Contextual risk among adolescents receiving opioid prescriptions for acute pain in pediatric ambulatory care settings

Abstract

Many adults with opioid use disorder (OUD) report that their first exposure to opioids was in the course of routine pain treatment in medical care settings. Adolescents receive opioid prescriptions with frequency, but are susceptible to a constellation of unique risks in the context of pain management. This empirical study presents the first cohort of adolescents recruited from ambulatory medical care within 72 hours of their receipt of opioids to treat acute pain. The primary aim was to capture a time-sensitive report of the intersection of prescription opioid receipt and contextual risks for opioid misuse related to pain experience, mental health symptoms, and substance use at the adolescent and parental levels. Data were collected from 70 14–19-year-old adolescents and their accompanying parent. Results reflected that 90% of this sample of adolescents reported 2 or more risks and 35% reported 5 or more risks for future opioid misuse. Pain catastrophizing (46%) and alcohol use (40%) and were the most common adolescent-level risk factors; mother history of chronic pain (32%) and parent anxiety (21%) were the most common parent-level risk factors. Past-week parent pain intensity showed the strongest association with adolescent past-week pain intensity; neither was associated with adolescent OUD symptoms. Adolescent pain catastrophizing most reliably predicted OUD symptoms; parent pain interference was also associated. Seventy-one percent of parents reported keeping opioids at home, a relevant risk factor for future misuse. These findings illuminate the intersection between adolescent and parental risks in the context of pediatric opioid prescribing for acute pain management, and provide initial insight into potential points of prevention early in adolescent pain treatment, including avenues by which to inform and enhance prescriber decision-making regarding factors to be weighed in adolescent candidacy for opioid therapy.

1. Introduction

Adolescent opioid use remains an issue of critical public health concern. A growing body of literature highlights the risks and morbidity associated with adolescent opioid prescribing, as well as the serious and sustained impact of the United States’ opioid epidemic on this age group (1). The pediatric medical community is thus faced with the complicated responsibility of balancing adequate pain management with risks conferred by opioid exposure at a relatively early age (2, 3). Increasing evidence suggests that use of opioids during adolescence, even as prescribed by a medical professional to treat pain, is associated with later opioid misuse and transition into opioid use disorder (OUD; 4), highlighting the critical nature of understanding adolescent risk at the time of opioid prescription. Further, adolescents reporting high-severity pain are at greater risk for sustained opioid use after initial prescription regardless of injury severity (5). In adults, on-label use periods as short as 5 days increase risk for continued opioid use (6). While long-term opioid use often begins with opioids prescribed to treat acute pain (6, 7), we could find no published empirical studies that have captured a comprehensive assessment of adolescent risk for future opioid misuse at the time of prescription in routine outpatient pediatric ambulatory care settings.

Because legitimate medical use of prescription opioids is often the first introduction to a substance of abuse for many otherwise non-substance-seeking adolescents (4, 8), this pharmacological approach to pediatric pain management must be considered within the framework of adolescent substance use models in addition to pain management guidelines. This perspective requires examination of individual risk factors, the role of parents, and the familial environment (9, 10). Factors at both the individual and parental levels can confer risk for adolescent substance use, particularly in a context in which barriers to accessing substances (e.g., being under 21, high parental monitoring, values against alcohol and drug use) are circumvented by the medical system providing treatment. As such, it is critical to identify what individual and familial factors could play a role in risky opioid-related trajectories.

A theoretical model of individual- and parent-level risk factors for adverse pain outcomes and persistent opioid misuse in the context of pain experience has posited a number of specific risks (see Figure 1; 10). At the individual level, adolescent risk factors include pain-related cognitions and behaviors (distress intolerance, pain catastrophizing, and low pain self-efficacy); mental health symptoms (anxiety and depression); and alcohol, tobacco, and cannabis use. At the parental level, risk factors include pain experience (pain intensity, history of chronic pain), pain cognitions and behaviors, mental health symptoms, history of substance-related problems, and presence of opioids (surplus and/or not prescribed to the adolescent) in the home. The purpose of this study was to explicitly examine the empirical strength of each of these interrelationships proposed in this theoretical model (10). Understanding the risks for adolescent substance use not only after initiation of opioid therapy, but in the immediate context of approved candidacy for opioid therapy to manage acute pain, can provide insight into potential approaches for developing individualized pediatric prescribing practices aimed at mitigating risk for adverse pain- and opioid-related outcomes.

