Abstract
Introduction: Prescription psychotherapeutic medication misuse is a growing problem in the United States, but no method exists to routinely screen for this in primary care. Our study sought to (1) describe the prevalence of prescription psychotherapeutic medication misuse in primary care and the characteristics of patients who misuse and (2) compare 2 screening instruments modified to identify prescription medication misuse in primary care.
Methods: Primary care patients from underserved, urban clinics within a health system completed anonymous computer-directed health screens that included standard questions about prescription medication misuse. They were also administered the 4-item Cut down, Annoyed, Guilty, and Eye-opener questionnaire modified to focus on prescription medications (RxCAGE) and a 6-item Prescription Opioid Misuse Index (POMI-e) expanded to include other prescription medications.
Results: Of 2,339 respondents, 15.3% were positive for at least 2 items on the RxCAGE and 18.6% were positive for at least 2 items on the POMI-e. Using our computer-directed health screen as a comparison, we found that POMI-e had a higher area under the curve (0.63). A positive POMI-e was associated with being male, white and unemployed, having depression and anxiety, and currently using illicit substances, smoking, and misusing alcohol.
Conclusions: Rates of prescription medication misuse were substantial with both RxCAGE and POMI-e showing promise as screening instruments. Future studies are needed to test prescription medication misuse screening tools in broader populations and pilot interventions for those screening positive.
In 2016, an estimated 115 million Americans or 42.8% of the adult population used prescription psychotherapeutic drugs in the past year. Of those, an estimated 18.6 million or 16.2% misused prescription psychotherapeutic drugs.1⇓–3 Misuse is defined by the National Survey on Drug Use and Health (NSDUH) as medication use in any way not directed by a prescribing clinician, including using another's prescription, greater amounts, more often, or longer than directed.4 Current attention has focused heavily on the growing opioid epidemic.2,5⇓–7 As prescription opioid misuse continues to grow each year, overdose deaths from prescription opioids, the most commonly misused prescription drug, have almost quadrupled in the past decade and are paralleled by a similar increase in emergency department visits, falls, and fractures.8 Furthermore, the misuse of prescription opioids is a significant risk factor for future or concurrent heroin use.9,10
In addition to prescription opioid misuse, misuse of tranquilizers (defined as benzodiazepines and muscle relaxants), stimulants (such as amphetamines), and sedatives is growing, albeit largely unrecognized.11⇓⇓⇓⇓–16 Although it is well known that tranquilizers, stimulants, and sedatives can have substantial negative effects that include fatal overdose in some cases, they have not been well-studied to date.13,16,17
Although recognizing prescription opioid misuse as a problem in the patients they are serving,18 primary care clinicians continue contributing to the growing problems with misuse by writing 45% of all opioid prescriptions19 and substantial proportions of benzodiazepines and other psychotherapeutic drugs.20,21 A few studies have documented the prevalence of prescription psychotherapeutic medication misuse in general primary care or community pharmacy settings, but studies to date have not documented the characteristics of patients who are likely to misuse prescription psychotherapeutic drugs and the prevalence of misuse in predominantly underserved primary care practices.22,23 Our study aimed to (1) describe the characteristics of primary care patients who misuse prescription psychotherapeutic agents and (2) compare rates of misuse obtained from 2 standardized screening tools modified to focus on prescription medication misuse.
Methods
Participants and Data Collection
We conducted a cross-sectional survey of patients in underserved, urban primary care clinics within a university-based health system. Clinics were located both adjacent to the hospital or at satellite, community sites. Patients were recruited in clinic waiting areas for an anonymous survey focused on health behaviors. Patients between the ages of 18 and 70 years who could understand English were eligible for the study. Interested participants were escorted to a private area adjacent to the clinic waiting room either before or after their scheduled medical appointments. After providing verbal consent, participants completed the 15-minute Health Cheq survey on a tablet computer, guided by a 3-D avatar (Peedy the Parrot) who read each question aloud and kept participants engaged in the process.24,25 This was followed by a brief research assistant-administered survey. Patients received $10 for their participation in the study. Data were collected between July 2012 and December 2013. This study was approved by the Virginia Commonwealth University Institutional Review Board.
Outcome Measures
Health Cheq collected information on patient demographics, general health behaviors, general medical concerns, mental health, and psychosocial issues. It also included 4 questions about prescription drug misuse in the past month: (1) taking more pills than prescribed, (2) taking pills more often than prescribed, (3) using medicine prescribed for someone else, or (4) getting the same medicine from more than 1 doctor. These questions were considered standard at the time and were based on the NSDUH questionnaire. Health Cheq also collected the standard Cut down, Annoyed, Guilty, and Eye-opener (CAGE) questionnaire for alcohol if male26 or the Tolerance, Annoyance, Cut down, and Eye-opener (TACE) questionnaire if female,27 as well as the CAGE Adapted to Include Drugs (CAGE-AID) for all illicit drugs.28 Research assistants then screened patients in person using 2 prescription drug misuse screening instruments: (1) prescription medication CAGE (RxCAGE), a 4-item screener, using CAGE alcohol use disorder items modified to focus on prescription drugs26 and (b) Prescription Opioid Misuse Index (POMI-e), expanded from the original focus on prescription opioid misuse to include all prescription drugs.29
Data Analysis
Using SPSS v22, item frequencies and scale scores for the 2 screening instruments were calculated. Sensitivity and specificity were calculated, along with a receiver operating characteristic curve analysis to calculate area under the curve (AUC), for cutoff scores of 1 and 2 in both RxCAGE and POMI-e. The anonymous Health Cheq questions that were standard care at the time were used for comparison, because they were computer administered. McNemar tests were then used to test for associations between the screening instrument and cutoff score with the highest AUC with demographic, medical, mental health, and substance use characteristics, as identified in Health Cheq.
