Abstract
Black patients are less likely than white patients to receive pain treatment, especially opioids, for both acute and chronic pain. Black men are at higher risk than other populations of being “assumed criminal” regardless of any involvement in criminal activity. Additionally, certain injury and patient characteristics such as intentionality of injury and substance use history may lead providers to suspect criminal involvement and impact pain treatment decisions. The purpose of this study was to describe factors that predict receipt of opioid prescription at hospital discharge. We conducted a secondary analysis of data from a cohort of 623 seriously injured Black men treated at trauma centers in Philadelphia between 2013–2017. Regression models were used to examine relationships between discharge opioid prescriptions, injury intent and substance use history. Controlling for age, injury severity, pain score, length of hospital stay (LOS), insurance type, and year of study, receipt of opioids was not impacted by injury intent. However, patients who self-reported substance overuse were less likely to receive opioids than those who did not. Patients with higher injury severity, pain scores, and longer LOS were more likely to receive opioids. Of patients who received opioids, patients with higher pain scores and longer LOS received higher dosages than those with lower scores and shorter LOS. While previous research highlights stigmatization experienced by intentionally injured patients, injury intent did not impact receipt of discharge opioid prescriptions in this study. Future research should continue to explore the effect of injury intent on patients’ experiences in the healthcare system.
Keywords: injury, pain, analgesia, opioid
Introduction
As emergency care for severely injured patients improves, more patients are surviving serious injuries and reentering the community[1,2]. Many of these patients require ongoing pain management upon discharge from the hospital to prevent the negative physical, mental, social, and financial outcomes associated with inadequately treated pain[3,4]. At the same time, fears about potential for misuse or diversion of opioid analgesics have impacted clinical practice and providers have become more cautious and restrictive in their opioid prescribing[5,6].
Certain patients are at higher risk for undertreatment of pain; for example, Black patients are less likely to be treated for traumatic, surgical, and chronic pain than white patients with comparable illness and injury[7,8]. Black patients are also less likely to be prescribed opioids than white patients, despite a lack of evidence that Black patients are at higher risk of opioid misuse[8,9], and physicians are likely to underestimate the pain of their Black patients versus white patients[10]. It is possible that the assumed criminality due to racial stereotyping and bias is one factor that may influence the receipt of pain treatment, especially opioid analgesia[11]. However, little is known about how patient and injury characteristics, especially those that diverge on health and criminal justice concerns, like substance use history or intentionality of injury (intentional injury being a gunshot/stab injury vs. unintentional injury due to motor vehicle accident or fall), may impact the pain treatment of injured Black men.
Black men are more likely than other demographic groups to be assumed criminal in US society, regardless of actual criminal behavior[12,13,14]. The conflation of Blackness with criminality has pervaded the breadth of US history[12,15]. Its evidence in the present day is diverse but can be seen through its consequences which range from overwhelming overrepresentation of Black people in jails and prisons[12] to recent media that highlights instances across the country in which police are called to respond Black people engaging in non-criminal, everyday activities[16,17,18].
As in larger society, the consequences of assumed criminality can play out in healthcare encounters. Previous research has highlighted some ways in which Black patients hospitalized with intentional injuries like gunshot wounds perceive that they are stigmatized based on their race and how they were injured[19,20,21]. John Rich’s in-depth research on the lived experience of Black men injured by gun violence illustrated examples of healthcare providers making assumptions about these patients’ involvement in criminal activities[22]. Rich hypothesized that providers may have difficulty empathizing with intentionally injured patients, since providers themselves are unlikely to live in neighborhoods with high risk of injury due to gun violence. This lack of common experience, he theorized, contributes to assumptions that these patients “did something” to warrant their injuries[22].
The interplay between racial bias, assumed criminality and traumatic injury may have direct implications on how Black patients are prescribed opioids. Black people are six to ten times more likely than white people to be incarcerated for substance-related offenses, despite evidence that Black and white people use illicit substances (including opioids) at comparable rates[23,24,9]. These statistics point to the racialized responses to substance use. While substance misuse is an ongoing public health issue across the US population, it is much more likely to be treated as a criminal justice concern when it affects people of color[24].
The aim of this study was to examine the factors associated with the receipt of a discharge opioid prescription in a cohort of seriously injured Black men. We hypothesized that men with intentional injuries and self-reported history of substance use are less likely to be discharged on opioids than men with unintentional injuries and no self-reported substance use histories. Further, of those discharged on opioids, it was hypothesized that men with intentional injuries and substance use histories would receive lower morphine equivalent doses than men with unintentional injuries and no substance use history.
Methods
Description of Parent Study
Parent Study Design.