Figure 1.

Integrated theoretical model of risk and protective factors for pain and opioid use outcomes among youth receiving an opioid prescription

Note. From Dash, Wilson, Morasco, and Feldstein Ewing (2018). Reproduced with author permission.

1.2. Individual-Level Risk Factors

1.2.1. Pain-related beliefs and cognitions.

Adolescents’ perceptions of their pain and beliefs about their ability to tolerate and manage pain appear highly relevant in opioid use outcomes (10, 11). Higher pain catastrophizing impacts pain interference among adolescents (12), and is associated with higher perceived pain intensity, pain-related disability, and risk for future opioid misuse (13–15). Additionally, pain self-efficacy plays a stronger role in perceptions of pain intensity and trajectory of pain management than does injury severity (16). In fact, higher pain-self efficacy has been associated with comparably greater pain relief than opioid therapy in adults (17). As such, self-assessments of the adolescent pain experience, in terms of both pain severity and confidence in their ability to tolerate and manage it, are essential to consider when evaluating an adolescents’ candidacy for opioid treatment.

1.2.2. Mental health symptoms.

Psychopathology is a well-established risk factor for later substance use among adolescents (18, 19). While most research has been conducted in adults, mental health symptoms appear to confer substantial risk for opioid-related problems in adolescents as well (18, 20–22). One third of adolescents presenting for acute pain also report experiencing depressive symptoms (23), which can contribute to persistent pain and greater risk of opioid use as a mechanism to alleviate psychological and physical pain (20, 21). Adolescents with pre-existing anxiety and depression are more likely to receive an opioid prescription, report sustained opioid use, and transition to long-term opioid therapy (5, 24), indicating the importance of evaluating mental health symptoms prior to prescribing opioids for pain management.

1.2.3. Substance use.

An existing history of alcohol, tobacco, and cannabis use places adolescents at enhanced risk for opioid misuse even when receiving a legitimate prescription (5). This is an increasingly salient issue in light of changing cannabis legislation and the popularization of cannabis use (medical and/or recreational) in states with permissive legislation, which results in increased availability, higher perceived peer use, and decreased perception of harm from use (25–28). Further, introduction of opioids during the course of routine pediatric ambulatory care may introduce non-substance-seeking youth to positive opioid experiences (e.g., euphoric “high,” abatement of distress) that they would otherwise not be exposed to (29, 30). Such exposure, coupled with the low perceived harm of opioids that comes from being prescribed a medication by one’s healthcare provider, may place youth at risk for future opioid use (10, 31).

1.3. Parent-Level Risk Factors

1.3.1. Parent(s) with pain conditions.

Genetic and shared environmental factors associated with family history of chronic pain increase adolescent risk for pain conditions, many of which have historically been managed with opioids. Chronic pain conditions are moderately heritable (32, 33), and children of parents with chronic pain are at elevated risk for substance use in adolescence (34). With respect to familial environment, disrupted parenting styles, difficult social and emotional climates in the home, and chronic stressors within the family unit that result from parental pain can also impact adolescent risk for substance misuse (35–41). Parents with chronic pain may be more likely to catastrophize their child’s pain, communicating that pain is a threat that should be rapidly mitigated and subsequently influencing their child’s perception and management of pain (10, 36, 37). These parent behaviors can increase adolescents’ vulnerability to higher perceived pain and functional disability (35). More directly, parents with pain are more likely to have opioids at home. This can contribute to lower perceived harm, facilitate access to opioids, and create opportunities for future experimentation in a “safe” environment with a relatively low-risk method of acquisition (41). Furthermore, parental modeling of pain behavior influences adolescent pain responses and pain management choices, including analgesic use (35, 36, 42–44).

1.3.2. Parent mental health.

Mental health symptoms such as anxiety and depression are, to some degree, heritable, and environmental factors associated with parents who experience these symptoms also increase adolescent vulnerability to these factors (45). Parents with anxiety and/or depression can be emotionally and practically unavailable, irritable, self-preoccupied, and pessimistic (46). Furthermore, decreased parental monitoring that accompanies parental unavailability is itself a risk factor for adolescent opioid misuse (38). These parent behaviors can cascade into adolescent anger, depression, and anxiety, which are also associated with adverse substance use outcomes among adolescents (46). While adolescent doctor visits do not typically include conversations about parent mental health, these factors may be important in the calculation of risk inherent in pediatric opioid prescribing.

