Statistical considerations for medication adherence research

Association Between Buprenorphine Adherence Trajectories, Health Outcomes, and Health Care Costs Among Medicaid Enrollees

OBJECTIVES

To identify distinct buprenorphine adherence trajectories among patients with opioid use disorder (OUD) and evaluate their associations with health events and health care costs.

METHODS

A retrospective longitudinal cohort study was conducted using the Merative Multi-state Medicaid database. The study analyzed 12,244 Medicaid enrollees aged 18-64 years who were diagnosed with OUD and initiated buprenorphine treatment between July 1, 2017 and June 30, 2019. Group-based trajectory models were used to identify adherence patterns during the first 180 days of treatment. Cox proportional hazard models were used to evaluate the associations between adherence trajectories and time to opioid overdose, substance use disorder-related hospitalization, and all-cause hospitalization. Generalized linear models were used to compare health care costs across trajectories.

RESULTS

Four buprenorphine adherence trajectories were identified: completely adherent (50.8%), initially adherent with later decline (13.6%), increasing adherence with later decline (9.9%), and continuously declining nonadherence (25.8%). Compared to the completely adherent group, patients in other groups had a higher risk of opioid overdose, hospitalization and increased health care costs. The continuously declining nonadherent group demonstrated the highest risks, with an opioid overdose hazard ratio (HR) of 1.92 (95% CI, 1.46-2.39), all-cause hospitalization of HR of 1.71 (95% CI: 1.58-1.85), and substance use disorder (SUD)-related hospitalization HR of 2.01 (95% CI: 1.82-2.15). Additionally, healthcare costs were notably higher compared to the completely adherent group, with an increase of $1482.45 (95% CI: $745.45-$2756.01) in the increasing adherence with later decline group and $1698.46 (95% CI: $432.57-$3087.78) in the continuously declining nonadherence groups.

CONCLUSIONS

Almost half of Medicaid beneficiaries with OUD exhibited varying degrees of nonadherence to buprenorphine within 180 days of treatment initiation. This nonadherence was associated with adverse clinical outcomes and increased health care costs. Health care providers should consider adherence challenges when designing therapeutic interventions with buprenorphine.

Impact of the family doctor system on medication adherence among type 2 diabetes patients in China: a difference-in-differences analysis

BACKGROUND

Nonadherence to medication is a prevalent issue among patients with type 2 diabetes mellitus (T2DM). The family doctor system promotes continuous, integrated, and personalized primary care, which may improve medication adherence. However, more evidence is needed regarding causal association between family doctor system and medication adherence. This study is to assess the impact of the family doctor system on medication adherence among T2DM patients.

METHODS

This cohort study utilized data between 2014 and 2019 from three separate administrative databases of an eastern city in China. Adult patients with T2DM who continuously registered with family doctors from 2015 to 2019 (n = 18,841) were assigned to the intervention group, while those who never registered during this period (n = 1,429) were classified as the control group. A difference-in-differences design was employed to compare medication adherence between registered and unregistered T2DM patients after the first stage of family doctor system in 2015 and the second stage in 2018. Propensity score matching was applied to ensure the robustness. The mean proportion of days covered (PDC), and proportion of patients with good adherence (PDC ≥ 80%) were the outcomes of interest. All recommended T2DM medications were included to calculate PDC.

RESULTS

Compared to the unregistered group, PDC across all institutions for registered patients increased by 5.0% (p < 0.001) after the first stage of family doctor system, and by 5.9% (p < 0.001) after the second stage. The proportion of patients with good adherence increased by 9.5% (p < 0.001) and by 11.8% (p < 0.001) after two stages, respectively. The adherence improvement was more pronounced in community health centers. However, the overall good adherence rate among registered patients in 2019 remained relatively low, reaching only 59.77%.

CONCLUSIONS

The family doctor system significantly improved medication adherence among T2DM patients by providing patient-centered, continuous, and integrated primary services, especially in community health centers. Nevertheless, further efforts should be made to enhance medication adherence.

Identifying the Relative Importance of Factors Influencing Medication Compliance in General Patients Using Regularized Logistic Regression and LightGBM: Web-Based Survey Analysis

BACKGROUND

Medication compliance, which refers to the extent to which patients correctly adhere to prescribed regimens, is influenced by various psychological, behavioral, and demographic factors. When analyzing these factors, challenges such as multicollinearity and variable selection often arise, complicating the interpretation of results. To address the issue of multicollinearity and better analyze the importance of each factor, machine learning methods are considered to be useful.

