Using trajectory models to assess the effect of hydrocodone upscheduling among chronic hydrocodone users

Competitive effects of federal and state opioid restrictions: Evidence from the controlled substance laws

A significant concern in the policy landscape of the U.S. opioid crisis is whether supply-side controls can reduce opioid prescribing without harmful substitution. We consider an unstudied policy: the federal Controlled Substance Act (CSA) restrictions placed in August 2014 on tramadol, the second most popular opioid medication. This was followed seven weeks later by CSA restrictions for hydrocodone combination products, the leading opioids on the market. Using regression discontinuity design (RDD) models, based on the timing of the (up-)scheduling changes, to explore spillover effects, we find that tightening prescribing restrictions on one opioid reduces its use, but increases prescribing of close competitors, leading to no reduction in total opioid prescriptions.This suggests that supply restrictions are not effective in reducing opioid prescribing the presence of close substitutes that remain unrestricted.

Changes in Hydrocodone Misuse Exposures Reported to U.S. Poison Centers Following Rescheduling in 2014

BACKGROUND

In 2014, the Drug Enforcement Administration rescheduled hydrocodone combination products to Schedule II to reduce nonmedical use and diversion.

METHODS

The impact of rescheduling was assessed using quarterly data from 2011 through 2019 from the Researched Abuse, Diversion and Addiction-Related Surveillance (RADARS^®) System Poison Center Program and IQVIA^TM Longitudinal Prescription Data. Trends and immediate changes in prescriptions dispensed and misuse exposures before and after rescheduling involving hydrocodone, oxycodone, and other Schedule II opioid analgesics were calculated using segmented regression.

RESULTS

Hydrocodone prescriptions were stable pre-rescheduling, decreased by 2.7% (95% CI: -3.6%, -1.8%, p < 0.0001) per quarter post-rescheduling. Misuse exposures involving hydrocodone were decreasing by 3.2% (95% CI: -3.9%, -2.4%, p < 0.0001) per quarter pre-rescheduling and decreased by 4.9% (95% CI: -5.5%, -4.2%, p < 0.0001) post-rescheduling. Immediate decreases in hydrocodone prescriptions and misuse exposure rates in 2014Q4 compared to 2014Q3 were significant and different from oxycodone or other Schedule II opioids. Schedule II opioid analgesics prescriptions in aggregate were stable prior to rescheduling, decreased by 10.8% (95%CI: -14.0%, -7.6%, p < 0.0001) immediately after the rescheduling, and decreased by 2.3% per quarter (95% CI: -3.1%, -1.5%, p < 0.0001) subsequently. Misuse exposures involving these opioids were decreasing by 3.3% (95% CI: -4.1%, -2.5%, p < 0.0001) prior to rescheduling then by 2.8%, (95% CI: -3.4%, -2.2%, p < 0.0001) after rescheduling. The immediate change in misuse was not significant.

CONCLUSIONS

Rescheduling corresponded with changes in hydrocodone prescribing and misuse not offset by increases in other Schedule II opioid analgesics. Misuse exposures for hydrocodone and comparators were decreasing prior to rescheduling with little change post-intervention.

Comparing Continuous and Binary Group-based Trajectory Modeling Using Statin Medication Adherence Data

BACKGROUND

Of 58 medication adherence group-based trajectory modeling (GBTM) published studies, 74% used binary and 26% used continuous GBTM. Few studies provided a rationale for this choice. No medication adherence studies have compared continuous and binary GBTM.

OBJECTIVE

The objective of this study was to assess whether continuous versus binary GBTM: (1) impacts adherence trajectory shapes; and (2) results in the differential classification of patients into adherence groups.

METHODS

Patients were prevalent statin users with myocardial infarction hospitalization, 66+ years old, and continuously enrolled in fee-for-service Medicare. Statin medication adherence was measured 6 months prehospitalization using administrative claims. Final GBTM specifications beyond default settings were selected using a previously defined standardized procedure and applied separately to continuous and binary (proportion of days covered ≥0.80) medication adherence measures. Assignment to adherence groups was compared between continuous and binary models using percent agreement of patient classification and the κ coefficient.

RESULTS

Among 113,296 prevalent statin users, 4 adherence groups were identified in both models. Three groups were consistent: persistently adherent, progressively nonadherent, and persistently nonadherent. The fourth continuous group was moderately adherent (progressively adherent in the binary model). When comparing patient assignment into adherence groups between continuous and binary trajectory models, only 78.4% of patients were categorized into comparable groups (κ=0.641; 95% confidence interval: 0.638-0.645). The agreement was highest in the persistently adherent group (∼94%).

CONCLUSIONS

Continuous and binary trajectory models are conceptually different measures of medication adherence. The choice between these approaches should be guided by study objectives and the role of medication adherence within the study-exposure, outcome, or confounder.