1.3.3. Family history of substance use.

Family history of substance use disorder (SUD) is a long-established risk factor for adolescent SUD. Risk for SUD is heritable (47, 48), as are various opioid use phenotypes (49). Adolescents living with parents who have SUD are exposed to increased environmental risk for substance use via both parental modeling and substance accessibility (low perceived harm around use, ease of access to substances of abuse; 50). Factors such as parental rules and attitudes regarding substance use can shape adolescents’ perception of substance use risk and norms (51–53), which in turn impacts potential adolescent misuse of prescribed opioid medication.

1.4. Present Study

The present study offers a novel contribution to our understanding of adolescent risk in the context of routine pediatric ambulatory care by capturing the interplay of pain, mental health symptoms, substance use characteristics, and family factors of adolescents immediately following receipt of opioids to treat pain as prescribed by a pediatric medical provider. The purpose of the study was to examine the presence and intersection of adolescent and parent characteristics through the lens of an integrated theoretical model of adolescent risk for opioid use outcomes that addresses the intersection of individual and familial factors (10), and to do so in a group of adolescents deemed by medical providers to be suitable candidates for pediatric opioid-based pain management. Our goals were threefold: 1) to examine cumulative risk in adolescents who were prescribed opioids in routine pediatric outpatient care, 2) to examine these risk factors concurrently to ascertain their relative importance in the presentation of adolescent-reported pain intensity and OUD symptoms, and 3) to assess the potential importance of parent-level risk factors in adolescent pain intensity and OUD symptoms. Based upon the previous literature in each of these areas, it was hypothesized that: 1) the sample would evidence high levels of catastrophizing, low levels of self-efficacy, high rates of depressive symptoms, and high rates of parental chronic pain, 2) adolescent pain-related cognitions, as well as parent pain history, pain interference, and pain catastrophizing would have the greatest predictive value for adolescent pain intensity, and 3) adolescent pain intensity, pain-related cognitions, depression, and substance use, as well as parent pain history, pain catastrophizing, mental health, substance use history, and presence of opioids in the home would most substantively predict adolescents’ OUD symptoms.

2. Methods

2.1. Participants

The sample included 70 adolescent-parent dyads in ambulatory medical care settings presenting for opioid treatment of acute pain in the adolescent. Reflecting the metropolitan northwest region where this sample was recruited, the adolescent sample was 49% female (M_age=16.4, SD=1.3; range=14–19), 69% non-Hispanic white, and 17% Hispanic/Latinx. The parent sample was 90% female (M_age=46.2, SD=6.5), 81% non-Hispanic white, and 10% Hispanic/Latinx; 81% were married and 54% worked full time.

A majority (71%) of adolescents were recruited from a day surgery unit, and a majority (80%) of opioid prescriptions were given for pain management after outpatient surgery. The most common prescriptions were 5mg Oxycodone (66%) and Hydrocodone-Acetaminophen 5–325mg (33%). Number of doses provided ranged from 2 (1%) to 60 (3%), with the most common amounts being between 20 and 40 doses (47%). Instructions most commonly indicated to take the opioid as needed for pain every 4–6 hours. Many adolescents in the sample (44%) had not taken an opioid in the 6 months prior to assessment; only 1 (1.5%) reported currently taking a prior opioid prescription at the time of assessment.

2.2. Procedures

All study procedures were approved by the institution’s IRB. Participants were identified within 72 hours of receiving a prescription for opioids as a component of the first wave of a larger ongoing longitudinal study (see Figure 2 for recruitment details). Participants were recruited from pediatric emergency departments, ambulatory day surgery settings, and general and specialty pediatric clinics. Manual chart reviews and electronic medical records system alerts identified youth who were prescribed opioids during the course of routine pediatric medical care. Inclusion criteria included adolescent age 14–18 at time of enrollment, adolescent and parent English speaking, and adolescent being prescribed opioid medication for acute pain. Potential participants were excluded if they had a life-threatening or serious chronic medical condition for which they would be receiving chronic opioid care (e.g., cancer, sickle cell disease, arthritis), were admitted to the hospital overnight, or had a diagnosis of cognitive disability or developmental delay. Potentially eligible youth were contacted in person or via telephone by study staff who completed a screening with families. Eligible youth and their parent provided consent either in person or electronically with a phone consent review. Both adolescent assent and parent consent were required for participation. Adolescents and their parents completed separate telephone surveys with trained research staff. Self-report measures were sent separately via e-mail to adolescents and their parents via the Research Electronic Data Capture system (RED-Cap; 54).