This study was a secondary analysis of data from the Psychological Effects of Injuries in Injured Black Men Study, which enrolled a cohort of 623 seriously injured Black men, 55.5% of whom were intentionally injured and 44.5% of whom were unintentionally injured[25]. Eligible patients were Black men, 18 years of age and older, English speaking, and residing in the Greater Philadelphia metropolitan area. All participants had sustained acute traumatic injuries (ICD-9 codes 800–995) that were severe enough to require hospitalization. Exclusion criteria were central nervous system (CNS) injury or current symptoms of serious psychotic disorder (DSM 5 axis 1 diagnoses), injury caused by suicide attempt/suicidal ideation, or if currently receiving treatment for PTSD or depression. These individuals were excluded from the parent study as emergence of PTSD or depression symptoms was the primary outcome of that study. Patients who had a history of psychotic disorders but were not currently experiencing psychotic symptoms were not excluded. The sample was comprised of consecutively consented and enrolled injured men who met study criteria during admission to a trauma center at two hospitals in Philadelphia from 2013–2017. Participants agreed to be interviewed during hospitalization and three months after hospital discharge, and to allow access to electronic health records (EHR) to obtain injury-related medical information.
Description of Study
This secondary data analysis included all 623 participants and used a cross-sectional design. With IRB approval, the parent study team created a de-identified dataset from baseline parent study interviews and the participants’ EHR. The final dataset included injury intent, mechanism and severity, self-reported substance use, urine drug screens, discharge analgesic orders, demographic characteristics, year of participation in parent study, length of stay (LOS) and in-hospital pain ratings for the 24 hours prior to discharge.
Instruments & Measurement
Independent variables.
Injury was coded as intentional (e.g. violent) and unintentional (e.g. motor vehicle crash) at the hospital where these data were collected by trained trauma registrars using patient descriptions of their injuries and following definitions from the statewide criteria for the Pennsylvania Trauma Systems Outcome Study[26]. This data collection is required for all patients admitted to the hospital’s trauma service. Coding is based on the patient’s descriptions of their injuries at the time of hospitalization. If the intent of injury was not clearly coded by the registrar, the PI and one Co-I (both experts in traumatic injury) and data analyst of the parent study reviewed all information to reach consensus on the classification as intentional and unintentional. For example, a jump out of a window was classed as intentional when during a home invasion an intruder was chasing the person. Similarly, a case “fall from 2ndfloor porch roof when being attacked by landlord with machete and knife” was coded as intentional. Conversely a gunshot wound that resulted from a discharge by a child playing with a gun in the home was coded as unintentional.
Substance use history was determined during the parent study survey interview by two measures: (1) self-report of substance use in the 6 months prior to injury and (2) self-report of substance overuse, which was determined by an affirmative answer to the question: “during the past 6 months, have you felt that you use too much alcohol or drugs?” If participants answered yes, they were asked which substances they overused. We examined the admission urine drug screens but only 278 patients were administered urine drug screens and many patients are given pain medications and/or benzodiazepines in transit to the hospital. In addition, we were unable to access data about substance use diagnoses as they are protected under Pennsylvania state law. We chose to use self-report from parent study interviews in model construction, recognizing the limitation that data collected during parent study interviews may not mirror what patients self-report to clinicians.
Outcome variables.
To determine discharge prescription dose, we first examined all discharge medication orders and extracted opioid medications. We used a morphine milligram equivalents (MME) calculator to convert all opioid dosages to morphine equivalents[27,28].
Covariates.
Injury severity.
Injury severity was calculated by trauma registrars using the Abbreviated Injury Scale (AIS). The AIS ranks the severity of injuries on a 1–6 scale (1 is minor, 6 is most severe), measuring injuries in 6 body systems (head/neck, face, thorax, abdomen, extremities, external)[29]. The AIS ratings are then used to calculate an Injury Severity Score (ISS), summing the square of the most severe injury across three body systems. The summed score ranges from 1–75 (1 is most minor and 75 is incompatible with life)[30]. These scores are grouped into categories: minor (<9), moderate (9–15), severe (16–25), and very severe (>25)[31]. Due to the small number of participants in the parent dataset with very severe injuries, we combined participants with injury severity scores ≥16 into one “severe” category.
Pain Severity.
The highest pain score reported (on a 0–10 numeric rating scale) during the 24 hours prior to discharge was used, as an average of all scores recorded may have been skewed by low scores recorded after receipt of pain medication. The hospitals from which participants were recruited for the parent study have similar policies about pain assessment, which involve mandatory assessment by RNs a minimum of every 8–12 hours. This variable was transformed into a categorical variable: mild (≤4), moderate (>4 & ≤7), and severe (>7).
Length of stay.
LOS was defined as the number of days of hospitalization and was categorized as 1–3 days, 4–7 days, 8–14 days, and >14 days. Participants were likely to have received opioid pain treatment during their hospitalization, and length of exposure to opioids may impact discharge prescriptions as patients can develop physical dependence and tolerance and providers might take this into consideration when planning discharge pain treatment. Unfortunately, we did not have access to data describing amount of opioid pain medication prescribed during hospitalization, thus LOS was used as a proxy.
Sociodemographic characteristics.
In that younger persons are more likely to misuse substances than older persons[32], age could possibly impact provider beliefs about risk of substance misuse or diversion. Insurance status was included because of the potential impact of patients’ ability to pay for medications on provider treatment decisions. The year that individuals participated in the parent study was included to account for changes in prescribing practices over time.
Power Analysis
With the available sample size, the power analysis, using PASS, indicated that we had 80% power to detect a difference of 0.11 in proportion of patients who received opioids at discharge.
