Measuring adherence with medications: time is of the essence

Real-World Dispensing Patterns of Inhalation Medication in Young Adult Asthma: An Inception Cohort Study

Purpose

The Global Initiative for Asthma (GINA) suggests a step-wise approach for pharmacological treatment of asthma. Valid study of real-world treatment patterns using dispensing databases includes proper measurement of medication adherence. We aim to explore such patterns by applying a time-varying proportion of days covered (tPDC)-based algorithm.

Patients and Methods

We designed a retrospective inception cohort study using the University of Groningen IADB.nl community pharmacy dispensing database. Included were 19,184 young adults who initiated asthma medication anywhere between 1994 and 2021, in the Netherlands. Main treatment steps were defined as: 1 - SABA/ICS-formoterol as needed, 2 - low dose ICS, 3 - low dose ICS + LABA or tiotropium, or intermediate dose ICS, 4 - intermediate to high dose ICS + LABA or tiotropium, triple therapy, or high dose ICS, 5 - treatment prescribed by a specialist. Changes in treatment steps were determined using a time-varying proportion of days covered (tPDC)-based algorithm. Individual drug treatment trajectories were visualized over time using a lasagna plot.

Results

At initiation, of the 19,184 included individuals, 52%, 7%, 15%, 16%, and 10% started treatment in steps 1 to 5, respectively. The median (IQR) follow-up time was 3 (1-7) years. Median (IQR) number of switches was 1 (0-3). Comparing starting step to last observed step, 37% never switched between treatment steps, 20% of individuals stepped down and 22% stepped up.

Conclusion

The low proportion of treatment switches between steps indicates that tailoring of treatment to patients' needs might be suboptimal. The tPDC-based algorithm functions well in translating dispensing data into continuous drug-utilization data, enabling a more granular assessment of treatment patterns among asthma patients.

Methodological considerations on estimating medication adherence from self-report, electronic monitoring, and electronic healthcare databases using the TEOS framework

AIM

Measuring adherence to medication is complex due to the diversity of contexts in which medications are prescribed, dispensed, and used. The Timelines-Events-Objectives-Sources (TEOS) framework outlined a process to operationalize adherence. We aimed to develop practical recommendations for quantification of medication adherence using self-report (SR), electronic monitoring (EM), and electronic healthcare databases (EHD) consistent with the TEOS framework for adherence operationalization.

METHODS

An adherence methodology working group of the International Society for Medication Adherence (ESPACOMP) analysed implications of the process of medication adherence for all data sources and discussed considerations specific to SR, ED, and EHD regarding the information available on the prescribing, dispensing, recommended and actual use timelines, the four events relevant for distinguishing the adherence phases, the study objectives commonly addressed with each type of data, and the potential sources of measurement error and quality criteria applicable.

RESULTS

Four key implications for medication adherence measurement are common to all data sources: adherence is a comparison between two series of events (recommended and actual use); it refers to one or more specific medication(s); it applies to regular repeated events coinciding with known recommended dosing; and it requires separate measurement of the three adherence phases for a complete picture of patients' adherence. We propose recommendations deriving from these statements, and aspects to be considered in study design when measuring adherence with SR, EM and EHD using the TEOS framework.

CONCLUSION

The quality of medication adherence estimates is the result of several design choices that may optimize the data available.

Accurate Medication Adherence Measurement Using Administrative Data for Frequently Hospitalized Patients

Background: Pharmacy administrative claims data remain an accessible and efficient source to measure medication adherence for frequently hospitalized patient populations that are systematically excluded from the landmark drug trials. Published pharmacotherapy studies use medication possession ratio (MPR) and proportion of days covered (PDC) to calculate medication adherence and usually fail to incorporate hospitalization and prescription overlap/gap from claims data. To make the cacophony of adherence measures clearer, this study created a refined hospital-adjusted algorithm to capture pharmacotherapy adherence among patients with end-stage renal disease (ESRD). Methods: The United States Renal Data System (USRDS) registry of ESRD was used to determine prescription-filling patterns of those receiving new prescriptions for oral P2Y12 inhibitors (P2Y12-I) between 2011 and 2015. P2Y12-I-naïve patients were followed until death, kidney transplantation, discontinuing medications, or loss to follow-up. After flagging/censoring key variables, the algorithm adjusted for hospital length of stay (LOS) and medication overlap. Hospital-adjusted medication adherence (HA-PDC) was calculated and compared with traditional MPR and PDC methods. Analyses were performed with SAS software. Results: Hospitalization occurred for 78% of the cohort (N = 46 514). The median LOS was 12 (interquartile range [IQR] = 2-34) days. MPR and PDC were 61% (IQR = 29%-94%) and 59% (IQR = 31%-93%), respectively. After applying adjustments for overlapping coverage days and hospital stays independently, HA-PDC adherence values changed in 41% and 52.7% of the cohort, respectively. When adjustments for overlap and hospital stay were made concurrently, HA-PDC adherence values changed in 68% of the cohort by 5.8% (HA-PDC median = 0.68, IQR = 0.31-0.93). HA-PDC declined over time (3M-6M-9M-12M). Nearly 48% of the cohort had a ≥30 days refill gap in the first 3 months, and this increased over time (P < .0001). Conclusions: Refill gaps should be investigated carefully to capture accurate pharmacotherapy adherence. HA-PDC measures increased adherence substantially when adjustments for hospital stay and medication refill overlaps are made. Furthermore, if hospitalizations were ignored for medications that are included in Medicare quality measures, such as Medicare STAR program, the apparent reduction in adherence might be associated with lower quality and health plan reimbursement.