Figure 2.

Flowchart of recruitment procedures

The 72-hour time frame for enrollment was selected with careful consideration. Targeting a subject pool of adolescents who received opioids for treatment of acute pain necessarily precludes sampling prior to prescription (that is, we cannot know which adolescents will be prescribed opioids prior to the prescription event). As such, it was the earliest time point possible to capture youth receiving opioids in order to examine risk in the immediate wake of prescription receipt. To our knowledge, this is the only study that has conducted in vivo empirical evaluation of adolescents in their course of receiving an opioid to treat pain. Importantly, these experiences are precisely the ones many opioid misusers report as setting them on their OUD trajectory (4, 7, 55).

2.3. Measures

2.3.1. Adolescent pain intensity.

Adolescents were asked to rate the typical intensity of their pain in the past week. Response options ranged from 0 (no pain) to 10 (worst pain possible; 56).

2.3.2. Distress Tolerance Scale (DTS).

Adolescents were administered an abbreviated 10-item version of the DTS (57), which assesses capacity to tolerate pain-related distress. Responses were rated on a 1 (strongly agree) to 5 (strongly disagree) scale, such that higher scores reflect higher distress tolerance. A mean distress tolerance score was derived from these ratings (range=1–5).

2.3.3. Pain Self-Efficacy Questionnaire (PSEQ).

Adolescents were administered the PSEQ (58), a 7-item questionnaire assessing adolescents’ confidence in their capacity to engage in a variety of activities (e.g., house chores, being with friends) when they have pain. Responses were rated on a 1 (very sure) to 5 (very unsure) scale, such that higher scores reflect lower self-efficacy. A mean pain self-efficacy score was derived from these ratings (range=1–5).

2.3.4. Pain Catastrophizing Scale (PCS).

Adolescents and their parents responded to the PCS-Child (59) and the adult PCS (60), respectively, regarding their own pain experience. Responses were rated on a 0 (not at all) to 4 (extremely/all the time; PCS-C/PCS) scale and summed to create a total score (range=0–52). A cut-score of 15 or higher was used to indicate clinical significance for the PCS-C (61) and a cut-score of 20 or higher (50^th percentile) was used to indicate clinical significance for the PCS (60).

2.3.5. Parent pain characteristics.

Parents were administered the National Institutes of Health’s (NIH) Patient-Reported Outcomes Measurement Information System (PROMIS-29) measures of their own pain intensity and pain interference in the past week (62). Average past-week pain intensity was rated on a 0 (no pain) to 10 (worst pain imaginable) scale. The pain interference scale is a 6-item measure assessing the extent to which pain interfered with completing activities of daily living (e.g., household chores, social activities) in the past week; responses were rated on a 1 (not at all) to 5 (very much) scale. Scores were t-score standardized for analysis per PROMIS scoring guidelines (M=50, SD=10).

2.3.6. Family history of pain.

Parents reported if their child’s biological mother and father have ever experienced chronic pain, defined as pain that lasted for three or more months.

2.3.7. Anxiety and depression.

Adolescents and their parents were administered the brief anxiety and depression symptom scales as part of the NIH PROMIS-29 (62, 63). Patients were queried with 4 items for each anxiety (e.g., “I felt like something awful might happen”) and depression (e.g., “I felt like everything in my life went wrong”) symptoms over the past week. Responses were rated on a 1 (never) to 5 (almost always) scale. Scores were t-score standardized for analysis per PROMIS scoring guidelines (M=50, SD=10).

2.3.8. Adolescent substance use.

Adolescents were asked whether they had ever used alcohol, tobacco, and cannabis in their lifetime. Adolescents responding positively were queried as to how many times they have used each substance in their life. A composite lifetime substance involvement variable was created by summing use events across substances.

