporting this potential association. To our knowledge, it is the first such review of the pain/suicidality literature.

Methods

Relevant references were located by the following procedure. MEDLINE, Embase, AMED, Psychological Abstracts, Science Citation Index, and the National Library of Medicine Physician Data Query databases were reviewed utilizing the following subject headings: pain, chronic pain, chronic pain patients, acute pain, low back pain, chronic low back pain, failed back surgery syndrome, soft tissue syndromes, fibromyalgia, fibromyalgia syndrome, musculoskeletal pain, chronic widespread pain, myofascial pain, and myofascial pain syndrome. Each of these was exploded with suicide, suicide completion, suicide attempts, and suicide ideation, and all retrieved references reviewed.

The searches were not restricted to the English language and conducted back to 1966, except for Science Citation index, which was conducted back to 1974. The upper limit of each search was March 2013.

A manual search was also performed using key pain journals, pain meeting abstracts, and textbooks. For the following journals, the following years were reviewed: Pain, 1975–March 2013; Spine. 1986–March 2013; The Pain Clinic, 1986–March 2013; Clinical Journal of Pain, 1985–March 2013; Pain Medicine, 2000–March 2013. Abstracts of the following meetings were reviewed for the following years: International Association for the Study of Pain 1981, 1984, 1987, 1990, 1993, 1996, 1999, 2002, 2005, 2008, 2011 and the American Pain Society Meetings 1982–2013.

One thousand six hundred and nineteen references were found and subjected to a cursory review. Studies were excluded from detailed review for the following reasons: 1) did not address acute or CP; 2) did not address SI, SA, or SC; and 3) the report was not a study, e.g., case reports. Previously reviewed studies and case reports were placed into a separate file. Studies were included for detailed review if they were not excluded by the above exclusion criteria. Of the original 1,619 references, 1,561 were excluded by this process or had been previously reviewed, leaving 58 studies that fulfilled these exclusion/inclusion criteria. These 58 studies [31–88] were reviewed in detail and sorted into seven lines of evidence: SI and pain by type of populations; SA and pain by type of population; SC and pain by type of populations; reported prevalence percentages of SI in CPPs and reported prevalent SI vs controls; reported prevalence percentages of SA in CPPs and reported prevalence of SA in CPPs vs controls; reported prevalence percentages of SC in CPPs and reported prevalence of SC in CPPs vs controls; and reported prevalence percentages of SI and SA in APPs vs controls.

Studies were also scrutinized for variables that were demonstrated to independently predict SI, SA, or SC in CPPs and APPs. Reported predictors were recorded.

Data from the above lines of evidence were then abstracted into seven Appendix tables (Tables A1–7) and labeled according to those lines of evidence. Whether each study did/did not support the association between pain and suicidality was placed into the table as a “Yes” vote.

The senior author independently abstracted the data into Tables A1–7. The studies were not rated for quality as this was not meant to be a systemic review. The senior author also categorized the type of evidence the studies represented in Tables A1–7. This was based on the guidelines developed by the AHRQ for categorizing the levels of evidence represented by the reviewed studies (Table A8, Subsection I) [30]. Studies were categorized I through V according to this scheme.

The percentage of studies supporting the association of pain and suicidality was than calculated for each table according to the Yes votes. The overall strength and consistency of the research evidence represented by the percentage of studies supporting the association between pain and suicidality (the Yes votes) (Tables A1–7 ) were then rated according to the AHRQ consistency of evidence guidelines (Table A8, Subsection II) [30]. These guidelines allow the researcher to categorize the reviewed evidence as being consistent, generally consistent, inconsistent, or demonstrating little or no evidence for the question associated with the studies in each table.

To render a consistency rating, one needs to have at least three studies that address the issue under study. As such, for this review, if there were not enough studies within the 58 studies that addressed the issue in question, studies from the previously reviewed literature were added in order to have a reliable consistency rating.

As a final step, data from Tables A1–7 were formatted into a summary table (Table ¹⁰⁰¹). This table was designed to summarize the overall findings of this narrative structured review. The headings for these tables are the following: table number, line of evidence, number of studies in this group, % of studies in the group with each type of evidence category, and strength and consistency of this evidence for the line of evidence addressed by the studies in that table according to AHRQ consistency guidelines. If it was necessary to tap previously reviewed studies, this was noted in the table.

