Chronic Pain Among Suicide Decedents, 2003 to 2014: Findings From the National Violent Death Reporting System.

We analyzed data from 18 states participating in the National Violent Death Reporting System (NVDRS) for at least 1 year from 1 January 2003 to 31 December 2014. The NVDRS, described elsewhere ( 15 ), is an active, state-based surveillance system that links data on violent deaths (including suicides) from death certificates, coroner or medical examiner reports, and law enforcement reports into a single incident record ( 15 ). Information is abstracted by using standard coding guidance developed by the Centers for Disease Control and Prevention (CDC). The NVDRS defines suicide as a death resulting from the intentional use of force against oneself, classified by the International Classification of Diseases, 10th Revision, cause-of-death codes X60 to X84, Y87.0, and U03 ( 16 ). Statewide NVDRS data have been collected in Maryland, Massachusetts, New Jersey, Oregon, South Carolina, and Virginia since 2003; Alaska, Colorado, Georgia, North Carolina, Oklahoma, Rhode Island, and Wisconsin since 2004; Kentucky, New Mexico, and Utah since 2005; Ohio since 2011; and Michigan since 2014. More than 600 standard variables are coded in the NVDRS, including circumstances identified as directly contributing to the death (such as a physical health problem). Data on these precipitating circumstances often originate from investigators’ interviews with informants who knew the decedent. In addition, the NVDRS abstractors enter separate narratives based on reports from both coroners or medical examiners and law enforcement officers to summarize the events of the fatal incident and other pertinent information, including health conditions affecting the decedent. We included decedents aged 10 years or older who died by suicide during 2003 to 2014 in this analysis. We identified cases with evidence of chronic pain by using keyword searches for 120 medical conditions and 9 pain types (allowing for common misspellings) in the coroner or medical examiner and law enforcement narratives. We selected these keywords on the basis of medical conditions and pain types associated with chronic pain from the American Chronic Pain Association (ACPA) ( 17 ). We selected cases 3 ways ( Figure 1 ). First, we selected cases that reported at least 1 medical condition primarily associated with chronic pain (such as “fibromyalgia”), pain classified by anatomical location (such as “back pain”) or organ system (such as “musculoskeletal pain”), or pain type consistent with chronic pain (such as “persistent pain”) (group 1). Second, we selected cases with medical conditions frequently associated with chronic pain, but not necessarily as the primary symptom (such as “sickle cell”), if the keyword “pain” (excluding “emotional” and “acute” pain) also was present and a contributing physical health problem was endorsed (group 2). Third, we selected cases with the keyword “pain” (excluding “emotional” and “acute” pain) and a contributing physical health problem (group 3) to maximize the capture of decedents with chronic pain. “Contributing physical health problem” is a standard NVDRS variable indicating that any physical health condition, such as terminal disease, debilitating condition, or chronic pain, was determined to be a contributing factor. This variable alone is too broad for defining cases of chronic pain but was used in groups 2 and 3 to increase specificity. We separately categorized pain by anatomical structure and medical condition by organ system by using a pain taxonomy prepared by the American Pain Society ( 18 ). Decedents may have had more than 1 medical condition in more than 1 pain category. We reviewed the narratives of randomly selected decedents ( n = 100 each from groups 1, 2, and 3) by using an iterative approach to further refine and validate our keyword search. Criteria for chronic pain cases were met if the decedent had any of the ACPA medical conditions or pain types described earlier, or if the narrative specifically indicated chronic pain. Cases were excluded if the pain duration was documented as less than 3 months, the described pain referred exclusively to emotional pain, or the condition or pain type described applied to someone other than the decedent. Rater pairs conducted a final review of 216 randomly selected cases (2.0%). Interrater agreement ranged from 92.6% to 96.2% (κ range, 0.63 to 0.65); discrepancies were discussed and coded to consensus. In this analysis, 199 decedents (92.1%) were true cases of chronic pain. We also reviewed narratives of a random sample of 200 decedents with chronic pain who left suicide notes to determine whether the notes offered insights regarding the role of pain in precipitating suicide. Pain was determined to be a contributing factor if the suicide note included documentation that the decedent’s decision to die by suicide was because of pain. (For coding guidance used for the narrative review of suicide notes, see Appendix Table 1 , available at Annals.org .) We assessed the percentage of suicide decedents with chronic pain over time as well as the subset who died of prescription or illicit opioid overdose. We examined suicide decedents with and without chronic pain by demographic characteristics, mechanism of death, toxicology results, and precipitating circumstances. We examined the percentage of suicide decedents with evidence of chronic pain over time and used descriptive statistics to examine characteristics of suicide decedents with and without chronic pain across 12 data years (2003 to 2014). We used U.S. Census population data to calculate crude suicide rates per 100 000 persons by age group. Because of small sample sizes during 2003 to 2004, resulting in unstable rates, all reported rates represent annualized averages across 10 data years (2005 to 2014). We used SAS, version 9.3 (SAS Institute), for all analyses. This study used data collected as part of routine injury surveillance and was not funded.