Data Analysis
STATA 15.1 was used for statistical analysis. Chi-square tests were used to compare the proportions of participants discharged on opioids with intent and substance use history respectively. Logistic regression models were constructed to estimate the proportion of men who were prescribed opioids with and without intentional injuries and substance use histories. The interaction term of intentional injury and substance use history was not significant (p=0.76) and was removed from the model. We then constructed the adjusted model, adjusting for age, injury severity, pain score, LOS, insurance status, and year of study. For those discharged on opioids, t-tests and linear regression models were used to compare the morphine milligram equivalent doses in men with and without intentional injury and substance use histories. Again, the interaction term of intentional injury and substance use history was not significant (p=0.80) and was removed from the model. The adjusted model included intentionality, age, injury severity, pain score, LOS, insurance category, substance use history, substance overuse history, and year of study. Multiple imputation by chained equations (MICE) was used to account for missing data, as 32 participants (5.1%) had missing injury severity data, 66 participants (10.6%) had missing pain score data, and 30 (4.8%) had missing LOS data. To check the validity of this approach, we then compared MI results to those obtained with complete case (CC) analysis.
Results
Participants who received a discharge opioid prescription differed significantly from those who did not receive a discharge opioid prescription (Table 1). Participants with intentional injuries were more likely to receive opioids than patients with unintentional injuries. Receipt of a discharge opioid prescription was more likely when injuries were penetrating and more severe, and for participants with higher pain ratings in the 24 hours before discharge, longer LOS, and younger age. Participants who reported alcohol overuse were less likely to receive opioids.
Table 1:
Descriptive Statistics
| Receipt of Opioids | ||||
|---|---|---|---|---|
| All Patients n=623 | Received Opioids n=481 | Did Not Receive Opioids n=142 | P Value | |
| Characteristic | N (%) | N (%) | N (%) | |
| Age, mean (SD) | 35.63 (14.89) | 34.51 (14.00) | 39.43 (17.08) | .002 |
| Insurance Type | .059 | |||
| Private | 69 (11.08) | 60 (12.47) | 9 (6.34) | |
| Public | 200 (32.10) | 146 (30.35) | 54 (38.03) | |
| Third Party | 61 (9.79) | 42 (8.73) | 19 (13.38) | |
| Self-Pay | 261 (41.89) | 208 (43.24) | 53 (37.32 | |
| Unknown | 32 (5.14) | 25 (5.20) | 7 (4.93) | |
| Injury Type | <.001 | |||
| Blunt | 329 (52.89) | 235 (48.86) | 94 (66.20) | |
| Penetrating | 293 (47.11) | 246 (51.14) | 47 (33.10) | |
| Injury Severity Score (ISS) | <.001 | |||
| Mild (ISS<9) | 293 (47.03) | 197 (40.96) | 96 (67.61) | |
| Moderate (ISS≥9 & ≤15) | 211 (33.87) | 192 (39.92) | 19 (13.38) | |
| Severe (ISS≥16) | 87 (13.96) | 71 (14.76) | 16 (11.27) | |
| No data | 32 (5.14) | 21 (4.37) | 11 (7.75) | |
| Highest Pain Score | <.001 | |||
| Mild (≤4) | 58 (9.31) | 30 (6.24) | 28 (19.72) | |
| Moderate (>4 & ≤7) | 180 (28.89) | 145 (30.15) | 35 (24.65) | |
| Severe (>7) | 319 (51.20) | 290 (60.29) | 29 (20.42) | |
| No data | 66 (10.59) | 16 (3.33) | 50 (35.21) | |
| Length of Stay | <.001 | |||
| 1–3 days | 263 (42.22) | 168 (34.93) | 95 (66.90) | |
| 4–7 days | 193 (30.98) | 166 (34.51) | 27 (19.01) | |
| 8–14 days | 97 (15.57) | 89 (18.50) | 8 (5.63) | |
| >14 days | 40 (6.42) | 34 (7.07) | 6 (4.23) | |
| No data | 30 (4.82) | 24 (4.99) | 6 (4.23) | |
| Reported substance use in past 6 mos. | 476 (76.53) | 369 (76.72) | 107 (75.35) | .737 |
| Alcohol | 374 (60.03) | 285 (59.25) | 89 (62.68) | .464 |
| Marijuana | 215 (34.51) | 174 (36.17) | 41 (28.87) | .108 |
| Opioids | 27 (4.33) | 24 (4.99) | 3 (2.11) | .139 |
| Cocaine | 30 (4.82) | 26 (5.41) | 4 (2.82) | .206 |
| Reported overuse in past 6 mos. | 157 (25.20) | 110 (22.87) | 47 (33.10) | .014 |
| Alcohol | 95 (15.25) | 58 (12.06) | 37 (26.06) | <.001 |
| Marijuana | 45 (7.22) | 37 (7.69) | 8 (5.63) | .405 |
| Opioids | 5 (0.80) | 5 (1.04) | 0 (0) | .223 |
| Cocaine | 12 (1.93) | 11 (2.29) | 1 (0.70) | .228 |
| Positive Urine Drug Screen* | ||||
| Marijuana | 113 (40.65) | 86 (43.43) | 27 (33.75) | .137 |
| Opioids | 70 (25.18) | 55 (22.78) | 15 (18.75) | .116 |
| Intent of Injury | .008 | |||
| Unintentional | 277 (44.46) | 200 (41.58) | 77 (54.23) | |
| Intentional | 346 (55.54) | 281 (58.42) | 65 (45.77) |
Due to missing data, totals may not equal 623. T-tests were used for continuous variables and chi-square tests were used for categorical variables.