2.3.9. DSM-5 Opioid Use Disorder (OUD) symptoms.

Adolescents were administered the 11-item DSM-5 diagnostic criteria for OUD (64). Responses were given as yes/no answers to each criterion. A total OUD symptom score was created by summing the number of symptoms endorsed (range=0–11).

2.3.10. Accessibility of opioids.

Parent assessments queried possession of prescription medications in the home, and those who responded positively were asked about several classes of prescription medications, including opioids.

2.3.11. Family history of substance use.

Parents reported whether the adolescent’s biological mother and father ever experienced 2 or more substance-related problems in their lifetime.

2.4. Analytic Approach

Analyses were conducted in SPSS version 25 (65). Cross tabulations and descriptive analyses were run to calculate the frequencies and means of pain characteristics, anxiety and depression scores, and substance use variables for both adolescents and parents. Individual and parent risk variables were coded as being significantly elevated or not based on existing clinical cut points, where available; continuous format was retained for the main analyses. To address aim 1, these variables were summed to calculate a cumulative adolescent risk score (range=0–21), and a simple regression was run to determine the influence of cumulative risk on OUD symptoms.

To address aims 2 and 3, a series of hierarchical regression models were built. These models tested the hypotheses that the adolescent risk factors identified by the contextual model of adolescent risk for opioid misuse (10) would contribute to past-week adolescent pain intensity ratings and cumulative adolescent OUD symptoms, and that parent-level risk factors would contribute to these outcomes over and above adolescent risk factors. First, we ran two full hierarchical models that included all predictors with past-week pain intensity and OUD symptoms as outcomes. At step 1, all adolescent characteristics were entered into each model. At step 2, all parent characteristics were added to the model with all adolescent variables retained. Next, reduced models utilizing a critical alpha of p=.25 were run to determine the most parsimonious model (66, 67).

3. Results

3.1. Cumulative Risk

Most adolescents in this sample (N=29, 41%) presented with 2–3 risk factors, and 25 youth (35%) had 5 or more risk factors (see Table 1). The most frequently occurring individual risk factors were high pain catastrophizing (N=32, 46%), history of alcohol use (N=28, 40%), elevated depression symptoms (N=24, 34%), and moderate-high pain intensity (N=23, 33%). The most frequently occurring parent and family risk factors were presence of prescription opioids in the household (N=50, 71%), chronic pain in biological mother (N=23, 32%), and parent anxiety symptoms (N=15, 21%). Cumulative risk predicted OUD symptoms, (F[1, 66]=22.49, p<.001).

Table 1.

Sample pain, mental health, and substance use characteristics

Adolescent
Pain Characteristics	Mean	SD	Range	Clinical Cutoff	N (%) with Risk
Pain Intensity- Past Week	2.82	2.28	0–10	≥5	23 (32.86)
Distress Tolerance	2.90	0.60	1–5	>1 SD above mean	5 (7.14)
Pain Self-Efficacy	2.42	0.93	1–5	>1 SD above mean	13 (18.57)
Pain Catastrophizing	15.07	11.18	0–44	≥15	32 (45.71)
Lifetime Mental Health Symptoms	Mean	SD	Range	Clinical Cutoff	N (%) with Risk
Anxiety Symptoms	50.52	11.65	*	>60	14 (20.00)
Depression Symptoms	52.70	11.59	*	>60	24 (34.29)
Substance Use					N (%) with Risk
Ever Used
Alcohol					28 (40.00)
Tobacco					7 (10.00)
Cannabis					14 (20.00)
Parent
Pain Characteristics	Mean	SD	Range	Clinical Cutoff	N (%) with Risk
Pain Intensity- Past Week	1.59	1.79	0–10	≥5	5 (7.14)
Pain Interference	48.84	9.10	*	>60	9 (12.86)
Pain Catastrophizing	8.73	8.06	0–44	≥20	8 (11.43)
History of Chronic Pain
Biological Mother					23 (32.39)
Biological Father					14 (19.72)
Biological Mother & Father					9 (12.68)
Lifetime Mental Health Symptoms	Mean	SD	Range	Clinical Cutoff	N (%) with Risk
Anxiety Symptoms	50.65	10.36	*	>60	15 (21.43)
Depression Symptoms	46.06	8.79	*	>60	6 (8.57)
Substance-Related Characteristics					N (%) with Risk
Lifetime Alcohol and/or Substance-Related Problems
Biological Mother					7 (10.00)
Biological Father					7 (10.00)
Biological Mother & Father					3 (4.29)
Prescription Opioids Kept in Household					50 (71.43)

