The Prospective Association between Personality Traits and Persistent Pain and Opioid Medication Use

Abstract

Objective:

Pain and prescription opioid medication use are prevalent and a major source of psychological and physical health burden. This research examines whether Five Factor Model personality traits prospectively predict who will experience persistent pain and use prescription opioid medication over a 10-year follow- up.

Methods:

Participants (N=8,491) were drawn from the Health and Retirement Study. At baseline, participants reported on their personality and whether they were in pain. Every two years, participants reported on their pain and, at the most recent assessment, their current use of prescription opioid medication. Logistic regression was used to test whether personality was associated with persistent pain over the up to 10-year follow-up and whether it predicted who would be taking prescription opioid medication.

Results:

Neuroticism was associated with higher risk of persistent pain (OR=1.44, 95% CI=1.38– 1.51) and opioid medication use (OR=1.21, 95% CI= 1.14–1.29) over the follow-up. Extraversion was associated with lower risk of persistent pain (OR=0.83, 95% CI=.80-.87) and opioid medication use (OR=0.92, 95% CI=.86-.97). Similarly, Conscientiousness was associated with lower risk of persistent pain (OR=0.83, 95% CI=.79-.87) and opioid medication use (OR=0.91, 95% CI=.86-.97).

Conclusions:

The findings suggest that personality traits are one psychological characteristic that modulates the likelihood of persistent pain and opioid medication use.

Pain and treatment for pain have a renewed urgency given the current opioid epidemic in the United States.¹ It is estimated that approximately one-third of American adults used prescription opioid medication in 2015.² Among those who misused it, two-thirds attributed the misuse to managing physical pain². Pain is multi-determined, with biological, social, and psychological antecedents.³ Among the psychological factors often implicated in health more broadly, an individual’s characteristic ways of thinking, feeling, and behaving (i.e., their personality) shape perceptions of health and well-being;⁴ this association may extend to the experience of pain.

The Five Factor Model (FFM)⁵ defines personality traits along five broad dimensions: Neuroticism (the tendency to feel negative emotions and vulnerability to stress), Extraversion (the tendency to be outgoing, social, and active), Openness (the tendency to be creative and unconventional), Agreeableness (the tendency to be trusting and altruistic), and Conscientiousness (the tendency to be organized, disciplined, and responsible). Several studies have examined the association between these traits and pain in specific patient populations. In a sample of participants with fibromyalgia, for example, those who scored higher in Neuroticism and lower in Conscientiousness had more chronic pain.⁶ Likewise, these two traits (along with higher Openness) were associated with more pain during labor for childbirth.⁷ Patients with chronic foot and ankle pain scored higher in Neuroticism than matched controls, whereas there was no difference in Extraversion (the other three FFM traits were not measured).⁸ In contrast, among patients treated for cancer, higher Openness was associated with higher average pain and higher Extraversion associated with lower current pain.⁹ Personality traits have likewise been associated with pain experienced in more general populations. In a large population-based sample, higher Neuroticism and lower Conscientiousness were both associated with greater pain (the associations with the other three FFM traits were not reported).¹⁰ Further, compared to matched controls, individuals with chronic wide-spread pain had higher scores in Neuroticism and lower scores in Extraversion (the other three FFM traits were not measured).¹¹ These studies have been critical for establishing an association between personality and the experience of pain. Such studies, however, have been cross-sectional and/or retrospective and thus do not address whether personality prospectively predicts who will experience pain over time. Given that pain may also influence personality ratings when assessed concurrently in cross-sectional study designs, the prospective study design reduces the likelihood of this potential bias and provides a more stringent test of these associations.