Longitudinal K-means approaches to clustering and analyzing EHR opioid use trajectories for clinical subtypes

Identification of patient subtypes from retrospective Electronic Health Record (EHR) data is fraught with inherent modeling issues, such as missing data and variable length time intervals, and the results obtained are highly dependent on data pre-processing strategies. As we move towards personalized medicine, assessing accurate patient subtypes will be a key factor in creating patient specific treatment plans. Partitioning longitudinal trajectories from irregularly spaced and variable length time intervals is a well-established, but open problem. In this work, we present and compare k-means approaches for subtyping opioid use trajectories from EHR data. We then interpret the resulting subtypes using decision trees, examining how each subtype is influenced by opioid medication features and patient diagnoses, procedures, and demographics. Finally, we discuss how the subtypes can be incorporated in static machine learning models as features in predicting opioid overdose and adverse events. The proposed methods are general, and can be extended to other EHR prescription dosage trajectories.

Patterns of opioid analgesic use in the U.S., 2009 to 2018

ABSTRACT

Although overall outpatient dispensing of opioid analgesic prescriptions has declined, there may still be overprescribing. Understanding how many opioid analgesic units, primarily tablets, are dispensed with the intention of shorter-vs longer-term use can inform public health interventions. We used pharmacy prescription data to estimate the number of opioid analgesic tablets dispensed annually in the U.S. We studied patterns of new use of opioid analgesics by evaluating how many opioid analgesic prescriptions and tablets were dispensed to patients with no opioid analgesic prescriptions in the previous year. Estimated opioid analgesic tablets dispensed declined from a peak of 17.8 billion in 2012 to 11.1 billion in 2018. Patients newly starting opioid analgesics declined from 47.4 million patients in 2011 to 37.1 million patients in 2017. Approximately 40% fewer tablets were dispensed within a year to patients starting in 2017 (2.4 billion) compared with 2011 (4.0 billion). In 2011, patients with ≥5 opioid analgesic prescriptions within a year were dispensed 2.2 billion tablets (55% of all tablets in our study). This declined by 52% to 1.1 billion tablets (44% of all tablets) in 2017. Tablets dispensed within a year to patients with <5 opioid analgesic prescriptions declined by 26% from 2011 to 2017. Patients with ≥5 prescriptions comprised a small and decreasing proportion of all patients newly starting therapy. However, these patients received almost half of all tablets dispensed within a year to patients in our study, despite a larger decline than tablets dispensed to patients with <5 prescriptions within a year.

Effects of hydrocodone rescheduling on opioid use outcomes: A systematic review

OBJECTIVE

To evaluate opioid prescribing, dispensing, and use in relation to hydrocodone-containing product (HCP) rescheduling.

METHODS

Seven biomedical databases and grey literature sources were searched with keywords and database-specific controlled vocabulary relevant to HCP rescheduling for items published between January 2014 and July 2019. We included English-language quasi-experimental studies that assessed changes in HCP and other opioid prescribing, dispensing, utilization, and opioid-related health outcomes before and after HCP rescheduling. A data extraction sheet was created for this review. Two authors evaluated risk of bias for each included study. Two of 4 authors each independently extracted patient demographics and opioid-related outcomes from the included studies. Conflicts were resolved by a third author.

RESULTS

All studies identified (n = 44) were quasi-experimental in design with 10 using an interrupted time series approach. A total of 24 studies reported a decrease in HCP prescribing by 3.1%-66.0%. Six studies reported a decrease in HCP days' supply or doses by 14.0%-80.8%. There was increased prescribing of oxycodone-containing products by 4.5%-13.9% in 5 studies, tramadol by 2.7%-53.0% in 9 studies, codeine-containing products by 0.8%-1352.9% in 8 studies). Five studies reported a decrease in morphine equivalents by at least 10%, whereas 2 studies reported an increase in morphine equivalents. Differences in populations, sample sizes, and approaches did not allow for a meta-analysis. Details regarding approach and findings were limited in published conference abstracts (n = 16).

CONCLUSIONS

Hydrocodone rescheduling was associated with reductions in prescribing and use of HCPs but was also associated with increased prescribing and use of other opioids, both schedule II and nonschedule II.

The state of the science in opioid policy research

OBJECTIVE

Characterize the state of the science in opioid policy research based on a literature review of opioid policy studies.

METHODS

We conducted a scoping review of studies evaluating the impact of U.S. state-level and federal-level policies on opioid-related outcomes published in 2005-2018. We characterized: 1) state and federal policies evaluated, 2) opioid-related outcomes examined, and 3) study design and analytic methods (summarized overall and by policy category).

RESULTS

In total, 145 studies were reviewed (79 % state-level policies, 21 % federal-level policies) and classified with respect to 8 distinct policy categories and 7 outcome categories. The majority of studies evaluated policies related to prescription opioids (prescription drug monitoring programs (PDMPs), opioid prescribing policies, federal regulation of prescription opioids, pain clinic laws) and considered policy impacts with respect to proximal outcomes (e.g., opioid prescribing behaviors). In total, only 29 (20 % of studies) met each of three key criteria for rigorous design: analysis of longitudinal data with a comparison group design, adjustment for difference between policy-enacting and comparison states, and adjustment for potentially confounding co-occurring policies. These more rigorous studies were predominately published in 2017-2018 and primarily evaluated PDMPs, marijuana laws, treatment-related policies, and overdose prevention policies.

CONCLUSIONS

Our results indicated that study design rigor varied notably across policy categories, highlighting the need for broader adoption of rigorous methods in the opioid policy field. More evaluation studies are needed regarding overdose prevention policies and policies related to treatment access. Greater examination of distal outcomes and potential unintended consequences are also warranted.