OBJECTIVE

This study aimed to identify key factors influencing medication compliance by applying regularized logistic regression and LightGBM.

METHODS

A questionnaire survey was conducted among 638 adult patients in Japan who had been continuously taking medications for at least 3 months. The survey collected data on demographics, medication habits, psychological adherence factors, and compliance. Logistic regression with regularization was used to handle multicollinearity, while LightGBM was used to calculate feature importance.

RESULTS

The regularized logistic regression model identified significant predictors, including "using the drug at approximately the same time each day" (coefficient 0.479; P=.02), "taking meals at approximately the same time each day" (coefficient 0.407; P=.02), and "I would like to have my medication reduced" (coefficient -0.410; P=.01). The top 5 variables with the highest feature importance scores in the LightGBM results were "Age" (feature importance 179.1), "Using the drug at approximately the same time each day" (feature importance 148.4), "Taking meals at approximately the same time each day" (feature importance 109.0), "I would like to have my medication reduced" (feature importance 77.48), and "I think I want to take my medicine" (feature importance 70.85). Additionally, the feature importance scores for the groups of medication adherence-related factors were 77.92 for lifestyle-related items, 52.04 for awareness of medication, 20.30 for relationships with health care professionals, and 5.05 for others.

CONCLUSIONS

The most significant factors for medication compliance were the consistency of medication and meal timing (mean of feature importance), followed by the number of medications and patient attitudes toward their treatment. This study is the first to use a machine learning model to calculate and compare the relative importance of factors affecting medication adherence. Our findings demonstrate that, in terms of relative importance, lifestyle habits are the most significant contributors to medication compliance among the general patient population. The findings suggest that regularization and machine learning methods, such as LightGBM, are useful for better understanding the numerous adherence factors affected by multicollinearity.

Adherence and Treat-to-Target Benchmarks in Older Adults With Gout Initiating Urate-Lowering Therapy in Ontario, Canada: A Population-Based Study

OBJECTIVE

We sought to evaluate urate-lowering therapy (ULT) adherence and treatment-to-target (T2T) serum uric acid (SUA) levels among older adults with gout starting ULT.

METHODS

We performed a population-based retrospective cohort study in Ontario, Canada in patients with gout aged ≥66 years newly dispensed ULT between 2010 and 2019. We defined successful T2T as patients having SUA levels <360 μmol/L (6 mg/dL) within 12 months after ULT dispensation. We also assessed adherence to ULT. Multilevel logistic regression clustered by ULT prescriber evaluated patient, physician, and prescription factors associated with reaching target SUA levels.

RESULTS

Among 44,438 patients (mean ± SD age 76.0 ± 7.3 years; 64.4% male), 30,057 (67.6%) patients had ≥1 SUA test completed. Overall, 52.3% patients reached SUA target within 12 months, improving from 45.2% in 2010 to 61.2% in 2019 (P < 0.0001). ULT adherence was 55.3% overall and improved annually. Key factors associated with achieving T2T included febuxostat treatment (odds ratio [OR] 11.40, 95% confidence interval [95% CI] 5.10-25.43) (was only dispensed in 88 patients), ULT adherence (OR 5.17, 95% CI 4.89-5.47), allopurinol starting doses >50 mg (OR 2.53, 95% CI 2.14-2.99), colchicine/oral glucocorticoids co-prescription (OR 1.24, 95% CI 1.14-1.34), and ULT prescription from a rheumatologist.

CONCLUSION

Only 52.3% of patients achieved an optimal SUA level within 1 year of ULT initiation. ULT adherence was suboptimal, although improving over time. ULT adherence and higher allopurinol starting doses had the strongest associations of achieving a target SUA level. This study highlights room for improvement in gout management and potential strategies to address care gaps.

Imatinib Adherence and Persistence in Patients with Chronic Myeloid Leukemia in Belgium: Evidence from Real-World Data

Purpose

Imatinib adherence and persistence are key components of the successful treatment of Chronic Myeloid Leukemia (CML). In Belgium, there is no study assessing these behaviors at a national level. Our study aimed to provide the first nationwide measure and to identify associated pharmacy-based predictors (age, gender, comorbidities). We also assessed mortality and transplantation incidence according to adherence status.

