Current CTN HSN-led studies

Health Insurance Instability and Mortality among Patients Receiving Buprenorphine Treatment for Opioid Use Disorder: Linkage of the Prescription Opioid Registry with the National Death Index, (CTN-0141) Anh Nguyen, Ingrid Binswanger, Jason Glanz, Co-PIs

This study builds on CTN-0084, which included an aim to assess the optimal length of BUP-MOUD. Preliminary data analysis indicates that disenrollment from health plans is a common occurrence among the study cohort, including during periods of active treatment, thus warranting a need to investigate potential consequences of coverage disruption. This study will extend this work in several ways: (1) link patients from four large health systems in CTN-0084 with the National Death Index, a centralized database of state vital statistics, to ascertain fact and cause of death not captured in health systems during disenrollment; (2) collect data through 2020 to capture coverage dynamics during the COVID-19 pandemic; and (3) match the study cohort to an all-payers claims database (APCD) and criminal justice data at the lead site in Colorado to assess health plan changes and incarceration associated with disenrollment. The study will contribute substantively and methodologically to the CTN. Findings will inform strategies to ensure treatment continuity for patients vulnerable to insurance instability. As EHR data are increasingly important for pragmatic trials, this study will advance research by identifying data sources and methods to address bias from loss to follow-up, a common concern across studies.

Cannabis Use among Pregnant Women with Psychiatric and Substance Use Disorders, (CTN-0140) Kelly Young-Wolff, PI, Gwen Lapham, Co-I.​

Pregnant women face barriers to substance use interventions, including stigma, mental health issues, cost, uncertainty about treatment, lack of programs, and competing demands on time.1It is unknown whether substance use treatment initiation in women with prenatal cannabis use varies among those with vs. without co-occurring psychiatric disorders or SUDs - predictors of treatment among this comorbid population has received little attention. Further, existing research has focused on utilization of specialty programs external to prenatal care and not on models of care integrated into medical settings. We address these important gaps with an unparalleled opportunity to study the relation between psychiatric disorders and SUDs and prenatal cannabis use and treatment utilization using valid and contemporary data. This study leverages Kaiser Permanente Northern California's (KPNC) well-established universal prenatal substance use screening program that includes self-report and urine toxicology tests. This program screens ~40,000 pregnant women annually, providing a large, diverse cohort with objective measures of prenatal cannabis use. KPNC has comprehensive electronic health records (EHR) enabling collection of rich clinical data, including psychiatric disorders and SUDs and treatment services.

The Impact of COVID-19 Pandemic on Substance Use Treatment Services, (CTN-0117) Cynthia Campbell​, PI​

As a result of the COVID-19 pandemic, we are witnessing profound societal and health system developments with significant implications for the treatment of drug use disorders: an increase in drug use, people of color at disproportionate risk for distress and drug use, and a potentially transformational move to virtual treatment modalities. The proposed study examines this major shift to virtual treatment for drug use disorders during the pandemic in a large, diverse health care system, specifically exploring potential disparities in accessing treatment. Virtual treatment is likely to persist as a major delivery model of alcohol treatment, and understanding potential inequities in access is critical to improving quality of care and outcomes for underserved popula​​tions.

Validating a DSM-5 Substance Use Disorder (SUD) Symptom Checklist in EHR Data from a Large Primary Care Sample to Support Future Pragmatic Trials and Chronic Care Management of SUDs in Primary Care, (CTN-0113) Emily Williams and Katharine Bradley Co-PIs

The specific aim of this 1-year research proposal is to evaluate whether the novel patient-reported DSM-5 SUD symptom checklist, used in routine primary care and documented in EHRs, is unidimensional, discriminative, and reflects a continuum of SUD severity, similar to how DSM-5 symptom measures function in epidemiological research. Additionally, we will evaluate whether the DSM-5 SUD symptom checklist functions similarly across age, sex, race, and ethnicity by testing of differential item functioning. This will be the first study to our awareness that evaluates a DSM-5 SUD symptom checklist used in routine care in a primary care setting for chronic disease management. Our evaluations will support efforts to improve the quality of SUD care in primary care and to validate common data elements that have potential for SUD research occurring in routine primary care (e.g., CTN implementation studies and pragmatic trials).

