Clinical practice guidelines recommend against co-prescribing of opioids and benzodiazepines due to increased risk of over-sedation and respiratory depression, particularly when it occurs in combination with alcohol and illicit drugs. Additional risks include falls in older Veterans, intentional overdose in those at risk of suicide-related behaviors, and exacerbation of respiratory issues in those with sleep apnea. Clinical practices that do not comply with evidenced-based care may result from time and information processing constraints, suggesting that strategies are needed to identify Veterans at increased risk of adverse events. This pilot quality improvement (QI) project evaluates the implementation of an advanced medication alert aimed at reducing co-prescribing of opioids and benzodiazepines among Veterans with substance use and other high risk conditions at one VA healthcare system.
The aims of this project include: 1) assess perspectives of prescribers, clinical pharmacist specialists and key facility leaders on concurrent use of opioids and benzodiazepines among Veterans with substance use and other high risk conditions, 2) identify contextual barriers and facilitators related to modifying prescribing practices and to use of the advanced medication alert in order to refine the alert for future multi-site implementation studies, and 3) evaluate the preliminary effectiveness of the advanced medication alert in reducing concurrent use of opioids and benzodiazepines among Veterans with substance use and other high risk conditions.
This prospective mixed-methods QI project used the Promoting Action on Research Implementation in Health Services model as the framework for evaluating the implementation of an advanced medication alert. Data sources included surveys, semi-structured interviews and pharmacy data from the VISN 20 Data Warehouse. All prescribers working in primary care and mental health services at the VA Puget Sound Healthcare System (VAPSHCS) were eligible to participate. Utilizing pharmacy data, we identified 182 providers who ordered opioids or benzodiazepines between January 1, 2014 and March 31, 2014 and invited them to complete an anonymous, electronic survey prior to launch of the alert (Aim 1). Following the alert launch, prescribers with multiple exposures to the alert were invited to participate in semi-structured interviews (Aim 2). Using an interrupted time series approach, we examined changes in co-prescribing rates for the 12-months after the alert launch after adjusting for the trend in co-prescribing in the 12 months prior to launch, as well as demographic and clinical covariates. Separate regressions modeled changes in co-prescribing rates for each risk condition. The effect of the alert was also evaluated using a control healthcare system (HCS) without the alert under evaluation (Aim 3).
Overall, 47.8% of primary care and 46.3% of mental health prescribers responded to the survey. Respondents' mean age was 50.9 (SD=9.6) and 57% were women. Eighty-seven percent of prescribers strongly agreed with clinical practice guidelines that recommend caution in co-prescribing opioids and benzodiazepines to patients with substance use conditions. Beliefs endorsed most often among prescriber groups as contributing to co-prescribing included tapering/discontinuing would be too difficult (52%), tapering/discontinuing would cause patients to suffer (55%) and patients are stable on these medications with no adverse effects (42%). Several resources were rated as quite to extremely helpful in reducing co-prescribing, including approved strategies for patients who refuse to taper these medications (>80%), access to alternative behavioral interventions (>70%) and more time with patients (>70%). Prescribers' perspectives on alerts were largely favorable with over 70% agreeing/strongly agreeing that alerts can identify patient risks that might otherwise go unnoticed and prevent serious adverse outcomes.
Several facilitators and barriers to using the alert and reducing opioid and benzodiazepine co-prescribing emerged from qualitative analyses. Factors that supported use of the alert included that it was easy to use, identified patients most at risk and provided useful information. Barriers to using the alert included general fatigue associated with alerts, additional time required to process the alert and report by some providers that the alert lacked new information. Prescribers expressed beliefs that concurrent prescribing was reduced when medication and behavioral treatment alternatives were available and seen as efficacious, and when potential harms associated with continued co-prescribing were evident, either due to patients' misuse of medications or conditions that clearly increased the risk of adverse events. Prescribers also expressed significant reluctance in starting patients on this medication combination. Barriers to reducing opioid and benzodiazepine co-prescribing included unpleasant interactions with patients who are reluctant to taper one or both of the medications, frustration with the responsibility of tapering patients who were inherited from other prescribers and lack of time.
Among Veterans with substance use disorders at VAPSHCS, there was a decreasing trend in opioid and benzodiazepine co-prescribing in the year after the alert launch of approximately 4% per month [Adjusted odds ratio (AOR)=0.98, 95% CI:0.97-1.00, p<0.05], compared to 2% per month in the year prior to launch. However, the trend in co-prescribing at VAPSHCS was not different from the trend in co-prescribing at the control HCS [AOR=0.98, (0.96-1.01)]. For Veterans with high suicide risk at VAPSHCS, there was a significant decreasing trend in concurrent opioid and benzodiazepine use after the alert was launched of approximately 6% per month (AOR=0.95, 95% CI:0.91-0.98, p<0.01), compared to 0% per month in the prior year. The change in trend at VAPSHCS significantly differed from the change at the control HCS (AOR=0.95, 95% CI:0.92-0.99, p<0.05). Likewise, there was a decrease in the trend in co-prescribing for those with sleep apnea in the year after launch of approximately 3% per month at VAPSHCS (AOR=0.97, 95% CI 0.96-0.98, p<0.001), compared to 0% per month in prior year. This change was greater than the change at the control HCS (AOR=0.98, 95% CI:0.96-1.00, p<0.05). Among those aged 65 and older at VAPSHCS, no change was observed in the trend in co-prescribing over the year after the alert launch, nor was a difference in trends observed between VAPSHCS and the control HCS.
Results of this QI project suggest advanced medication alerts may facilitate efforts to reduce co-prescribing of opioids and benzodiazepines among Veterans with substance use and other high-risk conditions and may reduce adverse events.
- Malte CA, Berger D, Saxon AJ, Hagedorn HJ, Achtmeyer CE, Mariano AJ, Hawkins EJ. Electronic Medical Record Alert Associated With Reduced Opioid and Benzodiazepine Coprescribing in High-risk Veteran Patients. Medical care. 2018 Feb 1; 56(2):171-178.
- Hawkins EJ, Malte CA, Hagedorn HJ, Berger D, Frank A, Lott A, Achtmeyer CE, Mariano AJ, Saxon AJ. Survey of Primary Care and Mental Health Prescribers' Perspectives on Reducing Opioid and Benzodiazepine Co-Prescribing Among Veterans. Pain medicine (Malden, Mass.). 2017 Mar 1; 18(3):454-467.
- Hawkins EJ, Malte CA, Saxon AJ, Sayre G. Evaluating a Medication Alert to Reduce Concurrent Opioid and Benzodiazepine Use at a Single VA Health Care System. Paper presented at: Society for Implementation Research Collaboration Biennial Conference on Advancing Efficient Methodologies Through Community Partnerships and Team Science; 2015 Sep 10; Seattle, WA.
- Hawkins EJ, Malte CA, Grossbard JR, Saxon AJ. Prevalence and Trends of Concurrent Opioid Analgesic and Benzodiazepine Use Among Veterans Affairs Patients with PTSD 2003-2011. Poster session presented at: VA HSR&D / QUERI National Meeting; 2015 Jul 10; Philadelphia, PA.