During 2003 to 2014, the NVDRS identified 123 181 suicide decedents aged 10 years or older, 10 789 (8.8%) of whom had evidence of chronic pain. The percentage of decedents with chronic pain increased from 7.4% in 2003 to 10.2% in 2014, but the percentage who died by opioid overdose remained low overall (<2.0%) ( Figure 2 ). Table 1 presents 7 pain categories and the 3 most common conditions within each category among suicide decedents with chronic pain. (For a full list of conditions and categories, see the Appendix and Appendix Tables 2 to 10 , available at Annals.org .) The most common categories were spine pain (24.4%) and musculoskeletal pain (20.8%). The most common conditions were back pain (22.6%), cancer (12.5%), and arthritis (7.9%). More than half (54.4%) of decedents with chronic pain had 1 medical condition, 15.7% had 2 conditions, and 5.7% had 3 or more conditions; the remaining decedents (24.2%) did not have a medical condition noted but had a qualifying keyword in the narrative. Table 2 presents characteristics of suicide decedents. Those with chronic pain were older on average, and suicide rates increased with age and were highest among those aged 80 years or older. Nearly 1 in 3 suicide decedents (31.9%) with chronic pain were female, as were 21% of those without chronic pain. Of decedents both with and without chronic pain, most were non-Hispanic white (91.7% and 83.6%, respectively). More than a quarter of decedents (27.1%) with chronic pain had ever served in the military. The most common mechanism of death was firearm among decedents with chronic pain (53.6%) and those without it (51.4%), followed by opioid overdose (16.2%), other poisoning (13.3%), and hanging or suffocation (12.9%) among decedents with chronic pain, versus hanging or suffocation (25.6%), other poisoning (11.8%), and opioid overdose (3.9%) among those without chronic pain. Of decedents who had drug testing, 51.9% of those with chronic pain, but only 18.8% of those without it, had results positive for opioids; nearly half (47.2%) of those with and 31.7% of those without chronic pain had results positive for benzodiazepines. More than half (51.7%) of suicide decedents with and 44.1% of those without chronic pain had a known mental health condition, with depression being the most common diagnosis in both groups. Most suicide decedents were not known to have problematic substance use. Approximately 20% of decedents with chronic pain and 31.2% of those without it had suspected alcohol use at the time of death. Abuse of substances other than alcohol was noted in 17.6% of decedents with and 14.8% of those without chronic pain. A subset of decedents with chronic pain also were known to have life stressors related to job (10.5%) or financial problems (11.6%) or to interpersonal issues, such as intimate partner problems (16.9%) or an argument preceding the death (7.2%). However, 31.9% of decedents without chronic pain had intimate partner problems, and 12.0% had an argument preceding the death. Nearly one third of decedents with and without chronic pain had recent crises (27.7% and 29.3%, respectively). Many decedents with and without chronic pain left suicide notes (41.4% and 32.8%, respectively), had a history of suicidal thoughts or plans (39.5% and 31.1%, respectively), had a history of suicide attempts (21.1% and 19.8%, respectively), or disclosed their intent to die by suicide (33.0% and 27.4%, respectively). In a random sample of 200 decedents with chronic pain who left suicide notes, approximately half (47.5%) of the narratives reported the contents of the notes; among these, 67.4% of decedents indicated that a pain condition or pain itself played a role in their decision to die by suicide. In the remaining notes, pain was not mentioned or the extent to which pain was a precipitating factor for the suicide was unclear.