Results

The results of this structured review from Tables A1–7 are summarized in Table ¹⁰⁰¹. These results were as follows:

For SI and pain by type of population (25 studies, Table A1), of the four types of studies (including those that were controlled for behavioral issues), three received an A consistency rating for an association between SI and pain. The fourth type of study group (prospective) did not have enough studies for a consistency rating, but all three studies in this group indicated an association. Of all studies combined (N = 25), 23 studies or 91.7% supported this association for an A consistency rating.

For SA and pain by type of population (eight studies, Table A2), of the four types of study groups, one received an A consistency rating for an association between SA and pain. The two other study types (prospective and Type 3) did not have enough studies for a consistency rating. Of all studies combined (N = 8), five studies or 87.5% supported this association for an A consistency rating.

For SC and pain by type of population (nine studies, Table A3), of the four types of study groups, three received an A consistency rating for an association between SC and pain. The fourth type of study group (prospective) did not have enough studies for a consistency rating, but of the two studies in this group, 100% supported an association. Additionally, previously reviewed literature contained two prospective studies that were added to the two prospective SC studies in this review. This gave a total of four studies, all of which indicated an association between pain and SC for an A consistency rating. Of all studies combined (N = 9), eight or 88.8% supported this association for an A consistency rating.

For reported prevalence of SI in CPPs and reported prevalence of SC in CPPs vs controls (10 comparisons, Table A4), 80.0% of the comparisons supported an association between CP and SI (consistency A). SI prevalence in CPPs ranged from 7.92% to 40.9%.

For the prevalence of SA in CPPs vs controls (Table A5), there were only three studies, 100% of which supported an association between CP and SA. The number of studies here was not enough for a consistency rating. However, the previously reviewed literature contained three studies [11,15,89] that had utilized a control group for SA. All three of these studies had demonstrated a greater prevalence of SA in CPPs vs controls. This made a total of six studies, 100% of which showed a statistically greater prevalence of SA within CPPs vs controls for an A consistency rating. History of SA ranges from 14.1% to 38% (different for men and women).

For the prevalence of SC in CPPs vs controls (Table A6), there was only one study with a comparison group, resulting in an unknown consistency rating. That study indicated an association between CP and SC. However, previously reviewed literature contained three studies [9,16,28] where CPPs had been compared for SC with control groups. All three of these studies demonstrated a greater prevalence of SC within CPPs. This gave a total of four studies, 100% of which demonstrated this association for an A consistency rating.

For the prevalence of SI in APPs vs controls (Table A7), 100% supported the association of SI and acute pain, leading to an A consistency rating. For the prevalence of SA in APPs vs controls (Table A7), there was only one study with one comparison, leading to an unknown consistency rating. However, that study indicated an association.

The reviewed studies indicated that the following variables were predictors for SI in CPPs. Demographic variables were age [85], sex [85], worker compensation status [36], and in litigation [36]. Pain-related variables were pain severity [85], pain duration [57,85], pain catastrophizing [85], degree of pain-related catastrophizing [88], and pain status [36]. Health-related variables were functional limitation [85], patient health questionnaire scores [85], disability perception [84], and smoking [53]. Psychiatric/behavioral variables were history of sexual/physical abuse [86], family history of depression [86], social withdrawal [86], Behavioral Health Inventory 2 (BHI-2) Borderline Scale scores [84], magnitude of depression [88], and the Interpersonal Relationship Scale scores [85]. Patient perception variables were treated unfairly by family [84], thoughts of revenge killing [54], motivated to seek revenge [54], self-perceived burdensomeness [52,85,87], and BHI-2 Doctor Dissatisfaction scale scores [54].

The reviewed studies also found the following variables to be predictors for SA in CPPs. One demographic variable was age greater than 35 [50]. Psychiatric variables were BHI-2 survival of violence scale scores [55], hearing voices others do not hear [55], lots of energy [55], and no misuse of alcohol (this last predictor appears to be counterintuitive.). For SA in fibromyalgia pain, the following psychiatric variables were predictive: poor sleep quality, anxiety, and depression [51].