Harnessing Digital Health to Understand Clinical Trajectories of Opioid Use Disorder (D-TECT) (CTN 0084-A-2), Cynthia Campbel and Lisa Marsch, Co-PIs

Given the ubiquity of access to digital technologies worldwide, digital tools allow for the examination of health behavior and clinical trajectories within-individuals through intensive collection of individual-level, real-time data collected via surveys on mobile device (referred to as Ecological Momentary Assessment [EMA]), wearable sensors (on smartphones and/or smartwatches), and mapping digital footprints. As applied to persons with OUD, digital data that offers ongoing assessment of behavior as individuals live their daily lives can help us better understand the trajectory of clinically important behaviors (e.g., treatment retention; medication adherence over time) and identify fluctuating contextual factors that greatly influence such behaviors, (e.g., patterns leading to relapse or treatment dropout). The primary objective of this study is to evaluate the feasibility of utilizing digital health technology (EMA and digital sensing) continuously over a 12-week period and collecting social media data. The secondary objective is to examine the utility of EMA, digital sensing, and social media data (separately and compared to one another) in predicting OUD treatment retention, buprenorphine medication adherence (as captured in EHR, medical claims, and EMA data).

Analyses of Length of Buprenorphine Use and its Association with Mortality, (CTN-0084-A-1) Ingrid Binswanger and Jason Glanz, Co-PIs

This study is addressing the high priority issue regarding the optimal duration of buprenorphine treatment to reduce the risk of relapse, overdose, and mortality outcomes using observational data from CTN-0084's prescription opioid registry. Answering these questions with a randomized trial would raise ethical concerns but observational studies with large datasets can address these important questions relatively quickly. At the same time, observational studies pose their own methodologic challenges related to confounding, misclassification of exposure and outcome, and informative loss-to-follow-up. We are identifying and quantifying the potential for these sources of bias and then conducting analyses to address primary question of interest. KP Colorado, the site of this project is an integrated health system that has claims data, EHR data, buprenorphine dispensations, and information about methadone referrals and treatment. It is critical to understand these challenges and assess data quality in a rich data environment and potential sources of bias to inform future trials, such as the CTN duration trial, and comparative effectiveness studies to definitively answer these crucial questions regarding buprenorphine treatment.
​

Primary Care Opioid Use Disorder (PROUD) Treatment Trial Economic Analysis Study, (CTN-0074-A-1) Sean Murphy, PI

Effective treatment for opioid use disorder (OUD) has been shown to improve patient outcomes and reduce health care costs; however, evidence of this effect in primary care settings is severely limited. The health economic findings from this study will supplement the parent PROUD trial's results regarding clinical effectiveness and implementation outcomes and provide critical contextual information for health systems and other healthcare stakeholders. The study wille valuate the economic viability of the PROUD collaborative care model for OUD; that is, from the perspective of the health care sector, to what extent do the downstream cost savings associatd with improved patient outcomes offset the additional costs of the PROUD intervention. The specific aims are to: 1) estimate the start-up and ongoing management costs of the PROUD intervention; 2) assess costs associated with health care utilization for patients who receive primary care treatment in PROUD and usual care clinics and have been identified with recognized OUD prior to clinic randomization; 3) estimate the economic value of the PROUD intervention, measured as net monetary benefit (NMB, incremental benefit-incremental cost), from the health care sector perspective.​​

Developing a Prescription Opioid Registry across Diverse Health Systems (CTN-0084), Cynthia Campbell, PI

The opioid use disorder crisis continues in the U.S., in which prescription opioid use has played a key role. This study builds on CTN0061, which developed a prescription opioid registry at Kaiser Permanente Northern California and identified important predictors of opioid misuse and overdose across the years 2011-2014.?The current study aims to develop the prescription opioid registry across 10 diverse health systems with harmonized EHR data from years 2012-18, and leverage it to answer several key 'next step' research questions in response to the opioid crisis.?Specific questions include examining the association of opioid tapering to key health care utilization measures and mortality, and examining initial prescribing of opioids to subsequent opioid use, and the association of buprenorphine retention to mortality. To our knowledge, no study has established an EHR-based prescription opioid registry across several diverse health systems with common data algorithms with the flexibility to address varied questions.