We found that nearly 9% of suicide decedents in 18 states from 2003 to 2014 had documentation of chronic pain in their incident records and that the percentage of decedents with chronic pain increased during the study period. Our results highlight the importance of pain in quality of life and premature death, and contribute to the growing body of evidence indicating that chronic pain might be an important risk factor for suicide ( 11 – 14 ). We identified a wide variety of medical conditions and pain types among decedents, with back pain, cancer, and arthritis being the most common. Many chronic pain conditions have been associated with suicidality ( 9 , 19 – 23 ), but the strength of association might differ by the specific pain condition ( 8 ). We did not stratify by medical condition or pain type when describing the characteristics of decedents, but future research might use the NVDRS to learn more about the pain conditions most strongly associated with suicide. Establishing the causal role of pain in suicide was beyond our study’s scope. However, our results were consistent with the known epidemiology of chronic pain in terms of temporal trends, age, and sex ( 24 ). Other factors known to contribute to suicide risk, including interpersonal problems ( 25 ) and life stressors ( 26 ), were less frequent among decedents with chronic pain, which might indicate that chronic pain was the dominant stressor in this group. Also, although we cannot draw definitive conclusions regarding the proportion of suicides directly attributable to chronic pain, our narrative review of suicide notes suggests that the proportion was not trivial. The proportion of suicide decedents with chronic pain increased during the study period. Chronic pain has increased in the general population by a similar magnitude ( 24 ), which may partially explain this finding. In general, suicide rates are higher among males and peak in middle age ( 5 ). However, suicide rates among decedents with chronic pain increased by age group and remained high among older persons. Health problems late in life are particularly associated with elevated suicide risk ( 27 ), and the prevalence of chronic pain, especially severe chronic pain, increases with age ( 24 ). Overall, we found considerably more suicides among males than females, regardless of chronic pain status, in agreement with the suicide literature ( 5 ). However, the difference was smaller among decedents with chronic pain, consistent with reports that chronic pain is more common in females ( 24 ). Although opioid prescribing to treat chronic pain has increased in recent years ( 28 ), we found that the percentage of decedents with chronic pain who died by opioid overdose did not change over time. This finding suggests that increases in opioid availability are not associated with greater suicide risk from opioid overdose among patients with chronic pain. Additional research is needed to explore factors other than opioid availability that might be associated with suicide among persons living with chronic pain. Mental health disorders, particularly depression, were prevalent among decedents with chronic pain, consistent with previous findings ( 21 , 29 ). Chronic pain and depression have complex and bidirectional associations: Depression is a risk factor for chronic pain ( 30 ), and chronic pain, in turn, is a risk factor for depression ( 31 , 32 ). Disentangling this complexity was not possible in our analysis. However, providers should be alert to and possibly screen for depression and suicidal behaviors among patients with chronic pain ( 29 ), especially older patients ( 27 , 29 ). Indeed, a history of suicidal thoughts, plans, and attempts and disclosure of suicidal intent were more common among decedents with chronic pain than those without it, indicating that opportunities for intervention may have been available. A recent study suggests that among patients with chronic pain, a belief in a medical cure for their pain may act as a protective factor against suicidal ideation ( 33 ). This finding underscores the importance of access to quality, comprehensive pain care and adherence to best practices to improve both pain management and patient safety. The “CDC Guideline for Prescribing Opioids for Chronic Pain,” released in 2016, provides recommendations for prescribing opioid pain medication for patients aged 18 years or older in primary care settings ( 34 ). The recommendations, based on a systematic review of the existing scientific evidence, focus on the use of opioids in treating chronic pain outside active cancer treatment, palliative care, and end-of-life care ( 34 ). The guideline’s goal is to ensure that patients receive appropriate treatment for pain while considering the risks and benefits of treatment options ( 34 ). Integrated pain management and coordination of primary care, mental health care, and specialist services are essential for these patients. Other guidelines recommend behavioral health consultation for any patient with a history of suicide attempt or psychiatric disorder and advise against initiating opioid therapy during acute psychiatric instability or uncontrolled suicide risk ( 35 ). In addition, the CDC’s suicide prevention technical package describes several evidence-based prevention strategies relevant to patients with chronic pain: strengthening economic supports (such as disability benefit programs), strengthening access to and delivery of suicide care (such as mental health insurance coverage), and creating protective environments (such as reducing access to lethal means) ( 36 ). More than half of all suicide decedents, with and without chronic pain, died by firearm, whereas approximately 16% of those with chronic pain died by opioid overdose. Safe storage of firearms might reduce the risk for suicide in this population by separating persons at risk for suicide from easy access to lethal means ( 36 ). Safe storage practices may include education and counseling for storing firearms locked in a secure place, unloaded, and separated from ammunition for persons who may be at risk for suicide ( 36 ). This analysis had limitations. First, data were obtained from 18 states and are not nationally representative, and the states included in the study varied over time. Second, our results probably underrepresent the true percentage of suicide decedents who had chronic pain, given the nature of the data and how they were captured. The reliability and validity of the medical diagnoses endorsed in the investigative reports also are unknown. Historically, patient reports and self-reported survey data have been the primary sources for pain estimates ( 1 ). Third, our case selection method led to relatively few false-positive results in our narrative review of selected cases, but a small percentage may have been misclassified, especially if pain was not a prominent issue reported by informants who knew the decedent. Fourth, we did not have information to assess pain characteristics, such as intensity and duration, or to determine medical treatment, and we could not distinguish prescription from illicit opioids. Finally, we could not assess other factors commonly associated with chronic pain that affect quality of life, such as disability, sleep disturbance, and in the case of terminal illness, the mental and physical effects of coping with end-of-life concerns. Chronic pain is a public health problem that has been garnering increased attention since the 2016 publication of “National Pain Strategy: A Comprehensive Population Health-Level Strategy for Pain” ( 4 ). Chronic pain also has implications for health care. Providers caring for patients with chronic pain should be aware of the potential increased risk for suicide, and more effort may be needed to diagnose, manage, and treat chronic pain and comorbid mental health conditions. Continued surveillance and research are needed to better understand the burden of chronic pain in the United States.