Of the FFM traits, Neuroticism, Extraversion, and Conscientiousness are the most likely to have the strongest association with pain over time. Both Neuroticism and Conscientiousness, for example, are associated with a greater burden of disease,¹² including diseases that have a significant pain component, such as arthritis.¹³ As such, individuals higher in Neuroticism and lower in Conscientiousness may have greater likelihood of experiencing pain through more disease burden. Neuroticism and Extraversion may be related to greater likelihood of pain through their well-established association with emotion.¹⁴ In particular, state negative affect, such as depressive symptoms and anxiety, are robust predictors of future pain.¹⁵ Negative affect, for example, predicts increases in pain over time among patients with chronic pain.¹⁶ Neuroticism and Extraversion are the two emotional traits that have the strongest associations with negative affect.¹⁷ The frequent experience of negative affect is, in fact, one of the core defining components of Neuroticism.¹⁸ Individuals higher in Extraversion, in contrast, tend to experience less negative affect concurrently and over time.¹⁷ As such, state depressive symptoms and anxiety, two common measures of negative affect, may contribute to and explain the association between these two traits and pain. Openness and Agreeableness are not associated consistently with either disease burden or state negative affect and thus these traits are less likely to be associated with greater pain.

In contrast to personality and pain, less research has addressed whether personality is associated with the use of opioid medication to manage pain. Much of what is known about personality and opioid use comes from either samples of patients in treatment for opioid addiction or community samples that include individuals who use illicit opioids. In patient samples, for example, those being treated for opioid dependence scored higher in Neuroticism and lower in Extraversion, Agreeableness, and Conscientiousness than matched controls.¹⁹ Studies of personality and illicit substance use in the community have generally combined individuals who use heroin and/or cocaine (which is not an opioid) and compared them with non- users. These studies have found that individuals who use these substances score higher in Neuroticism and lower in Conscientiousness than individuals who do not use such substances.^{20, 21} Some studies have further found that individuals who use these illicit substances also score lower in Agreeableness;²¹ although not all find this relation.²⁰ These associations tend to be consistent with the broader literature on personality and substance use.²² Conspicuously missing from this literature is the relation between personality traits and prescription opioid use. Such information is critical for better understanding the psychological dispositions of who is most likely to use prescription opioids. Identifying prospective personality trait predictors will help inform risk models of opioid use and help improve patient care.

The purpose of the present research was to examine the relation between personality traits and persistent pain and opioid use over an up to 10-year follow-up in a population-based sample of older adults in the United States. We extend the literature on personality and pain in several ways. First, we test whether personality at baseline prospectively predicts who will experience persistent pain over 10 years. Based on the literature on personality and pain in specific patient populations,^6–9 we hypothesized that participants higher in Neuroticism or lower in Extraversion or Conscientiousness would be at greater risk of persistent pain. Second, because disease burden and negative affect may explain the relation between these traits and risk of persistent pain, we tested whether either disease burden or depressive symptoms and anxiety could account for the association between personality and persistent pain. We hypothesized the association between personality (higher Neuroticism and lower Extraversion and Conscientiousness) and persistent pain will be explained, in part, by these additional risk factors. Third, we test whether personality at baseline prospectively predicts who will be taking prescription opioids at the 10-year follow-up. Given the literature on personality and illicit opioid use,^19–21 we hypothesized that higher Neuroticism and lower Conscientiousness would also be associated with greater likelihood of use of prescription opioids for pain. We also test whether the associations between personality and persistent pain and prescription opioid use vary by age or sex because of age and sex differences in both personality²³ and pain.²⁴