Medical Cannabis Use Among Primary Care Patients: Using Electronic Health Records to Study Large Populations (CTN-0077), Gwen Lapham, PI

The United States is seeing a rapid rise in medical and non-medical cannabis use as states expand legal access, particularly in those states with legal recreational use. Yet, remarkably little is systematically known about the prevalence and long-term benefits and risks of medical cannabis use, as well as the levels of patient exposure to different cannabis products for medical use. Research is needed to identify the conditions and symptoms for which medical cannabis is most commonly used by patients who are engaged in primary care. Patient electronic health records (EHRs) have the potential to be a rich source of data about medical cannabis use among primary care patients. This study describes medical use of cannabis, including symptoms, conditions and use patterns common to medical use, among patients in a single large health system that routinely documents primary care patients' self-reported frequency of past-year cannabis use in a system-wide EHR. This study relies on structured EHR data for cannabis screen results and the use of natural language processing (NLP) methods whereby a computer "reads" narrative clinical text in EHR notes to distinguish medical cannabis use from other cannabis use.

PROUD Trial (Primary Care Opioid Use Disorder (CTN-0074), Katharine Bradley, PI

The US is facing an epidemic of opioid use disorders (OUDs). Treatment with medications markedly improves outcomes for OUDs, but OUDs remain under-diagnosed and under-treated. In Massachusetts, a nurse-based collaborative care model of OUD primary care treatment was used to increase 1) the number of patients who initiated OUD treatment with medications and 2) persistence of OUD treatment, known to improve health outcomes for patients with OUDs. In the PROUD trial, we are testing whether the Massachusetts model of OUD treatment in primary care, compared to usual primary care increases the number of patients who receive medication treatment for OUDs, and/or increases the persistence of OUD treatment in those who initiate treatment, in medical settings. The trial is conducted in 6 diverse health systems in Florida, Texas, New York, Michigan and Washington state, with observational analyses of health systems in Colorado and Oregon. All data are obtained from electronic health records, claims, and other secondary data sources. Phase 1 of the PROUD trial evaluated OUD diagnosis and treatment across 11 health care systems. Phase 2 was a pragmatic, cluster randomized trial that ended 2/29/2020. The PROUD intervention was continued 1 year after the trial (ending 2/28/2021) to evaluate longer term impacts.

Examine Patient and System-Level Factors Associated with HEDIS AOD-IET Measure Performance across Health Systems, (CTN-0072) Constance Weisner, PI

Access to care for alcohol and drug disorders (AOD) is problematic in the U.S.—only 10% of patients receive recommended care, and most of it is in specialty programs. Little is known about the HEDIS AOD Initiation and Engagement in Treatment (IET) measures and factors related to higher performance, but understanding their scope in identifying problems, referral to treatment, and engagement in AOD treatment (whether within primary care or in specialty AOD care) could be a first step for improving AOD care nationwide. In seven health plans, this study examined how: 1) performance on the AOD initiation and engagement measures varies by characteristics of the eligible population; 2) performance on the AOD initiation and engagement measures varies by system characteristics; and 3) structural differences and policies within each region and for specific facilities may explain variations in performance on the HEDIS AOD Initiation and Engagement measures.

Evaluation of Drug Screening Implementation in Primary Care (CTN-0065) Katharine Bradley, PI

Recent trials of screening and brief intervention for drug use have no demonstrated benefit compared to screening alone. However, the value of screening for drug use, compared to no screening, has not been evaluated. Screening for drug use may identify patients with drug use disorders who might need more intensive treatments than brief intervention. In states where marijuana use is now legal for recreation, as well as medical uses, clinicians want to be aware of and assess risks of marijuana use. Further, little is known about the predictive validity of screens for marijuana and/or drug use/misuse for subsequent adverse health outcomes. To that end, this study evaluated implementation of drug screening, among all Group Health patients who sought care in the Group Health primary care clinics to: 1) describe rates of drug and marijuana screening, and positive screens, as well as barriers and facilitators to population-based screening; 2) assess changes in rates of assessment and identification; and 3) assess whether drug and marijuana use is associated with increased subsequent urgent care, ED, and hospitalization.

Kaiser Virtual Data Warehouse - Prescription Opioid Users (CTN-0061) Cynthia Campbell, PI

To address the ongoing opioid crisis, this study used electronic health record data to develop a large, prescription opioid registry in a large, health care delivery system to examine predictors of chronic opioid use, opioid misuse, and adverse events such as overdose and mortality. We found higher risk was associated with gender, race/ethnicity, and by characteristics of opioid use, such a length of use and dose, as well as concurrent benzodiazepine use.