Treatment Sequences After Discontinuing a Tumor Necrosis Factor Inhibitor in Patients With Rheumatoid Arthritis: A Comparison of Cycling Versus Swapping Strategies

OBJECTIVE

To evaluate the sequences of tumor necrosis factor inhibitors (TNFi) and non-TNFi used by rheumatoid arthritis (RA) patients whose initial TNFi therapy has failed, and to evaluate effectiveness and costs.

METHODS

Using the Truven Health MarketScan Research database, we analyzed claims of commercially insured adult patients with RA who switched to their second biologic or targeted disease-modifying antirheumatic drug between January 2008 and December 2015. Our primary outcome was the frequency of treatment sequences. Our secondary outcomes were the time to therapy discontinuation, drug adherence, and drug and other health care costs.

RESULTS

Among 10,442 RA patients identified, 36.5% swapped to a non-TNFi drug, most commonly abatacept (54.2%). The remaining 63.5% cycled to a second TNFi, most commonly adalimumab (41.2%). For subsequent switches of therapy, non-TNFi were more common. Patients who swapped to a non-TNFi were significantly older and had more comorbidities than those who cycled to a TNFi (P < 0.001). Survival analysis showed a longer time to discontinuation for non-TNFi than for TNFi (median 605 days compared with 489 days; P < 0.001) when used after initial TNFi discontinuation, but no difference in subsequent switches of therapy. Although non-TNFi were less expensive for adherent patients, cycling to a TNFi was associated with lower costs overall.

CONCLUSION

Even though patients are more likely to cycle to a second TNFi than swap to a non-TNFi, those who swap to a non-TNFi are more likely to persist with the therapy. However, cycling to a TNFi is the less costly strategy.

Classification of drug use patterns

Characterizing long‐term prescription data is challenging due to the time‐varying nature of drug use. Conventional approaches summarize time‐varying data into categorical variables based on simple measures, such as cumulative dose, while ignoring patterns of use. The loss of information can lead to misclassification and biased estimates of the exposure‐outcome association. We introduce a classification method to characterize longitudinal prescription data with an unsupervised machine learning algorithm. We used administrative databases covering virtually all 1.3 million residents of Manitoba and explicitly designed features to describe the average dose, proportion of days covered (PDC), dose change, and dose variability, and clustered the resulting feature space using K‐means clustering. We applied this method to metformin use in diabetes patients. We identified 27,786 metformin users and showed that the feature distributions of their metformin use are stable for varying the lengths of follow‐up and that these distributions have clear interpretations. We found six distinct metformin user groups: patients with intermittent use, decreasing dose, increasing dose, high dose, and two medium dose groups (one with stable dose and one with highly variable use). Patients in the varying and decreasing dose groups had a higher chance of progression of diabetes than other patients. The method presented in this paper allows for characterization of drug use into distinct and clinically relevant groups in a way that cannot be obtained from merely classifying use by quantiles of overall use.

Dynamic monitoring of the effects of adherence to medication on survival in heart failure patients: A joint modeling approach exploiting time-varying covariates

Adherence to medication is the process by which patients take their drugs as prescribed, and represents an issue in pharmacoepidemiological studies. Poor adherence is often associated with adverse health conditions and outcomes, especially in case of chronic diseases such as heart failure (HF). This turns out in an increased request for health care services, and in a greater burden for the health care system. In recent years, there has been a substantial growth in pharmacotherapy research, aimed at studying effects and consequences of proper/improper adherence to medication both for the increasing awareness of the problem and for the pervasiveness of poor adherence among patients. However, the way adherence is computed and accounted for into predictive models is far from being informative as it may be. In fact, it is usually analyzed as a fixed baseline covariate, without considering its time-varying behavior. The purpose and novelty of this study is to define a new personalized monitoring tool exploiting time-varying definition of adherence to medication, within a joint modeling approach. In doing so, we are able to capture and quantify the association between the longitudinal process of dynamic adherence to medication with the long-term survival outcome. Another novelty of this approach consists of exploiting the potential of health care administrative databases in order to reconstruct the dynamics of drugs consumption through pharmaceutical administrative registries. In particular, we analyzed administrative data provided by Regione Lombardia - Healthcare Division related to patients hospitalized for HF between 2000 and 2012.

Factors associated with medication persistence among ischemic stroke patients: a systematic review

OBJECTIVE

An investigation of the prevalence of medication persistence and associated factors in order to inform effective strategies for improving medication persistence.