Urine drug screen was performed on 278 patients in the sample.
Discharge Opioid Prescriptions
While injury intent was significant in the bivariate model (participants with intentional injuries were more likely to receive opioids at discharge OR=1.66, 95% CI 1.14–2.43), intent was no longer significant in the adjusted model. Substance use was not significant in bivariate or adjusted models. In the adjusted model, substance overuse, injury severity, pain, year of study, and LOS were significant (Table 2). Substance overuse was associated with lower odds of receiving opioids at discharge (OR=0.50, 95% CI 0.26–0.81). Moderate injuries were associated with significantly higher odds of receiving opioids at discharge (OR=4.43, 95% CI 2.30–8.52). Increasing pain scores were also associated with significantly higher odds of receiving opioids at discharge (moderate pain: OR=4.55, 95% CI 2.27–9.11; severe pain: OR=13.33, 95% CI 6.45–27.55). Increasing LOS was associated with higher odds of receiving opioids at discharge (LOS 4–7 days: OR=3.34, 95% CI 1.86–5.99; LOS 8–14 days: OR=6.84, 95% CI 2.58–18.12; LOS > 14 days: OR=4.27, 95% CI 1.39–13.13). Patients discharged in the year 2014 were less likely to receive opioids than those discharged in 2013 (OR=0.56, 95% CI 0.31–0.99). These results differed slightly from the complete case (CC) analysis; in the CC, LOS >14 days was not significantly associated with increased odds of opioid receipt.
Table 2:
Odds of Receipt of Opioids at Discharge
| Bivariate | Adjusted Model | |||||
|---|---|---|---|---|---|---|
| Odds Ratio | Standard Error | 95% CI | Odds Ratio | Standard Error | 95% CI | |
| Intentional Injury | 1.66 | 0.32 | (1.14, 2.43)b | 1.17 | 0.33 | (0.67, 2.05) |
| Age | 0.98 | 0.01 | (0.97, 0.99)c | 0.99 | 0.01 | (0.98, 1.01) |
| ISS | ||||||
| (Ref group: minor) Moderate (≥9 & ≤15) | 4.92 | 1.33 | (2.90, 8.37)d | 4.43 | 1.48 | (2.30, 8.52) c |
| Severe (≥16) | 2.16 | 0.66 | (1.19, 3.91)b | 0.82 | 0.36 | (0.35, 1.92) |
| Pain Rating | ||||||
| (Ref group: minor) Moderate (>4 & ≤7) | 3.87 | 1.25 | (2.05, 7.29)d | 4.55 | 1.61 | (2.27, 9.11) c |
| Severe (>7) | 9.33 | 3.05 | (4.92, 17.72)d | 13.33 | 4.92 | (6.45, 27.55) c |
| Length of Stay | ||||||
| (Ref group: 1–3 days) 4–7 days | 3.48 | 0.85 | (2.15, 5.61)c | 3.34 | 0.99 | (1.86, 5.99) c |
| 8–14 days | 6.29 | 2.46 | (2.92, 13.53)c | 6.84 | 3.40 | (2.58, 18.12) c |
| >14 days | 3.20 | 1.48 | (1.30, 7.91)a | 4.27 | 2.45 | (1.39, 13.13) a |
| Insurance | ||||||
| (Ref group: private) Public | 0.41 | 0.16 | (0.19, 0.87)a | 0.43 | 0.20 | (0.17, 1.08) |
| Third-Party | 0.33 | 0.15 | (0.14, 0.80)a | 0.54 | 0.29 | (0.19, 1.57) |
| Self-Pay | 0.59 | 0.23 | (0.27, 1.26) | 0.67 | 0.31 | (0.27, 1.67) |
| Unknown | 0.54 | 0.30 | (0.18, 1.60) | 0.62 | 0.44 | (0.16, 2.47) |
| Reported substance use in past 6 months (Ref group: no reported use) | 1.08 | 0.24 | (0.70, 1.67) | 1.19 | 0.36 | (0.65, 2.16) |
| Reported overuse in past 6 months (Ref group: no reported overuse) | 0.60 | 0.13 | (0.40, 0.90)a | 0.50 | 0.133 | (0.26, 0.81) a |
| Year of Study | ||||||
| (Ref group: 2013) 2014 | 0.60 | 0.14 | (0.38, 0.95)a | 0.56 | 0.17 | (0.31, 0.99) a |
| 2015 | 1.71 | 0.53 | (0.93, 3.14) | 1.46 | 0.56 | (0.68, 3.11) |
| 2016 | 1.31 | 0.40 | (0.72, 2.37) | 1.28 | 3.40 | (0.61, 2.68) |
| 2017 | 0.56 | 0.32 | (0.18, 1.73) | 0.37 | 0.26 | (0.09, 1.47) |
Note: p<.05,
p<.01,
p<.001,
p<.0001;
significant values are bolded in adjusted column
Opioid Dose at Discharge
Four hundred and eighty-one participants received opioids at discharge. In bivariate models, increasing injury severity, pain severity and LOS were associated with higher discharge dosages. Increasing age and discharge in the year 2016 were associated with lower discharge dosages (Table 3). In the adjusted model, increasing age was associated with decreased discharge dosages (β=−0.39, 95% CI −0.75- −0.21). Severe pain was associated with increased discharge dosages (β=30.76, 95% CI 10.83–50.69), as was increasing LOS (LOS 8–14 days: β=29.02, 95% CI 14.48–43.55; LOS>14 days: β=23.92, 95% CI 2.30–45.54). In addition, patients discharged during the year 2016 received lower opioid dosages than those discharged during 2013 (β= −17.60, 95% CI −31.80- −3.40) (Figure 1). Results obtained using complete case analysis did not differ.