Method

Participants and Procedure

Participants were drawn from the Health and Retirement Study (HRS). HRS is a longitudinal study of Americans aged 50 years and older and their spouses.²⁵ HRS data are publically available from here: http://hrsonline.isr.umich.edu/. Five factor model personality traits were first assessed on a random half of the HRS sample in 2006 as part of a leave-behind questionnaire; the other half of the sample first completed the personality scale in 2008. These two assessments were combined as the baseline sample. Pain was measured at every two-year assessment. Participants were first asked about opioid use at the 2016 assessment. Of the 14,189 participants who had complete baseline personality data in 2006/2008, 1,966 died before a follow-up assessment. The 3,732 participants who survived but did not have follow-up data in 2016 were older (d=.58; p<.001), more likely to be male (χ²=15.08, p<.001), a race other than white (χ²=7.12, p=.028), had fewer years of education (d=.16, p<.001), and scored higher in Neuroticism (d=.04, p=.05) and lower on Extraversion (d=.12, p<.001), Openness (d=.14, p<.001), Agreeableness (d=.07, p<.001), and Conscientiousness (d=.15, p<.001). A total of 8,491 participants reported on their personality and experience of pain in 2006 or 2008 and had the relevant follow-up data through 2016 to be included in the analyses. Participants with missing data on any of the study variables were not included in the analysis. Participants who completed the baseline assessment in 2006 were younger (d=.25, p<.01) and scored higher in Neuroticism (d=.06, p<.001) and were more likely to report persistent pain (χ²=23.11, p<.001) than participants in the 2008 baseline sample, perhaps due to the extra pain assessment compared to participants in the latter sample (i.e., the 2006 sample had 5 follow-ups while the 2008 sample had 4 follow-ups).

Measures

Personality.

The 2006 and 2008 leave-behind questionnaire included the Midlife Development Inventory (MIDI),²⁶ a measure of FFM personality traits. Participants rated 26 items that measured Neuroticism (e.g., moody; Cronbach’s alpha [alpha]=.70), Extraversion (e.g., talkative; alpha=.73), Openness (e.g., creative; alpha=.80), Agreeableness (e.g., helpful; alpha=.63), and Conscientiousness (e.g., organized; alpha=.66). Personality in the HRS is fairly stable, with retest correlations that range between .62 (Agreeableness and Conscientiousness) and .68 (Extraversion) over four years ²⁷. Items were rated on a scale that ranged from 1 (a lot) to 4 (not at all). Relevant items were reverse scored in the direction of the label of the trait, and the mean was taken across items for each trait and standardized to a mean of zero and standard deviation (SD) of one.

Persistent pain.

During the health module, participants were asked, “Are you often troubled with pain?” and responded yes or no to the item. This question was asked at every two- year assessment between the baseline (2006/2008) personality assessment and the most recent (2016) assessment. Thirty-three percent of participants reported pain at the 2006/2008 baseline assessment, 35% of participants reported pain at the 2010 assessment, 34% of participants reported pain at the 2012 assessment, 38% of participants reported pain at the 2014 assessment, and 40% of participants reported pain at the 2016 assessment. Consistent with research on pain in the HRS,²⁸ persistent pain was defined as reporting yes to this question on at least two adjacent assessments between the 2006/2008 baseline and the most recent follow- up (2016). Although single-item measures of pain have somewhat lower reliability than measures with multiple items, the validity is similar ²⁹, especially when aggregated.³⁰ For descriptive analyses, participants were classified into three pain categories: no pain reported at any assessment, infrequent pain (pain reported at least once but not on adjacent assessments), and persistent pain (pain reported on at least two adjacent assessments).

Opioid use.

At the 2016 HRS assessment, participants were asked for the first time to report on their recent prescription opioid use. Specifically, participants were asked about their use of a “class of pain medications, called “opioids”, [which] includes such things as Vicodin, OxyContin, codeine, morphine, or similar medications. In the past three months, have you taken any opioid pain medications?” Participants responded yes or no to this question.

Disease burden.

Disease burden was the sum of 7 significant conditions (hypertension, diabetes, cancer, lung disease, heart disease, stroke, arthritis) measured concurrent with baseline personality. Two other diseases (psychiatric, dementia) were reported but not included in the disease burden total.

Negative affect.

Two aspects of negative affect were assessed concurrent with baseline personality: depressive symptoms and anxiety. Depressive symptoms were measured as the sum of eight items, answered yes/no, from the Center for Epidemiologic Studies Depression scale (CESD).³¹ Anxiety in the last week was measured with five items from the Beck Anxiety Inventory³² that were rated on a scale from 1 (never) to 4 (most of the time). Sample size for negative affect was 8,200 due to missing data on depressive symptoms and anxiety.