Results
Respondent Characteristics
A total of 2,339 patients participated in the study (Table 1). The mean age was 46, with 61.7% of respondents who were female and 70.5% identified as African American. A total of 79.4% were uninsured with 71.6% of those receiving care through the health system's indigent care program.
Demographics of Respondents to Health Cheq Survey (n = 2,339)
Screening Tools
On the RxCAGE, 37.8% answered yes to at least 1 item and 15.3% answered yes to 2 or more items (Table 2). Feeling the need to cut down on the use of prescription drugs was the most frequently endorsed (22.6%), followed by ever using prescription drugs to get going or calm down (21.2%). On the POMI-e, 36.2% answered yes to at least 1 item and 18.6% to 2 or more items (Table 2). Ever feeling high or getting a buzz after using medication was the most frequently positive in POMI-e (16.7%), followed by using medication more frequently than prescribed (15.7%). When comparing RxCAGE and POMI-e as well as cutoff scores of 1 or higher and 2 or higher, POMI-e with a cutoff score of 2 or higher had the largest area under the curve (AUC) (0.63, P < .001) (Table 3).
Positive Screens from Prescription Drug Misuse Screening Instruments
Sensitivity, Specificity, and ROC curve analyses for Rx-CAGE and POMI-e
Associations with Prescription Drug Misuse
The POMI-e was used to characterize associations because it had the largest AUC with a cutoff score of 2 or more (hereafter referred to as a “positive POMI-e”). Multiple demographic characteristics along with medical, mental health, and substance use comorbidities were associated with a positive POMI-e (Table 4). Demographic characteristics associated with a positive POMI-e included being younger, male, white, and unemployed. Although arthritis was the only chronic condition with POMI-e positive and negative group differences (44.8% vs 35.9% respectively, P = .001), POMI-e positive patients had a higher number of chronic diseases than POMI-e negative patients (1.62 vs 1.49 respectively, P = .037).
Participant Characteristics of Those at Risk for Prescription Misuse Based on POMI-e Scores of ≥2
Both depression and anxiety independently were associated with a positive POMI-e. Likewise, lifetime cigarette use and current smoking status were associated with a positive POMI-e. Those who had used illicit drugs in the past 30 days as well as those who screened positive in the CAGE-AID were more likely to have a positive POMI-e. Furthermore, positive alcohol TACE for women, CAGE for men, or a positive CAGE-AID were more likely for the positive POMI-e group as compared with the negative POMI-e group (64.1% vs 32.2% respectively, P < .001).
Discussion
Substantial rates of prescription psychotherapeutic medication misuse were identified in this urban primary care population. The prevalence of prescription medication misuse was similar for the 2 instruments, but POMI-e had a higher AUC than RxCAGE. Given the high association of prescription drug misuse with mental health and substance use comorbidities, patients identified with prescription drug misuse might merit special attention in primary care. Routine screening for prescription drug misuse might have a substantial impact in reducing misuse if clinicians had evidence-based interventions for those who screened positive.
Future studies to validate and test implementation of the screening tools could have substantial impact on the way clinicians identify patients with prescription psychotherapeutic drug misuse. Furthermore, successful interventions to treat patients identified with misuse and at risk of medication harm will need to be tested in primary care settings. To be successfully implemented, the development of validated screening tools and evidence-based interventions should be combined with clinician education in prescription drug misuse.
Several limitations exist. First, participants were all from urban, underserved primary care clinics, and only English speakers were included, which may limit the generalizability to the general population. This may be mitigated by the large sample size. Second, those who have more difficulty with using a computer-administered screening tool or have poor literacy may not have answered the questions or declined participation in the study. Third, there may be a response bias because a research assistant administered instruments with patients knowing that results would not be shared with their clinician. Patients may respond differently if these screening tools were administered in person by their clinicians.
Conclusions
Both RxCAGE and POMI-e show promise as screening tools for prescription psychotherapeutic medication misuse and warrant further study. Developing and implementing validated screening tools for prescription psychotherapeutic medication misuse may help primary care clinicians identify and care for patients with prescription drug misuse and reduce the morbidity, societal impacts, and costs of the current prescription drug misuse epidemic.
Notes
This article was externally peer reviewed.
Funding: Funding for this study was provided by the National Institute on Drug Abuse (RO1 DA026091).
Conflict of interest: none declared.
To see this article online, please go to: http://jabfm.org/content/32/2/272.full.
- Received for publication June 12, 2018.
- Revision received November 5, 2018.
- Accepted for publication November 7, 2018.