Cumulative Adolescent Risk
Risk Factors	1	2	3	4	5	6	7	8	9	13
N of Sample	7	12	17	9	4	3	8	7	2	1
% of Sample	10.00	17.14	24.29	12.86	5.71	4.29	11.43	10.00	2.86	1.43

Note. Possible cumulative risk factor range=0–21;

^*measure is t-score standardized.

3.2. Pain Intensity

The full and reduced hierarchical regression models examining the role of adolescent and parent risk factors in adolescent-reported past-week pain intensity are presented in Table 2. Pain catastrophizing emerged as a significant predictor at step 1 of the full model. At step 2, past-week parent pain intensity also emerged as a significant predictor of adolescent pain. Inclusion of parent factors doubled the variance explained. In the more parsimonious reduced model, adolescent pain catastrophizing, adolescent depression, parent pain intensity, and parent history of chronic pain emerged as significant predictors of adolescent pain. Parent mental health and substance use history did not appear to enhance prediction of adolescent pain intensity.

Table 2.

Full and reduced models predicting past-week pain intensity

Full Model					Reduced Model
Step 1	B (SE)	β	t	p	B (SE)	β	t	p
Pain Catastrophizing (A)	.11 (.05)	.51	2.23	.03*	.07 (.04)	.33	1.97	.05*
Distress Tolerance (A)	−.52 (.66)	−.13	−.78	.44	-	-	-	-
Pain Self-Efficacy (A)	−.88 (.64)	−.36	−1.39	.18	−.56 (.46)	−.22	−1.21	.23
Anxiety (A)	−.04 (.05)	−.19	−.72	.48	-	-	-	-
Depression (A)	.07 (.05)	.36	1.25	.22	.08 (.03)	.40	2.59	.01*
Substance Use (A)	.01 (.01)	.11	.70	.49	-	-	-	-
	F(6, 32)=2.28, p=.06, R2=.30				F(3, 50)=5.31 p=.003, R2=.24
Step 2	B (SE)	β	t	p	B (SE)	β	t	p
Pain Catastrophizing (A)	.10 (.05)	.48	2.16	.04*	.06 (.03)	.30	21.99	.05*
Distress Tolerance (A)	−.14 (.75)	−.04	−.19	.85	-	-	-	-
Pain Self-Efficacy (A)	−.72 (.71)	−.29	−1.00	.33	−.22 (.44)	−.09	−.50	.62
Anxiety (A)	−.04 (.05)	−.19	−.80	.43	-	-	-	-
Depression (A)	.09 (.05)	.50	1.74	.10	.06 (.03)	.32	2.16	.04*
Substance Use (A)	.00 (.01)	−.00	.00	.99	-	-	-	-
Pain Intensity (P)	.91 (.34)	.75	2.67	.01*	.70 (.17)	.53	4.02	<.001
Chronic Pain (P)	−1.50 (.87)	−.33	−1.72	.10	−1.29 (.61)	−.28	−2.11	.04*
Pain Interference (P)	.00 (.08)	.00	.00	.99	-	-	-	-
Pain Catastrophizing (P)	.00 (.05)	−.02	−.08	.94	-	-	-	-
Anxiety (P)	.00 (.04)	−.02	−.08	.94	-	-	-	-
Depression (P)	−.06 (.06)	−.23	−.92	.37	-	-	-	-
Substance Problems (P)	.14 (.83)	.03	.17	.87	-	-	-	-
Opioids at Home (P)	.53 (1.20)	.07	.44	.66	-	-	-	-
			ΔR2=.29				ΔR2=.19
	F(14, 24)=2.34, p=.03*, R2=.59				F (5, 48)=7.37, p<.001, R2=.43

Note. Bold indicates significance at Bonferroni-corrected p<.003;

^*indicates significance at p<.05;

A=adolescent factor; P=parent factor; reduced model=predictors with p>.25 (at step 1 for A variables) not carried forward.