Methods

Based on medico-administrative database linkage, we identified a retrospective Belgian cohort of 1194 patients diagnosed with CML between 2004 and 2016 and treated with imatinib. Adherence was measured over 24 months, considering the proportion of days covered (PDC). Persistence was measured as the time until discontinuation (gap of ≥90 days). Multivariable Poisson regression models with robust standard error were conducted to identify predictors associated with adherence (≥90% PDC). To identify factors associated with persistence, a multivariable Cox regression was performed.

Results

At six months, 60.3% of patients were adherent, declining to 41.5% at 12 months, and to 30.1% at two-year follow-up (n=998). Adherence was greater at a younger age (eg 31-40 years vs ≥75 years, adjusted prevalence ratio (aPR) 1.73; 95% confidence interval (CI): 1.09-2.77) and among patients with no comorbidity (0 vs ≥2 comorbidities (aPR 1.56; 95% CI: 1.11-2.19). The median persistence was 334.5 days (Q1:200-Q3:505.5); persistence at 24 months was 83.6% (n=998). Only age was associated with higher risk of discontinuation, with adjusted hazard ratio (aHR) of 6.05 for patients ≥75 years (95% CI: 5.52-6.58). Transplants and deaths mainly occurred in patients defined as non-adherent at 24 months.

Conclusion

This Belgian nationwide representative study revealed a critical low level of imatinib adherence, which decreased over time even though persistence was high at six months. We pinpointed pharmacy-based predictors that were easily identifiable by health care stakeholders in order to undertake interventions to improve adherence.

Estimating medication adherence from Electronic Health Records: comparing methods for mining and processing asthma treatment prescriptions

BACKGROUND

Medication adherence is usually defined as the extent of the agreement between the medication regimen agreed to by patients with their healthcare provider and the real-world implementation. Proactive identification of those with poor adherence may be useful to identify those with poor disease control and offers the opportunity for ameliorative action. Adherence can be estimated from Electronic Health Records (EHRs) by comparing medication dispensing records to the prescribed regimen. Several methods have been developed in the literature to infer adherence from EHRs, however there is no clear consensus on what should be considered the gold standard in each use case. Our objectives were to critically evaluate different measures of medication adherence in a large longitudinal Scottish EHR dataset. We used asthma, a chronic condition with high prevalence and high rates of non-adherence, as a case study.

METHODS

Over 1.6 million asthma controllers were prescribed for our cohort of 91,334 individuals, between January 2009 and March 2017. Eight adherence measures were calculated, and different approaches to estimating the amount of medication supply available at any time were compared.

RESULTS

Estimates from different measures of adherence varied substantially. Three of the main drivers of the differences between adherence measures were the expected duration (if taken as in accordance with the dose directions), whether there was overlapping supply between prescriptions, and whether treatment had been discontinued. However, there are also wider, study-related, factors which are crucial to consider when comparing the adherence measures.

CONCLUSIONS

We evaluated the limitations of various medication adherence measures, and highlight key considerations about the underlying data, condition, and population to guide researchers choose appropriate adherence measures. This guidance will enable researchers to make more informed decisions about the methodology they employ, ensuring that adherence is captured in the most meaningful way for their particular application needs.

Variation in missed doses and reasons for discontinuation of anti-tuberculosis drugs during hospital treatment for drug-resistant tuberculosis in South Africa

BACKGROUND

Updated World Health Organization (WHO) treatment guidelines prioritize all-oral drug-resistant tuberculosis (DR-TB) regimens. Several poorly tolerated drugs, such as amikacin and para-aminosalicylic acid (PAS), remain treatment options for DR-TB in WHO-recommended longer regimens as Group C drugs. Incomplete treatment with anti-TB drugs increases the risk of treatment failure, relapse, and death. We determined whether missed doses of individual anti-TB drugs, and reasons for their discontinuation, varied in closely monitored hospital settings prior to the 2020 WHO DR-TB treatment guideline updates.

METHODS

We collected retrospective data on adult patients with microbiologically confirmed DR-TB between 2008 and 2015 who were selected for a study of acquired drug resistance in the Western Cape Province of South Africa. Medical records through mid-2017 were reviewed. Patients received directly observed treatment during hospitalization at specialized DR-TB hospitals. Incomplete treatment with individual anti-TB drugs, defined as the failure to take medication as prescribed, regardless of reason, was determined by comparing percent missed doses, stratified by HIV status and DR-TB regimen. We applied a generalized mixed effects model.