Analytic Strategy

We used a t-test to descriptively examine mean differences in the traits by reported pain categorized into no pain, infrequent pain, and persistent pain. We then used logistic regression to examine the association between personality traits and persistent pain. We predicted persistent pain from each trait separately, adjusting for age, sex, race, and education, given the demographic differences in both the experience of pain²⁴ and personality.²³ As a sensitivity analysis, we repeated the analyses excluding individuals who reported pain at baseline to test whether personality prospectively predicted incident persistent pain among participants who were pain-free at baseline. Age and sex were tested as moderators of the relation between personality and persistent pain. We then repeated the analysis of personality on persistent pain including either disease burden or negative affect (depressive symptoms, anxiety) in the model to test whether these factors explained the association between personality and likelihood of persistent pain.

We also used logistic regression to predict prescription opioid use, adjusting for the same covariates, and tested age and sex as moderators of these associations. As a sensitivity analysis, we included persistent pain as an additional covariate for the prescription opioid analysis to test whether personality predicted opioid use adjusting for individual differences in pain. We tested for an interaction between personality and persistent pain to assess whether the traits either amplified or reduced the association between pain and risk of opioid use.

Results

Of the 8,491 participants with the requisite data, 40% (n=3,440) experienced persistent pain, and 14% (n=1,166) of participants reported taking opioid pain medication within the previous three months (Table 1). Descriptively, participants who reported persistent pain scored higher in Neuroticism (almost .5 SD) and lower in Extraversion, Openness, and Conscientiousness (about .2 SD) than participants who never reported pain; participants who reported infrequent pain across the study period also scored higher in Neuroticism and lower in Extraversion and Conscientiousness relative to those without experiences with pain (Table 1). A dose-response gradient for personality traits and prescription opioid use was evident across the 3 pain groups. Table 1 also indicates that women and individuals with lower education were more likely to report pain.

Table 1.

Descriptive Statistic for Study Variables and by Reported Pain

Variables	Full Sample	Pain
		No pain	Infrequent pain	Persistent pain
Age (years)	65.93 (8.54)	65.92 (8.60)	66.37 (8.58)	65.93 (8.54)
Sex (female)	61%	55%	60%*	67%*
Race (African American)	13%	13%	13%	13%
Race (other)	3%	3%	3%	3%
Education (years)	12.87 (3.06)	13.31 (2.96)	12.86 (3.09)*	12.50 (3.08)*
Personality1
Neuroticism	2.04 (.61)	1.90 (.56)	2.02 (.58)*	2.18 (.64)*
Extraversion	3.23 (.54)	3.29 (.52)	3.24 (.53)*	3.18 (.56)*
Openness	2.97 (.55)	3.01 (.53)	2.98 (.55)	2.94 (.56)*
Agreeableness	3.54 (.46)	3.53 (.46)	3.54 (.47)	3.56 (.46)
Conscientiousness	3.40 (.46)	3.45 (.44)	3.40 (.44)*	3.34 (.47)*
Disease burden	1.78 (1.24)	1.33 (1.10)	1.73 (1.19)*	2.20 (1.23)*
Depressive symptoms	1.23 (1.84)	.65 (1.25)	1.08 (1.66)*	1.82 (2.16)*
Anxiety	1.53 (.56)	1.38 (.45)	1.51 (.53)*	1.67 (.63)*
Persistent pain	40%	0%	0%	100%
Prescription Opioid Use	14%	4%	11%*	24%*

Note. N=8,491; n=3,016 for no pain, n=2,035 for infrequent pain, and n=3,440 for persistent pain. Values are either Means (standard deviations) or percentages.

¹For the logistic regressions, personality was standardized to have a mean of 0 and a standard deviation of 1. For descriptive purposed here, we report the untransformed means.

^*p<.05 for difference with no reported pain.