Table 3:
Change in Morphine Milligram Equivalents (MMEs)
| Bivariate | Adjusted Model | |||||
|---|---|---|---|---|---|---|
| β | Std. Error | 95% CI | β | Std. Error | 95% CI | |
| Intentional Injury | 5.68 | 5.01 | (−4.17, 15.53) | 0.08 | 5.71 | (−11.14, 11.29) |
| Age | −0.37 | 0.18 | (−0.72, −0.25)a | −0.39 | 0.19 | (−0.75, −0.21)a |
| ISS | ||||||
| (Ref group: minor) Moderate (≥9 & ≤15) | 10.31 | 5.46 | (−0.42, 21.04) | 6.00 | 5.57 | (−5.00, 16.91) |
| Severe (≥16) | 20.95 | 7.45 | (6.30, 35.60)b | 5.41 | 8.28 | −10.87, 21.69) |
| Pain Rating | ||||||
| (Ref group: minor Moderate (>4 & ≤7) | 12.74 |
10.75 | (−8.39, 33.86) | 13.40 | 10.55 | (−7.35, 34.14) |
| Severe (>7) | 28.46 | 10.28 | (8.26, 48.67)b | 30.76 | 10.14 | (10.83, 50.69) b |
| Length of Stay | ||||||
| (Ref group: 1–3 days) 4–7 days | 6.72 | 5.82 | (−4.71, 18.14) | 7.16 | 5.97 | (−4.57, 18.89) |
| 8–14 days | 31.41 | 6.97 | (17.72, 45.09)c | 29.02 | 7.39 | (14.48, 43.55) c |
| >14 days | 22.04 | 10.00 | (2.40, 41.67)a | 23.92 | 10.99 | (2.30, 45.54) a |
| Insurance | ||||||
| (Ref group: private) Public | −6.86 | 8.33 | (−23.23, 9.50) | −3.23 | 8.30 | (−19.47, 13.01) |
| Third-Party | −9.22 | 10.93 | (−30.70, 12.25) | −5.16 | 10.99 | (−26.56, 16.25) |
| Self-Pay | −2.35 | 7.96 | −18.00, 13.30) | −3.22 | 8.04 | (−19.10, 12.57) |
| Unknown | 6.34 | 12.93 | (−19.07, 31.75) | 8.61 | 13.11 | (−17.14, 34.38) |
| Reported substance use in past 6 months (Ref group: no reported use) | 0.82 | 5.86 | (−10.68, 12.33) | −0.44 | 6.21 | (−12.64, 11.75) |
| Reported overuse in past 6 months (Ref group: no reported overuse) | −1.03 | 5.89 | (−12.61, 10.55) | −3.53 | 6.25 | (−15.82, 8.75) |
| Year of Study | ||||||
| (Ref group: 2013) 2014 | −0.96 | 6.56 | (−13.85, 11.92) | −1.05 | 6.45 | (−13.74, 11.63) |
| 2015 | −1.39 | 6.81 | (−14.77, 12.00) | −5.67 | 6.89 | (−19.21, 7.87) |
| 2016 | −15.59 | 7.16 | (−29.66, −1.52)a | −17.60 | 7.23 | (−31.80, −3.40)a |
| 2017 | −10.94 | 17.63 | (−45.58, 23.71) | −19.71 | 18.01 | (−55.11, 15.54) |
Note: p<.05,
p<.01,
p<.001;
significant values are bolded in adjusted column
Figure 1:
Effect of Time on Morphine Milligram Equivalent (MME) Dosages
Discussion
The major findings of this study in light of our hypotheses are mixed. The hypothesis that men with intentional injuries would be less likely to receive opioids at discharge was not supported. However, the hypothesis that men who report a history of substance overuse would be less likely to receive opioids at discharge was supported. Characteristics suggestive of higher levels of pain, such as increasing injury severity and pain scores in the 24 hours prior to discharge, along with longer LOS, were associated with greater likelihood of discharge opioid prescription. Increased pain severity and longer LOS were associated with higher doses of opioids at discharge. Overall, these results may suggest that providers treating this group of Black men prioritized patients’ pain level, severity of injury, LOS, and substance overuse history when deciding to prescribe or not prescribe opioids at hospital discharge. In addition, the prescribing practices of providers treating the participants in our study did not seem impacted by patient self-reported substance use history not characterized as overuse, although it is not clear that providers were aware of their patients’ substance use histories.