RESULTS

Among 242 patients, 131 (54%) were male, 97 (40%) were living with HIV, 175 (72%) received second-line treatment prior to first hospitalization, and 191 (79%) died during the study period. At initial hospitalization, 134 (55%) patients had Mycobacterium tuberculosis with resistance to rifampicin and isoniazid (multidrug-resistant TB [MDR-TB]) without resistance to ofloxacin or amikacin, and 102 (42%) had resistance to ofloxacin and/or amikacin. Most patients (129 [53%]) had multiple hospitalizations and DST changes occurred in 146 (60%) by the end of their last hospital discharge. Incomplete treatment was significantly higher for amikacin (18%), capreomycin (18%), PAS (17%) and kanamycin (16%) than other DR-TB drugs (P<0.001), including ethionamide (8%), moxifloxacin (7%), terizidone (7%), ethambutol (7%), and pyrazinamide (6%). Among the most frequently prescribed drugs, second-line injectables had the highest rates of discontinuation for adverse events (range 0.56-1.02 events per year follow-up), while amikacin, PAS and ethionamide had the highest rates of discontinuation for patient refusal (range 0.51-0.68 events per year follow-up). Missed doses did not differ according to HIV status or anti-TB drug combinations.

CONCLUSION

We found that incomplete treatment for second-line injectables and PAS during hospitalization was higher than for other anti-TB drugs. To maximize treatment success, interventions to improve person-centered care and mitigate adverse events may be necessary in cases when PAS or amikacin (2020 WHO recommended Group C drugs) are needed.

Effect of CGM Access Expansion on Uptake Among Patients on Medicaid With Diabetes

OBJECTIVE

Current studies on continuous glucose monitor (CGM) uptake are revealing for significant barriers and inequities for CGM use among patients from socially underprivileged communities. This study explores the effect of full subsidies regardless of diabetes type on CGM uptake and HbA1c outcomes in a U.S. adult patient population on Medicaid.

RESEARCH DESIGN AND METHODS

This retrospective cohort study examined 3,036 adults with diabetes enrolled in a U.S. Medicaid program that fully subsidized CGM. CGM uptake and adherence were assessed by CGM prescription and dispense data, including more than one fill and adherence by medication possession ratio (MPR). Multivariate logistic regression evaluated predictors of CGM uptake. Pre- and post-CGM use HbA1c were compared.

RESULTS

CGM were very well received by both individuals with type 1 diabetes and individuals with type 2 diabetes with similar high fill adherence levels (mean MPR 0.78 vs. 0.72; P = 0.06). No significant difference in CGM uptake outcomes were noted among major racial/ethnic groups. CGM use was associated with improved HbA1c among those with type 2 diabetes (-1.2% [13.1 mmol/mol]; P < 0.001) that was comparable between major racial/ethnic groups and those with higher fill adherence achieved greater HbA1c reduction (-1.4% [15.3 mmol/mol]; P < 0.001) compared with those with lower adherence (-1.0% [10.9 mmol/mol]; P < 0.001).

CONCLUSIONS

CGM uptake disparities can largely be overcome by eliminating CGM cost barriers. CGM use was associated with improved HbA1c across all major racial/ethnic groups, highlighting broad CGM appeal, utilization, and effectiveness across an underprivileged patient population.

Sacubitril/valsartan use patterns among older adults with heart failure in clinical practice: a population-based cohort study of >25 000 Medicare beneficiaries

AIMS

Sacubitril/valsartan is strongly supported in guidelines for the management of heart failure, but suboptimal adherence and treatment non-persistence may limit the population-level benefit that this therapy might otherwise offer.

METHODS AND RESULTS

We identified a cohort of Medicare beneficiaries (2014-2017) initiating sacubitril/valsartan after ≥6 months of continuous enrolment. We assessed adherence as the proportion of days covered (PDC) and proportion of patients non-persistent (having no prescription available) at 180 days after initiation. We fit a multivariable negative binomial model with a count of adherent days to evaluate independent factors associated with of sacubitril/valsartan adherence. Among 27 063 new sacubitril/valsartan users, most (n = 17 663, 65%) were prescribed low-dose at 24 mg/26 mg and most (n = 19 984, 74%) were switched from prior angiotensin-converting enzyme inhibitor/angiotensin receptor blocker (ACEi/ARB) rather than being RASi treatment naïve. Median 180-day PDC was 86% (25th-75th percentiles 58-98%). Black patients, those with high comorbid disease burden (≥8 comorbidities), and patients with recent hospitalization within 30 days had fewer adherent days, while those treated with preceding ACEi/ARB had more adherent days. Thirty-four percent of patients did not have an active sacubitril/valsartan prescription at day 180. Among these, few had preceding dose down-titrations (6% among patients on 49 mg/51 mg and 9% among patients on 97 mg/103 mg) and 68% did not have a subsequent ACEi/ARB prescription. Among patients who remained persistent, dose up-titrations occurred in 29% of patients who started on 24 mg/26 mg and 27% of patients on 49 mg/51 mg.