3.3. Opioid Use Disorder Symptoms

Ten percent of adolescents endorsed two or more OUD symptoms. Hierarchical regression models predicting OUD symptoms are presented in Table 3. At step 1 of the full model, no individual adolescent characteristics predicted OUD symptoms, despite a marginally significant model overall. The model was nonsignificant at step 2, although parent pain interference emerged as a potentially influential factor. In the reduced model, adolescent pain catastrophizing predicted OUD symptoms at step 1, and adolescent pain catastrophizing and parent pain interference predicted OUD symptoms at step 2, accounting for an additional 6% of the variance. Adolescent and parent mental health and substance use factors appeared unrelated to adolescent OUD symptoms.

Table 3.

Full and reduced models predicting OUD symptoms

Full Model					Reduced Model
Step 1	B (SE)	β	t	p	B (SE)	β	t	p
Pain Intensity	−.12 (.18)	−.12	−.69	.49	-		-	-
Pain Catastrophizing (A)	.08 (.05)	.38	1.60	.12	.10 (.03)	.54	3.30	.002
Distress Tolerance (A)	−.66 (.68)	−.16	−.97	.34	-	-	-	-
Pain Self-Efficacy (A)	−.98 (.66)	−.38	−1.47	.15	−.50 (.38)	−.21	−1.31	.19
Anxiety (A)	.05 (.05)	.24	.97	.34	-	-	-	-
Depression (A)	.06 (.06)	.31	1.12	.27	-	-	-	-
Substance Use (A)	.00 (.01)	.06	.36	.72	-	-	-	-
	F(7, 31)=2.63, p=.03*, R2=.37				F(2, 63)=6.51, p=.003, R2=.17
Step 2	B (SE)	β	t	p	B (SE)	β	t	p
Pain Intensity	−.27 (.23)	−.26	−1.23	.23	-		-	-
Pain Catastrophizing (A)	.11 (.06)	.51	1.98	.06	.10 (.03)	.52	3.30	.002
Distress Tolerance (A)	−.85 (.83)	−.20	−1.03	.31	-	-	-	-
Pain Self-Efficacy (A)	−1.03 (.81)	−.40	−1.28	.22	−.43 (.37)	−.19	−1.18	.24
Anxiety (A)	.06 (.06)	.26	1.02	.32	-	-	-	-
Depression (A)	.07 (.06)	.35	1.10	.28	-	-	-	-
Substance Use (A)	.00 (.02)	.01	.04	.97	-	-	-	-
Pain Intensity (P)	−.25 (.43)	−.20	−.59	.56	-	-	-	-
Chronic Pain (P)	.07 (1.02)	.01	.07	.95	-	-	-	-
Pain Interference (P)	.19 (.08)	.66	2.23	.04*	.06 (.03)	.24	2.13	.04*
Pain Catastrophizing (P)	−.03 (.06)	−.09	−.46	.65	-	-	-	-
Anxiety (P)	−.02 (.04)	−.08	−.42	.68	-	-	-	-
Depression (P)	−.03 (.07)	−.10	−.37	.72	-	-	-	-
Substance Problems (P)	.77 (.92)	.13	.83	.42	-	-	-	-
Opioids at Home (P)	.80 (1.34)	.10	.60	.56	-	-	-	-
			ΔR2=.19				ΔR2=.06
	F (16, 21)=1.99, p=.07, R2=.56				F (3,62)=6.09, p=.001, R2=.23

Note. Bold indicates significance at Bonferroni-corrected p<.003;

^*indicates significance at p<.05;

A=adolescent factor; P=parent factor; reduced model=predictors with p>.25 (at step 1 for A variables) not carried forward.

4. Discussion

This study presents the initial evaluation of a time-sensitive picture of adolescents who were recently prescribed an opioid during routine ambulatory care for acute pain. Quite uniquely, it represents the first empirical profiling of the intersection of pain, mental health, and substance use characteristics of adolescents and their parents in the immediate context of opioid prescribing. This approach offers critical insight into the profile of risk factors among adolescents who have received opioids to treat acute pain. The influential factors in risk for pain and OUD symptoms appeared quite divergent, with the exception of the role of pain catastrophizing. Adolescent pain catastrophizing and mental health, as well as parent pain intensity and history, predicted adolescent pain intensity, while only adolescent catastrophizing and parent pain interference, but not actual pain characteristics, predicted adolescents’ OUD symptoms. These results indicate that a relatively brief risk clinical screening to assess adolescent pain catastrophizing, adolescent depression, and parent pain history may be a valuable start to protecting adolescents at risk for opioid use outcomes in the context of pain management.