Personality and Persistent Pain

Personality was associated with persistent pain (Table 2). Independent of age, sex, race, and education, every SD higher score in Neuroticism was associated with an over 40% higher risk of persistent pain and every SD higher score in either Extraversion or Conscientiousness was associated with an approximately 20% lower risk of persistent pain. The associations were similar when the five traits were entered simultaneously into the analysis. Of note, the pattern was similar when the sample was limited to participants (n=5,718) who did not report pain at baseline (OR_Neuroticism=1.32, 95% CI=1.23–1.41; OR_Extraversion=.88, 95% CI=.83-.94; OR_{Conscientiousness}=.89, 95% CI=.83-.95). Openness was also associated modestly with lower risk of persistent pain in the full sample (Table 2) but not among participants who did not report pain at baseline (OR_Openness=.96, 95% CI=.90–1.03). There was little evidence that age and sex moderated these associations (Table 3). The association between Conscientiousness and persistent pain was slightly attenuated at older ages; there was a similar pattern for Openness. Sex did not moderate any of the associations.

Table 2.

Logistic Regression Predicting Persistent Pain from FFM Personality Traits

Trait	Model 1		Model 2		Model 3
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Neuroticism	1.44 (1.38–1.51)	<.001a	1.37 (1.31–1.44)	<.001a	1.15 (1.08–1.21)	<.001a
Extraversion	.83 (.80-.87)	<.001a	.86 (.82-.90)	<.001a	.94 (.90-.98)	.011
Openness	.95 (.91-.99)	.021	.96 (.92–1.01)	.13	1.03 (.98–1.08)	.28
Agreeableness	1.00 (.96–1.05)	.91	1.01 (.96–1.06)	.82	1.09 (1.04–1.14)	.001
Conscientiousness	.83 (.79-.87)	<.001a	.88 (.84-.92)	<.001a	.94 (.89-.98)	.011

Note. N=8,491. N=8,200 for Model 3 due to missing data on anxiety and depressive symptoms. Coefficients are odds ratios (95% confidence intervals) from logistic regressions when the personality traits are entered individually. Model 1 adjusts for age, sex, race, and education. Model 2 adjusts for Model 1 covariates and disease burden. Model 3 adjusts for Model 1 covariates and depressive symptoms and anxiety. OR=Odds Ratio. CI=Confidence Interval.

^aSignificant when sample is limited to participants who reported no pain at baseline (n=5,758).

Table 3.

Personality x Age and Personality x Sex Interactions Predicting Persistent Pain

Trait	Age Interactions		Sex Interactions
	OR (95% CI)	p	OR (95% CI)	p
Neuroticism	1.00 (.99–1.01)	.22	.98 (.89–1.08)	.70
Extraversion	1.01 (1.00–1.01)	.08	1.00 (.92–1.10)	.96
Openness	1.05 (1.01–1.10)	.030	1.00 (.92–1.10)	.95
Agreeableness	1.00 (.99–1.01)	.99	1.04 (.95–1.14)	.41
Conscientiousness	1.05 (1.01–1.10)	.021	1.00 (.92–1.10)	.97

Note. N=8,491. Coefficients are odds ratios (95% confidence intervals) for interactions between each trait and age and sex from logistic regressions when the personality traits are entered individually. CI=Confidence Interval.

We next addressed whether the association between personality and pain persisted after adjusting for shared risk factors. Disease burden was associated with persistent pain: For each additional disease, risk of persistent pain increased by 72% (OR_{Disease Burden}=1.72; 95% CI=1.65– 1.79). After adjustment for disease burden, Neuroticism, Extraversion, and Conscientiousness remained significant predictors of persistent pain (Model 2, Table 2). Disease burden accounted for 16% of the association between Neuroticism and persistent pain, 18% of the association between Extraversion and persistent pain, and 29% of the association between Conscientiousness and persistent pain. Both measures of negative affect were also associated with persistent pain: Every SD increase in depressive symptoms and anxiety was associated with a 56% and 24%, respectively, increased risk of persistent pain (OR_{Depressive Symptoms}=1.56, 95% CI=1.48–1.65 and OR_Anxiety=1.24, 95% CI=1.18–1.31). Including these two aspects of negative affect in the model accounted for 66% of the association between Neuroticism and persistent pain, 65% of the association between Extraversion and persistent pain, and 35% of the association between Conscientiousness and persistent pain. These traits remained significant predictors of persistent pain (Model 3, Table 2). When both disease burden and negative affect were entered into the model simultaneously, the association between Conscientiousness and persistent pain was reduced to non-significance (OR=.97, 95% CI=.92–1.02); Neuroticism (OR=1.14, 95% CI=1.08– 1.21) and Extraversion (OR=.95, 95% CI=.90-.99) remained significant.