The only predictor for SC in CPPs that was reported was that of the diagnosis of fibromyalgia [59], but for females only.

In APPs, the following variables were reported to be predictors of SI. Psychiatric variables were a depression item [55], a borderline item [55], a hostility item [55], anger [55], a family dysfunction item [55], and scores on the BHI-2 Family Dysfunction Scale [84]. Health-related variables were disability perception [84], scores on the BHI-2 somatic complaints scale [55], a somatic complaints item [55], and a muscular bracing item [55].

For SA in APPs, the following variables were reported to be predictors. Demographic variables were worker compensation status [36], history of personal injury [36], and in litigation [36]. Psychiatric variables were anger [55] and BHI-2 chronic maladjustment scale scores [55]. The only pain variable found to be predictive was pain range scores [55].

Discussion

A number of observations can be derived from the results in Table ¹⁰⁰¹, which divides the available studies by SI, SA, and SC. Firstly, the association between SI and pain was consistent (A) in “by type of population” and for greater prevalence percentages in CPPs vs controls (Table ¹⁰⁰¹). There was also a consistently (A) greater prevalence of SI within APPs vs controls (Table ¹⁰⁰¹).

Secondly, for SA, the association between SA and pain was also consistent (A) by “type of population.” For SA prevalence in CPPs vs controls, there were not enough studies for a consistency rating. Adding three studies that had previously been reviewed, however, gave an A consistency rating here. In APPs, for SA, there were not enough studies for a consistency rating.

Thirdly, for SC, the association between SC and pain was also consistent (A) in “by type of population.” However, for SC for CPPs vs controls, there were not enough studies for a consistency rating. Adding three studies that had previously been reviewed gave four studies, all of which demonstrated a greater prevalence of SC in CPPs vs controls for an A consistency rating.

Overall, combining the research reviewed here plus, if necessary, previously reviewed literature, there is strong consistent evidence for an association between SI, SA, and SC and pain in different types of populations and CPPs.

Lack of studies comparing CPPs with controls for various forms of suicidality was previously identified as a problem with this literature [22]. As demonstrated here, there are now more of these types of studies that are consistent in their results. However, these studies are still rare. There is an obvious reason for lack of studies in reference to SA and SC where CPPs are compared with non-pain controls. SA and SC are much rarer events than SI in both the general population and in CPPs, while SC is a much rarer event than SA. As such, studies for SA and SC are much more difficult to do than for SI. As most patients who proceed to SA and SC first demonstrate SI, the results here are nevertheless important for the association of suicidality and pain. However, risk factors that predict SI may not necessarily predict SA and/or SC. As such, more studies on the association of pain and SA and SC are needed.

It is to be noted that the data in Table ¹⁰⁰¹ for by type of population was subdivided into studies that were prospective and studies that were controlled for behavioral variables such as depression. This is because to really demonstrate an association between pain and suicidality, ideally one should do a prospective study to see if suicidality develops as a consequence of pain development controlling for various variables. Such a design controls for self-reporting errors, recall bias, and memory failure [22]. As pointed out by Tang and Crane [22], the studies available for their review did not control for psychiatric variables and were rarely prospective. In the new studies reviewed here for SI, SA, and SC by type of population, there were a significant number of studies that had controlled for behavioral issues. In each of these subgroups, an A consistency rating was generated. Thus, controlling for these variables that have a significant impact on suicidality, there is still an association between these types of suicidality and pain. In reference to prospective studies, there were not enough studies for SI and SA to reach a consistency rating, but enough studies for SC to do so. SC prospective studies were consistent in demonstrating that there is an association between SC and pain in by type of population.

It is also to be noted that some of the studies for SI, SA, and SC by type of population controlled for disability, social, and demographic factors. Failure to control for these variables has previously been pointed out as a major problem with this literature [22]. As noted here, this new literature has addressed this potential confounder.