We are unaware of any other studies exploring the impact of injury intent on receipt of opioids at discharge. While previous qualitative studies have described ways in which trauma patients perceive that they are stigmatized on the basis of their injury intent[33,20,21], the results of our analysis did not support the impact of injury-intent based biases on pain treatment decisions. This may be due in part to the increasing protocolization of pain care, particularly since the emergence of the opioid crisis[34]. It is also possible that providers at the hospitals where the parent study took place treat so many intentional injuries that they do not view these patients differently than those injured unintentionally.
As discussed earlier, it is possible that providers were not aware of patients’ substance use histories, as these self-reported data were collected by the study team during the parent study interviews. If providers were aware of their patients’ substance use histories, our results reflect that providers considered potential substance overuse and misuse as an important but not sole concern when creating pain treatment discharge plans. It is important to note that substance use and misuse/addiction are distinct phenomena, as individuals who use substances may not engage in behaviors that qualify as problematic. While the use of illicit substances is generally not viewed favorably in healthcare environments, some providers appropriately differentiate between use and addiction and may not have the same concerns about patients who report using substances versus those with problematic use history (i.e., self-described overuse). The only substance in this study that had an impact on receipt of opioids (when reported as overuse) was alcohol, however this is likely due to the small numbers of patients who reported overusing substances other than alcohol. This aligns with other studies that have identified that patients are more likely to underreport use of substances generally regarded as more harmful or criminal (e.g. opioids vs. marijuana or alcohol)[35,36]. Providers likely understand that the disease of addiction often generalizes across substance, such that patients who misuse alcohol are more likely to misuse opioids and other substances[37].
Although much of the media attention to the opioid crisis has highlighted cases of “drug naïve” individuals who become addicted to opioids after receiving a prescription, nearly 90% of patients who develop what is often labelled an iatrogenic opioid use disorder report a history of psychoactive substance use prior to first using opioids[37]. A recent study exploring provider decision-making related to pain treatment found that emergency medicine physicians consider past or present substance use when making pain treatment decisions[38]. In addition, the use of prescription drug monitoring programs (PDMPs) has increased since the early 2000’s and may reduce opioid prescribing by allowing providers to determine if patients are receiving multiple prescriptions for controlled substances[39].
Our analysis indicated a trend towards decreasing MME dosages across all patients in the over time. Patients discharged in 2016 received significantly lower dosages than in 2013. No participants received opioid dosages greater than 200 MME after 2015. The pattern of decreasing MME dosages seen in this study reflects national trends in opioid prescribing; opioid prescriptions have been decreasing in both number of prescriptions per capita and MME dosage per prescription since 2012, with a 20% reduction in MME/day dosages per capita in the US between the years 2015–2017[40]. These trends may also have been impacted by the release of the 2016 CDC guidelines which recommend 90 MME daily limits for most patients with chronic pain, and a limit on the duration of prescriptions for acute pain to 7 days or less[41].
This study had several limitations. First, comparing the treatment of patients with penetrating versus blunt injuries is difficult, as these injuries are medically quite different. In the group of unintentionally injured patients, the number of patients with penetrating injuries was very small (n=9). The number of intentionally injured patients with blunt injuries was larger (n=61), but intentionality and mechanism were closely related in this sample. Second, our evaluation of the impact of substance use history on receipt of opioids at discharge was limited by how we measured substance use. Due to lack of access to any other way to measure this variable, we had to use data that were collected during the parent study and were not necessarily the same as what was disclosed by participants to providers or noted in patients’ EHR. We acknowledge the limitation in assuming that providers were aware of participants’ substance use histories when making their discharge pain treatment decisions. In addition, participants may have underreported substance use to parent study staff in an attempt to provide socially desirable responses or because they were not comfortable sharing this information.
We were also limited by the inclusion/exclusion criteria of the parent study. Due to increased risk of substance use disorders among people with depression or PTSD[42], the parent study exclusion of patients currently receiving treatment for these conditions may have resulted in lower rates of substance use/overuse in this sample, and also limits generalizability of the findings. In addition, the parent study did not include individuals in police custody, as these individuals are considered a vulnerable population. The exclusion of this population is significant as individuals in police custody may be regarded and treated very differently by healthcare providers than other patients. Finally, the parent dataset only included patients from two trauma centers in one northeastern US city where gun violence is extremely common. It is unclear if intent of injury might have a stronger impact on pain treatment in a setting where providers are less experienced and/or comfortable treating patients with gunshot wounds.
Future studies should explore the potential impact of injury intent and substance use on receipt of opioids in racially and geographically diverse samples. Because a large number of intentional injuries in our sample were gunshot wounds, we are interested in whether results would differ at hospitals in areas where gunshot wounds are less common or areas where gun ownership is widely accepted and perceived as associated with legal ownership for the purposes of recreation or hunting. Intentionality of injury might have less of an impact in patients who already experience bias and are less likely to receive pain treatment due to their race, and we are curious if injury intent impacts pain treatment in patients of other racial groups. We acknowledge that assumed criminality of patients may impact care in subtle and pernicious ways that are not reflected in simple quantitative data about receipt of opioid prescriptions, and do not believe that our results suggest that healthcare environments are free from damaging stigma. As previously discussed, qualitative work in this area reveals that intentionally injured Black patients report feeling stigmatized in healthcare settings, which can impact the patient-provider relationship and patient trust in the healthcare system[33,20,21,22]. Continued research investigating health outcomes as well as qualitative studies exploring patient experiences and provider biases can help fill this knowledge gap. Finally, future studies should more explicitly study how providers integrate knowledge of substance use histories into their discharge pain treatment plans, using substance use history data that is self-reported directly to providers and/or substance use history diagnosis codes.