Unexpectedly, lower distress tolerance and lower pain self-efficacy were not significantly associated with past-week adolescent pain intensity, although the effect for pain catastrophizing aligned with hypotheses and extant literature. The evidence presented here indicates that pain catastrophizing, but not youth’s own perception of pain intensity, may be associated with risk for OUD symptoms. In light of evidence that perception of pain can be a better predictor of pain outcomes than injury severity (5), it may be the case that pain catastrophizing is a better indicator of risk for future opioid misuse than pain intensity.

While the retrospective report of past-week pain intensity implemented may not fully capture the role of moment-to-moment pain variations that can influence the decision to use an opioid, it is essential to note that adolescent ratings of past-week pain intensity in this sample were not uniformly high. This may reflect a reduction in pain intensity due to using opioids appropriately, but it may also reflect opioid prescribing for pain that is only mild to moderate. This highlights that, within this sample recruited in the midst of the 2018–2019 opioid epidemic, many adolescents were prescribed opioids, at times in great quantity, by caring professionals despite modest pain ratings. In other words, these pain conditions may have been well-managed with non-steroidal anti-inflammatory drugs (NSAIDs) and/or nonpharmacological approaches (68–73) rather than with opioids, which have a very high potential for addiction. This is an issue of particular urgency given that early exposure to opioids is a highly prevalent risk factor in the continued trajectory into adulthood.

4.1. Limitations

While this study has many strengths, including the novel empirical design and the timely capture of an otherwise non-existent sample of adolescents in routine medical care, results should be interpreted with the following limitations. First, our ability to detect and explain the relationships between the factors examined here with precision was likely limited by our relatively small sample size. Further work is needed to further support these initial findings and to examine the role of pain-related psychological factors in adolescent candidates for opioid-based pain management. Our sample was also predominantly white and accompanying parents were almost exclusively female, potentially limiting the generalizability of the presented findings; however, it should be noted that many of the regions impacted by opioid epidemic have demographics that parallel our sample. Despite this, future studies would benefit from examining similar patterns within other geographic regions to improve generalizability. Additionally, data were collected from one parent, leaving information about the pain characteristics, attitudes, and behaviors of other family members unknown. Information about physician-rated pain severity was not collected, so it is not possible to disentangle adolescents’ perception of their pain from a more objective rating of its severity. Future studies in this area may consider integrating physician reports into their assessments in addition to data collected from adolescents and their parents.

4.2. Conclusions

Despite these limitations, the present study provides insight into a range of characteristics of adolescents and their parents relevant to risk in the context of opioid prescribing during routine ambulatory care via evaluation of these risks in a compelling, time-sensitive window (within 72 hours after prescription). The results highlight the importance of understanding the intersection of pain, mental health, and substance use at both adolescent and parental levels to more fully understand the risks that adolescents face upon initiating opioid therapy. It will be important to continue to follow these adolescents through the course of opioid treatment to monitor trajectories of pain, mental health, and substance use to gain a more thorough picture of how baseline factors at the point of prescription may or may not influence substance use outcomes. More broadly, understanding how these types of risks, both of the individual adolescent and those conferred by parents, can aid in informing prescribing practices and evaluation of risk with opioid treatment. Integrating alternative, nonpharmacological therapies and/or over-the-counter medications into pain management may be a viable alternative for adolescents screening positive for many of these potential risk factors.

Highlights.

• Despite the proliferation of opioid prescriptions and associated sequalae, rates of adolescent opioid prescribing for acute pain in ambulatory care remain steady.

• Parent factors appear to drive adolescent experience of pain intensity, while adolescent factors primarily drive opioid use disorder symptoms.

• Adolescent pain catastrophizing may be a more robust indicator of risk for future opioid use than pain intensity.

• Pediatric health providers are uniquely well-positioned to integrate information about adolescent opioid use risk at initial treatment contact prior to developing treatment plans for pediatric patients with acute pain.