Besides the large number of studies added to this literature since the last review, there is another positive aspect to this literature not discussed above. The studies for SI, SA, and SC by type of population addressed a number of different population groups, subgroups, and countries: US adults, older home meal recipients, patients with major depressive disorder from Germany, adolescents, terminally ill patients, cancer patients, Canadian adults, US veterans, Korean adults, burn patients, medical students in Taiwan, Australian primary care patients, patients from emergency rooms, adults from Hong Kong, Scottish cancer patients, depressed geriatrics, survivors of childhood cancer, human immunodeficiency virus patients, suicidal attempters in India, internal medicine outpatients, World Health survey group (14 countries), Finnish children, suicides in India, Japanese adults, and veterans on opioids. These studies also addressed various forms of pain: headache, pain in general, abdominal pain, etc. The vast majority of these studies demonstrated an association between pain and suicidality. This indicates that the association between pain and suicidality relates to various subgroups, countries, and cultures and may apply to different forms of pain.

Another interesting aspect to these new pain–suicidality studies are the reports on the association of suicidality and pain in APPs. If the association between suicidality and pain is true, then one should also be able to demonstrate such an association in APPs. The studies reviewed here indicate that such an association does exist for SI in APPs. This finding then indirectly supports the concept that there is an association between pain and forms of suicidality.

This new literature has also added a host of predictor variables for SI, SA, and SC in CPPs, which are presented under the results. A number of general observations can be derived from this list. Firstly, a predictor for one type of suicidality, e.g., SI does not necessarily predict other types. Secondly, because of the difficulty in performing SC studies, few predictors are identified here. Thirdly, a number of pain and depression variables are similarly identified as per previous reviews [22]. Fourthly, smoking is identified as one of the new predictors for SI. This is important because smoking has been consistently identified as a variable associated with suicidality in psychiatric patients [53]. As such, this finding also translates to CPPs. Fifthly, disability has long been known to be a predictor of suicidality in psychiatric patients [84]. It appears that this variable is also a predictor within CPPs for SI [84,85]. Sixthly, an interesting new variable found to be predictive for SI in CPPs is that of burdensome perception [52,85,87]. This variable dovetails in with another variable found to be predictive for SI: treated unfairly by family [84]. Finally, the variables of sexual/physical abuse have long been identified as being associated with suicidality in psychiatric patients [86]. Sexual/physical abuse also often leads to posttraumatic stress disorder, which itself is associated with suicidality [90] and CP conditions [91]. As noted, sexual/physical abuse was found to be a predictor for SI [86]. The above studies from the psychiatric literature then indirectly support this finding. However, the psychiatric literature is inconsistent as different authors have defined sexual/physical abuse in different ways, probably leading to inconsistent results [92]. Lastly, the variables of male sex, older age, functional impairment, poor relationship with family, family history of mental disorders, aggression, depression, anxiety, borderline syndrome, and hallucinations have all previously been identified in the psychiatric literature as being associated with increased risk of suicide [21] and were identified as predictor variables in the reviewed studies. Thus, the previous psychiatric literature indirectly supports the results of these studies.

What are the potential confounders to the results of this review? First, there is significant difficulty in making a distinction between true SA and a suicide gesture or non-suicidal self-injury, e.g., self-cutting [93]. It is unclear if the studies included in this review under SA were able to make this distinction or simply classified non-suicidal self-injury as SA. Second, many CPPs are treated with drugs for their pain, e.g., anticonvulsants. Suicidality has been reported as a consequence of antiepileptic drugs [94] and other drugs [95]. None of the reviewed studies controlled for this issue. Thus, this could have been a potential confounder to some of the CPP studies. Third, in reference to SC in CPPs, the one new study [58] related to fibromyalgia and did not deal with men but only women. Men have a suicide rate twice that of women. Thus, this study may have limited applicability in answering the SC and CP association question. Fourth, this was a structured narrative review, which used a “ballot box” or “vote counting” approach to count the studies that did/did not support the association of pain and suicidality. It then utilized the AHRQ guidelines to rate the consistency of the studies in supporting the pain–depression association. Although this approach raises this review above that of a simple narrative review, it does not meet criteria for an evidence-based systemic review and is not being presented as such. This is because any and all studies addressing this association were included without regard to the quality of the studies, which was not rated. As such, this issue is a significant potential confounder to the results of this review. Finally, the first author abstracted the data (as to whether the study demonstrated/did not demonstrate the association of pain and suicidality), which was not checked or verified by an independent reviewer. Although the results were checked twice, there could have been errors in the abstraction process leading to potential confounding.