Conclusion
In a cohort of injured Black men hospitalized in a large northeastern US city between the years of 2013–2017, receipt of discharge opioid prescriptions was not impacted by intentionality of injury but by characteristics associated with increased levels of pain such as severity of injury, severity of pain, and length of hospital stay. In addition, men who reported substance overuse were less likely to receive opioids at discharge. Providers prescribed lower dosages of opioids over time, which mirrors national trends in opioid prescribing. While previous research highlights the stigmatization experienced by intentionally injured patients in their interactions with the healthcare system[20,21], our hypothesis that injury intent would impact receipt of discharge opioid prescriptions was not upheld.
Funding:
This work was supported by the Robert Wood Johnson Foundation Future of Nursing Scholars Fellowship; and by the National Institute of Nursing Research of the National Institutes of Health (grant R01NR013503 [Dr. Richmond]).
Footnotes
Conflicts of interest/Competing interests: None.
References
- 1.Celso B, Tepas J, Langland-Orban B, et al. A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma Inj Infect Crit Care. 2006;60(2):371–378. [DOI] [PubMed] [Google Scholar]
- 2.Gruen RL, Gabbe BJ, Stelfox HT, Cameron PA. Indicators of the quality of trauma care and the performance of trauma systems. Br J Surg. 2012;99:97–104. [DOI] [PubMed] [Google Scholar]
- 3.Holmes A, Williamson O, Hogg M, Arnold C, O’Donnell M. Determinants of chronic pain 3 years after moderate or serious injury. Pain Med. 2013;14(3):336–344. [DOI] [PubMed] [Google Scholar]
- 4.Velmahos CS, Herrera-Escobar JP, Al Rafai SS, et al. It still hurts! Persistent pain and use of pain medication one year after injury. Am J Surg. 2019;In Press. [DOI] [PubMed] [Google Scholar]
- 5.Clarke JL, Skoufalos A, Scranton R. The American opioid epidemic: Population health implications and potential solutions. Report from the National Stakeholder Panel. Popul Health Manag. 2016;19:S1–S10. [DOI] [PubMed] [Google Scholar]
- 6.Haytham K, Weil E, Wakeman S, Ring D. The opioid epidemic and new legislation in Massachusetts: Time for a culture change in surgery? Ann Surg. 2017;265(4):731–733. [DOI] [PubMed] [Google Scholar]
- 7.Anderson K, Green CR, Payne R. Racial and ethnic disparities in pain: Causes and consequences of unequal care. J Pain. 2009;10(12):1187–1204. [DOI] [PubMed] [Google Scholar]
- 8.Meghani SH, Byun E, Gallagher RM. Time to take stock: a meta-analysis and systematic review of analgesic treatment for pain in the United States. Pain Med. 2012;13(2):150–174. [DOI] [PubMed] [Google Scholar]
- 9.Substance Abuse and Mental Health Services Administration. Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings. US Department of Health and Human Services; 2014. [Google Scholar]
- 10.Staton LJ, Panda M, Chen I, et al. When race matters: Disagreement in pain perception between patients and their physicians in primary care. J Natl Med Assoc. 2007;99(5):532–538. [PMC free article] [PubMed] [Google Scholar]
- 11.Aronowitz S, McDonald C, Stevens R, Richmond T. Mixed Studies Review of Factors Influencing Receipt of Pain Treatment by Injured Black Patients. J Adv Nurs. 2020;(76):34–46. [DOI] [PubMed] [Google Scholar]
- 12.Muhammad KG. Where did all the White criminals go?: Reconfiguring race and crime on the road to mass incarceration. Souls. 2011;13(1):72–90. [Google Scholar]
- 13.Oliver MB. African American men as “criminal and dangerous”: Implications of media portrayals of crime on the “criminalization” of African American men. J Afr Am Stud. 2003;7(2):3–18. [Google Scholar]
- 14.Owusu-Bempah A Race and policing in historical context: Dehumanization and the policing of Black people in the 21st century. Theor Criminol. 2017;21(1):23–34. [Google Scholar]
- 15.Muhammad KG. The Condemnation of Blackness: Race, Crime, and the Making of Modern Urban America. Harvard University Press; 2011. [Google Scholar]
- 16.Caron C A Black Yale Student Was Napping, and a White Student Called the Police. The New York Times. https://www.nytimes.com/2018/05/09/nyregion/yale-black-student-nap.html. Published May 9, 2018. [Google Scholar]
- 17.Stevens M Starbucks C.E.O. Apologizes After Arrests of 2 Black Men. The New York Times. https://www.nytimes.com/2018/04/15/us/starbucks-philadelphia-black-men-arrest.html?action=click&module=inline&pgtype=Article. Published April 15, 2018. [Google Scholar]
- 18.Taylor OR. Even in Oakland, calling the cops on black people just living their lives. San Francisco Chronicle. https://www.sfchronicle.com/news/article/Even-in-Oakland-calling-the-cops-on-black-people-12920652.php. Published May 17, 2018. [Google Scholar]