CONCLUSIONS

Overall adherence to sacubitril/valsartan among Medicare beneficiaries is acceptable, but is lower in Black patients, those with higher comorbidities or those who started therapy after recent hospitalization. While broad implementation of guideline-directed medical therapy is a key priority, additional focused efforts to improve adherence early after hospitalization and among at-risk patients are needed in parallel.

Delta T, a Useful Indicator for Pharmacy Dispensing Data to Monitor Medication Adherence

INTRODUCTION

Calculating patients' medication availability from dispensing or refill data is a common method to estimate adherence. The most often used measures, such as the medication possession ratio (MPR), average medication supplies over an arbitrary period. Averaging masks the variability of refill behavior over time.

GOAL

To derive a new absolute adherence estimate from dispensing data.

METHOD

Dispensing histories of patients with 19 refills of direct oral anticoagulants (DOAC) between 1 January 2008 and 31 December 2017 were extracted from 39 community pharmacies in Switzerland. The difference between the calculated and effective refill day (ΔT) was determined for each refill event. We graphed ΔT and its dichotomized version (dΔT) against the MPR, calculated mean ΔT and mean dΔT per refill, and applied cluster analysis.

RESULTS

We characterized 2204 refill events from 116 DOAC patients. MPR was high (0.975 ± 0.129) and showed a positive correlation with mean ΔT. Refills occurred on average 17.8 ± 27.9 days "too early", with a mean of 75.8 ± 20.2 refills being "on time". Four refill behavior patterns were identified including constant gaps within or at the end of the observation period, which were critical.

CONCLUSION

We introduce a new absolute adherence estimate ΔT that characterizes every refill event and shows that the refill behavior of DOAC patients is dynamic.

Proportion of days covered as a measure of medication adherence

In an effort to expedite the publication of articles , AJHP is posting manuscripts online as soon as possible after acceptance. Accepted manuscripts have been peer-reviewed and copyedited, but are posted online before technical formatting and author proofing. These manuscripts are not the final version of record and will be replaced with the final article (formatted per AJHP style and proofed by the authors) at a later time.

Pharmacist Interventions to Improve Specialty Medication Adherence: Study Protocol for a Randomized Controlled Trial

BACKGROUND

The effectiveness of specialty medications in complicated clinical conditions depends on adherence to therapy. However, specialty medications pose unique barriers to adherence.

OBJECTIVE

This study aims to determine whether pharmacist interventions improve specialty medication adherence.

METHODS

This is a single-center, pragmatic, randomized controlled trial ongoing since 10 May 2019 at an integrated health system specialty pharmacy. This study evaluates usual care compared with usual care plus patient-tailored adherence interventions. Study design and procedures were informed by focus groups with patients and specialty pharmacists. Patients at Vanderbilt Specialty Pharmacy with a proportion of days covered (PDC) < 90% in the previous 4 months are identified by a daily query of the electronic pharmacy database. A pharmacist reviews these patients' electronic health records to identify and exclude ineligible patients. Eligible patients are randomized evenly to the control or intervention arm and stratified by historical clinic nonadherence rates. Patients randomized to the intervention arm undergo a baseline assessment to clarify reasons for nonadherence and subsequently receive patient-tailored interventions based on their specific reasons. Interventions and follow-up are provided at the discretion of the intervening pharmacist. The primary outcome is PDC calculated at 8 months post-enrollment. Enrollment of 438 participants will provide 90% power to detect a 5% difference in PDC between the two arms within each nonadherence risk stratum.

DISCUSSION

This trial will evaluate the effect of patient-tailored interventions on specialty medication adherence and will inform how often and why patients are misidentified as nonadherent.

REGISTRATION

The trial was deemed a quality improvement initiative by the Vanderbilt University Institutional Review Board. It was registered in ClinicalTrials.gov (NCT03709277) on 17 October 2018.