The results of this review also point to some new directions for research in this area. First, as pointed out, more prospective studies for SA and SC are needed for CPPs, and these studies should control for psychiatric and behavioral variables. Second, as noted for predictor variables, there were no studies that appeared to actually look at whether psychiatric diagnoses are predictive for SI, SA, and SC in CPPs. It is well known that psychiatric diagnoses such as major depression, alcohol use disorders, substance use disorders, etc., are associated with increased risk of suicidality [21]. As such, future studies need to actually control for individual diagnoses. Third, although some studies reviewed here had looked at pain subgroups such as headache pain, abdominal pain, etc., no studies had looked at the neuropathic pain subgroup. This subgroup is important as there is some recent evidence that it may be commonly found in CPPs [96]. Its relationship to suicidality is currently unclear. Fourth, physical diagnoses, such as chronic obstructive pulmonary disease, lupus, malignancy, chronic renal failure, etc., have also been identified as being factors associated with suicidality. Future studies in this area should also attempt to control for these variables. Finally, to take this research to a new level, a neurobiological mechanism for pain and depression needs to be clarified. Currently, it is believed that the neurological circuiting for emotions and pain in the nervous system are closely associated [97], and this may be the reason for the association of pain and depression, i.e., overstimulation of one circuit affects the other. Recently, it has also been demonstrated with functional magnetic resonance imaging in major depression that there are regions of the brain with abnormally enhanced activation at baseline. These regions are responsive to painful stimuli, and this responsiveness can be decreased by the antidepressant duloxetine [98]. These findings would lead one to design a functional magnetic resonance imaging study comparing the responsiveness of these regions to painful stimuli in three groups: CPPs without depression; CPPs with depression, but without suicidality; and CPPs with depression and SI or SA. Such a design could potentially lead to better understanding of the underlying neurologic mechanism for the association of pain and suicidality and perhaps a model.

What is the target audience for this review? The target audiences for this review are pain physicians, family practice physicians, psychiatric physicians, and any and all clinicians that deal with pain patients. Each of these groups has their own unique problems in dealing with pain patients with this association, which relates to the nature of their specialty. Pain physicians focus on pain and routinely fail to address or ask about depression or suicidality. The most common presentation to family practice physicians is pain. As such, and similar to pain physicians, this group focuses on pain and fails to evaluate the pain patient for depression and suicidality. Psychiatric physicians in turn focus on psychiatric pathology and thereby identify depression in their patients, but rarely inquire about pain status and thereby may not relate any suicidality they identify to pain [99]. This last observation also applies to other behavior clinicians such as psychologists. Other clinicians in contact with pain patients, such as physical therapists, focus on pain and activation and rarely inquire about depression and suicidality. Overall, the problem is one of division of medical care according to specialty training with the evaluation approach geared to that specialty. This often results in failure to identify the pain suicidality association.

Finally, when should clinicians worry and what should they do? There are two possible approaches to this problem. One approach is to routinely measure depression and suicidality in any CPP with such tools as the Beck Depression Inventory at every visit. Here, if any suicidality is recorded, then that clinician could immediately refer the CPP for psychiatric/behavioral evaluation for treatment recommendations. The second approach would be to routinely monitor pain and depression with such tools as the Brief Pain Inventory at every CPP visit. High scores on pain or high scores on depression on the rating scales in this tool should alert the clinician to inquire about suicidality or refer the CPP for psychiatric/behavioral evaluation. Direct inquiries about suicidality can be performed by asking, “Because of the pain, have you felt that life is not worth living (passive suicidality)”? A positive response should result in referral for a psychiatric/behavioral evaluation.

Conclusions

Since the last review on the association of pain and suicidality, there have been 58 studies added to this literature. Within the limits of this narrative review, overall, these studies plus, when necessary, previous studies are consistent utilizing AHRQ criteria in indicating that there is an association between SI, SA, and SC and pain.

References

Appendix