Personality and Prescription Opioid Use

Personality was also associated with prescription opioid medication use (Table 4): Neuroticism was associated with higher risk of using opioid medications in the last three months, whereas Extraversion and Conscientiousness were associated with lower risk of use. Agreeableness was also associated with slightly higher risk of opioid use. Similar to persistent pain, there was little evidence that age or sex moderated these associations (Table 5): Men who scored higher in Openness were more likely to use opioids, whereas there was no association between Openness and opioid use among women. Age did not moderate any of the associations. The association between Neuroticism and opioid use and Agreeableness and opioid use remained significant when adjusting for persistent pain; the other two associations were reduced to non- significance (Model 2, Table 4). Persistent pain accounted for 62% of the association between Neuroticism and Extraversion and opioid use and 56% of the association between Conscientiousness and opioid use; persistent pain did not account for the association between Agreeableness and opioid use. Finally, none of the interactions between personality and persistent pain on opioid use was significant, which indicated that personality did not amplify or reduce the relation between persistent pain and opioid use (Table 6).

Table 4.

Logistic Regression Predicting Prescription Opioid Use from FFM Personality Traits

Trait	Model 1		Model 2
	OR (95% CI)	p	OR (95% CI)	p
Neuroticism	1.21 (1.14–1.29)	<.001	1.08 (1.01–1.15)	.020
Extraversion	.92 (.86-.97)	.005	.97 (.91–1.04)	.38
Openness	1.01 (.94–1.08)	.80	1.03 (.96–1.10)	.41
Agreeableness	1.08 (1.01–1.15)	.030	1.08 (1.01–1.16)	.029
Conscientiousness	.91 (.86-.97)	.003	.96 (.90–1.03)	.28

Note. N=8,491. Coefficients are odds ratios (95% confidence intervals) from logistic regressions when the personality traits are entered individually. Model 1 adjusts for age, sex, race, and education. Model 2 adjusts for Model 1 covariates and persistent pain. OR=Odds Ratio. CI=Confidence Interval.

Table 5.

Personality x Age and Personality x Sex Interactions Predicting Prescription Opioid Use

Trait	Age Interactions		Sex Interactions
	OR (95% CI)	p	OR (95% CI)	p
Neuroticism	1.00 (.99–1.01)	.87	.96 (.84–1.08)	.49
Extraversion	1.00 (.99–1.01)	.99	.94 (.83–1.07)	.35
Openness	1.00 (.94–1.06)	.98	.85 (.74-.97)	.014
Agreeableness	1.00 (.99–1.01)	.77	.95 (.83–1.08)	.43
Conscientiousness	1.05 (.99–1.11)	.11	.93 (.82–1.05)	.26

Note. N=8,491. Coefficients are odds ratios (95% confidence intervals) for interactions between each trait and age and sex from logistic regressions when the personality traits are entered individually. CI=Confidence Interval.

Table 6.

Personality x Persistent Pain Interactions Predicting Prescription Opioid Use

Trait	Persistent Pain Interactions
	OR (95% CI)	p
Neuroticism	1.10 (.97–1.24)	.14
Extraversion	.95 (.84–1.07)	.40
Openness	1.05 (.92–1.19)	.48
Agreeableness	1.03 (.90–1.17)	.68
Conscientiousness	1.00 (.88–1.13)	.97

Note. N=8,491. Coefficients are odds ratios (95% confidence intervals) for interactions between each trait and persistent pain from logistic regressions when the personality traits are entered individually. CI=Confidence Interval.