- 19.Jacoby SF, Richmond TS, Holena DN, Kaufman EJ. A safe haven for the injured? Urban trauma care at the intersection of healthcare, law enforcement, and race. Soc Sci Med. Published online 2017:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Liebchutz J, Schwartz S, Hoyte J, et al. A chasm between injury and care: Experiences of Black male victims of violence. J Trauma. 2010;69(6):1372–1378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Patton D, Sodhi A, Affinati S, Lee J, Crandall M. Post-discharge needs of victims of gun violence in Chicago: A qualitative study. J Interpers Violence. 2019;34(1):135–155. [DOI] [PubMed] [Google Scholar]
- 22.Rich J Wrong Place, Wrong Time: Trauma and Violence in the Lives of Young Black Men. Johns Hopkins University Press; 2009. [Google Scholar]
- 23.Alexander M The New Jim Crow. The New Press; 2010. [Google Scholar]
- 24.Netherland J, Hansen H. White opioids: Pharmaceutical race and the war on drugs that wasn’t. Biosocieties. 2017;12(2):217–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Richmond TS, Wiebe DJ, Reilly PM, Rich J, Shults J, Kassam-Adams N. Contributers to Postinjury Mental Health in Urban Black Men with Serious Injuries. JAMA Surg. 2019;154(9):836–843. doi:doi: 10.1001/jamasurg.2019.1622 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Pennsylvania Trauma Systems Foundation. Pennsylvania Trauma Systems Foundation Operational Manual for the Data Base Collection System.; 2020:1–211. http://ptsf.org/upload/2020_July_update_PTOS_Manual.pdf [Google Scholar]
- 27.Agency Medical Directors’ Group. Opioid dose calculator. Published 2015. http://www.agencymeddirectors.wa.gov/calculator/dosecalculator.htm
- 28.Centers for Disease Control. Calculating total daily dose of opioids for safer dosage. Published April 13, 2018. https://www.cdc.gov/drugoverdose/pdf/calculating_total_daily_dose-a.pdf
- 29.Committee on Injury Scaling. The Abbreviated Injury Scale, 1990 Revision, Update 98. American Association of Automotive Medicine; 1998. [Google Scholar]
- 30.Baker SP, O’Neill B, Haddon W, Long WB. The injury severity score: A method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1947;14(3):187–196. [PubMed] [Google Scholar]
- 31.Brasel KJ, deRoon-Cassini T, Bradley CT. Injury severity and quality of life: whose perspective is important? J Trauma. 2010;68(2):263–268. [DOI] [PubMed] [Google Scholar]
- 32.Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2018 National Survey on Drug Use and Health. Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration; 2019. https://www.samhsa.gov/data/ [Google Scholar]
- 33.Jacoby S An Ethnographic Study of the Experience of Black Traumatically Injured Patients at a Trauma Center in Philadelphia. Published online 2015. [Google Scholar]
- 34.American Hospital Association. Stem the tide: Addressing the opioid epidemic. Published 2017. www.aha.org/opioidtoolkit [Google Scholar]
- 35.Bone C, Gelberg L, Vahidi M, Leake B, Yacenda-Murphy J, Anderson RM. Under-reporting of risky drug use among primary care patients in Federally Qualified Health Centers. J Addict Med. 2016;10(6):387–394. [DOI] [PubMed] [Google Scholar]
- 36.Garg M, Garrison L, Leeman L, et al. Validity of self-reported drug use information amoung pregnant women. Matern Child Health J. 2016;20:40–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cicero TJ, Ellis MS, Kasper ZA. Psychoactive substance use prior to the development of iatrogenic opioid abuse: A descriptive analysis of treatment-seeking opioid abusers. Addict Behav. 2017;65:242–244. [DOI] [PubMed] [Google Scholar]
- 38.Sinnenberg LE, Wanner KJ, Perrone J, Barg FK, Rhodes KV, Meisel ZF. What factors affect physicians’ decisions to prescribe opioids in emergency departments? Soc Med Decis Marking. 2017;2(1). doi: 10.1177/2381468316681006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bao Y, Pan Y, Taylor A, et al. Prescription drug monitoring programs are associated with sustained reductions in opioid prescribing by physicians. Health Aff (Millwood). 2016;35(6):1045–1051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Guy GP, Zhang K, Schieber LZ, Young R, Dowell D. County-level opioid prescribing in the United States, 2015 and 2017. J Am Med Assoc. 2019;179(4):574–576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain -- United States 2016. Centers for Disease Control; 2016:1–49. https://www.cdc.gov/mmwr/volumes/65/rr/rr6501e1.htm [DOI] [PubMed] [Google Scholar]
- 42.Davis L, Akihito U, Newell J, Frazier E. Major depression and comorbid substance use disorders. Curr Opin Psychiatry. 2008;21(1):14–18. [DOI] [PubMed] [Google Scholar]