Discussion

The present research used a large sample of middle aged and older adults and a long follow-up period to examine the prospective relation between personality and persistent pain and prescription opioid use. Higher Neuroticism and lower Extraversion and Conscientiousness were prospectively associated with persistent pain and prescription opioid use up to 10 years from the personality assessment. Higher Neuroticism and lower Conscientiousness are routinely associated with greater pain in patient populations^{6, 7} and there is growing evidence for the association in non-clinical samples.¹⁰ There is also some evidence that Extraversion is related to less pain,^{9, 11} although not all find this relation.⁸ The present research adds that personality prospectively predicts over 10 years who will experience persistent pain in a population-based sample of older adults, even among individuals who did not report pain at baseline.

There are both physical and psychological factors that may contribute to the relation between personality and pain. Individuals who suffer from a chronic burden of disease, for example, also tend to suffer from greater pain.³³ Negative affect is likewise a risk factor for greater pain over time.^{15, 34} Individuals higher in Neuroticism and lower in Conscientiousness tend to carry a greater burden of disease^{12, 13} and these two traits and lower Extraversion tend to also be associated with greater negative affect.¹⁷ As such, the association between these traits and persistent pain may be partly due to comorbid disease and negative affect. Adjusting for disease burden or depressive symptoms and anxiety did not account completely for the associations between personality and persistent pain. Disease burden and negative affect, however, could be considered more than confounding variables. That is, they might be mechanisms through which personality contributes to persistent pain over time. This pattern further suggests that even if disease burden and negative affect are either confounding variables or mechanisms of the associations, there are other mechanisms between personality and pain.

There are at least two other pathways through which personality traits may contribute to greater risk of persistent pain. The first pathway is through enhanced sensitivity to pain. Individuals report different perceptions of pain even when the stimulus is held constant.³⁵ Those higher in Neuroticism may have a greater sensitivity for feeling pain that leaves them more vulnerable to it. Experimental studies, for example, suggest that Neuroticism contributes to how pain is experienced. When induced to feel pain in the lab, individuals higher in Neuroticism reported more pain in response to the same stimulus than participants lower in Neuroticism.³⁶ In another experiment of laboratory-induced pain, Neuroticism was related to worse ratings of pain one week after the exposure.³⁷ This finding suggests that the experience of pain does not diminish with time for individuals high in Neuroticism. In these experimental studies, Extraversion tends to be unrelated to pain sensitivity: Extraversion was unrelated to pain rating in response to a stimulus³⁶ and to memory of laboratory-induced pain.³⁷ The association between Extraversion and pain may thus not be through greater sensitivity. And, since experimental studies have focused exclusively on Neuroticism and Extraversion, it is yet unknown whether pain sensitivity varies by level of Conscientiousness.

A second pathway in which personality may be associated with pain is through health-related behaviors. In contrast to research on pain sensitivity, there is considerable evidence that Conscientiousness and Extraversion, as well as Neuroticism, are associated with health-risk behaviors that increase risk of pain. For example, higher Neuroticism and lower Extraversion and Conscientiousness are associated with physical inactivity³⁸ and poor sleep patterns.³⁹ Neuroticism and Conscientiousness have further been implicated in smoking⁴⁰ and use of other substances²² that may also aggravate pain. Such behavioral factors, including lack of physical activity, poor sleep, and smoking and substance use, have been implicated in risk of pain.⁴¹

Personality and opioid use has been examined primarily in the context of illicit use and treatment for abuse and primarily among younger adults. In these studies, which often combine the use of opiates with other illicit substances (e.g., cocaine), higher Neuroticism and lower Conscientiousness are associated with greater use.^{20, 21} Further, these traits plus low Extraversion have been associated with substance use disorders.²² The findings from the current study are consistent with this literature: Participants higher in Neuroticism and lower in Extraversion and Conscientiousness were more likely to have used prescription opioid medication in the last three months. Even when adjusting for persistent pain, Neuroticism remained a significant predictor of opioid medication use. It is not surprising that the other traits were not significant predictors because inclusion of persistent pain as a covariate should substantially reduce associations with other variables (i.e., pain tends to be the reason why opioids are prescribed). As described above, the experimental studies on Neuroticism and pain indicate that individuals higher on Neuroticism have a greater sensitivity to pain that may lead them to pain medication at lower thresholds than individuals lower on this trait. A second, non-mutually-exclusive possibility is that individuals who score higher on Neuroticism may be more open or vulnerable (e.g., because of addiction) to receiving opioids to treat their pain. As such, the greater pain associated with lower Extraversion and Conscientiousness may be the predominant pathway through which these traits are associated with prescription opioid use. In contrast, for Neuroticism, there may be multiple pathways, including through greater pain, greater perceived need for taking prescription medication, and/or greater likelihood of being prescribed such a medication by a physician. These pathways are not mutually exclusive.

The present study had several strengths, including a large sample, a measure of all five major personality traits, a relatively long prospective assessment of pain and opioid use, and the focus on older adults, which complements past research on younger age groups. Limitations include (1) lack of specificity in the cause (e.g., back pain versus pain from arthritis) and course of pain, (2) the phrasing of the item assessing pain, (3) self-reported assessment of opioid prescription medication use instead of objective measures, and (4) the lack of information on prescription opioid before the 2016 assessment that does not allow for adjusting for previous opioid use. Another limitation is attrition from baseline to the 10-year follow-up, with a selection of participants with a more favorable trait profile. This attrition, however, may have led to an underestimate of the association between personality and pain/opioid use since the participants with the traits most associated with the outcomes were less likely to remain in the sample. Finally, we did not adjust the p-value for multiple comparisons. Although adjusting the p-value helps control for Type I errors, such adjustments can be overly conservative and inflate Type II errors. We note that some associations would not be significant if the p-value was adjusted (e.g., the association between Openness and persistent pain). Other associations, such as the relation between Neuroticism, Extraversion, and Conscientiousness and persistent pain, are consistent with the broader literature on personality and health and are less likely to be due to chance. Still, this statistical choice should be considered when evaluating the results.

These limitations suggest opportunities for future research. First, in future research it would be helpful to have a more detailed pain assessment to know more about the timing, location, and nature of the pain and whether these characteristics have differential associations with personality traits. Second, it is critical to have objective measures of opioid medication use, perhaps through claims data. Third, it would be useful to also have a measure of illicit opioid use to examine potential differences across legal and illicit use. Fourth, these associations should be examined in specific clinical populations, such as patients with arthritis or back pain. Such replication would also increase confidence that the associations are robust and not due to chance.

There are potential translational implications for clinical populations. This research helps to identify who is most at risk. In particular, it indicates that individual differences in psychological dispositions are associated with pain and prescription opioid use. Such information can be used in risk predictor models that incorporate multiple variables to determine risk. This information may also be useful for interventions. In addition to identifying who is most at risk and may benefit most from an intervention, interventions themselves may be based on the individual’s personality. Although not yet tested in the context of pain and prescription opioid use, personality-tailored interventions have been successful in other domains.^{42, 43} Interventions have also recently been found to be able to change maladaptive aspects of personality, including those found in the current study to increase risk of pain and prescription opioid use.⁴⁴ One downstream benefit of such interventions may be better pain and drug-use related outcomes. Despite the limitations and opportunities of future research, the present research indicates that personality is implicated in pain over time. Similar to genetics,⁴⁵ stable psychological functioning may be one factor that makes an individual more vulnerable to persistent pain and may be useful in developing targeted prevention and treatments programs to manage pain.

Highlights.

• Higher Neuroticism and lower Extraversion and Openness are associated with persistent pain

• These traits also predict use of prescription opioid medication

• These associations were apparent over a 10-year follow-up period

• Personality traits have a prospective association with who will experience pain

Abbreviations:

SD

standard deviation

OR

odds ratio

CI

confidence interval