METHODS

A systematic review of the literature from 2010 to the present was performed, using the PRISMA protocol. Primary and empirical observational studies of adult ischemic stroke or transient ischemic attack patients were included. PubMed, CINAHL, Web of Science, Cochrane Library, and PsycInfo databases were searched using the key terms stroke, ischemic stroke, medication persistence, medication adherence, and patient compliance.

RESULTS

Of four hundred twenty-eight journal articles retrieved, a final 18 articles were included. Short-term medication persistence was 46.2-96.7%, and long-term medication persistence was 41.7-93.0%. Identified hospital-related factors for medication persistence were stroke unit care, in-hospital medical complications, and early follow-up visit. Demographic factors for medication persistence were older age, and high/adequate financial status; disease-related factors were disease history, stroke subtype, and symptom severity. Age less than 75, female sex, comorbidity, antiplatelet medication switch, and polypharmacy were identified as factors of medication nonpersistence.

CONCLUSIONS

Stroke patients' medication persistence decreases over time, and persistence on antiplatelets, anticoagulants, and statin was poor. Several factors were associated with medication persistence, and these factors should be considered in future secondary preventative strategies.

Beyond Adherence Thresholds: A Simulation Study of the Optimal Classification of Longitudinal Adherence Trajectories From Medication Refill Histories

Background: The description of adherence based on medication refill histories relies on the estimation of continuous medication availability (CMA) during an observation period. Thresholds to distinguish adherence from non-adherence typically refer to an aggregated value across the entire observation period, disregarding differences in adherence over time. Sliding windows to divide the observation period into smaller portions, estimating adherence for these increments, and classify individuals with similar trajectories into clusters can retain this temporal information. Optimal methods to estimate adherence trajectories to identify underlying patterns have not yet been established. This simulation study aimed to provide guidance for future studies by analyzing the effect of different longitudinal adherence estimates, sliding window parameters, and sample characteristics on the performance of a longitudinal clustering algorithm.

Methods: We generated samples of 250–25,000 individuals with one of six longitudinal refill patterns over a 2-year period. We used two longitudinal CMA estimates (LCMA1 and LCMA2) and their dichotomized variants (with a threshold of 80%) to create adherence trajectories. LCMA1 assumes full adherence until the supply ends while LCMA2 assumes constant adherence between refills. We assessed scenarios with different LCMA estimates and sliding window parameters for 350 independent samples. Individual trajectories were clustered with kml, an implementation of k-means for longitudinal data in R. We compared performance between the four LCMA estimates using the adjusted Rand Index (cARI).

Results: Cluster analysis with LCMA2 outperformed other estimates in overall performance, correct identification of groups, and classification accuracy, irrespective of sliding window parameters. Pairwise comparison between LCMA estimates showed a relative cARI-advantage of 0.12–0.22 (p < 0.001) for LCMA2. Sample size did not affect overall performance.

Conclusion: The choice of LCMA estimate and sliding window parameters has a major impact on the performance of a clustering algorithm to identify distinct longitudinal adherence trajectories. We recommend (a) to assume constant adherence between refills, (b) to avoid dichotomization based on a threshold, and (c) to explore optimal sliding windows parameters in simulation studies or selecting shorter non-overlapping windows for the identification of different adherence patterns from medication refill data.

Patient-related characteristics associated with non-persistence with statin therapy in elderly patients following an ischemic stroke

PURPOSE

This study was aimed at evaluating the extent of non-persistence with statin therapy in elderly patients after an ischemic stroke and identifying patient-related characteristics that are risk factors for non-persistence.

METHODS

The evaluable study cohort (n = 2748) was derived from the database of the largest health insurance provider in the Slovak Republic. Patients aged ≥65 years who were initiated on statin therapy following the diagnosis of an ischemic stroke during one full year (1 January 2010 to 31 December 2010) constituted this cohort. Each patient was followed for a period of 3 years from the date of the first statin prescription. Patients with a continuous treatment gap of 6 months without statin prescription were designated as non-persistent. The Cox proportional hazard model was applied to determine patient-associated characteristics that influenced the likelihood of non-persistence.

RESULTS

During the 3-year follow-up period, 39.7% of patients in the study cohort became non-persistent. Factors associated with decreased probability of a patient becoming non-persistent were age ≥75 years (hazard ratio (HR) 0.75), polypharmacy (concurrent use of ≥6 drugs) (HR 0.79), diabetes mellitus (HR 0.80), dementia (HR 0.81) and hypercholesterolemia (HR 0.50). On the other hand, the presence of anxiety disorders (HR 1.33) predicted an increased likelihood of a patient being non-persistent.

CONCLUSIONS

Our findings suggest that patients aged ≥75 years or those with the presence of diabetes mellitus, dementia, hypercholesterolemia or polypharmacy were likely to be persistent with statin therapy, whereas those with anxiety disorders may need greater assistance with persistence of statin therapy. Copyright © 2016 John Wiley & Sons, Ltd.

Characteristics associated with nonadherence to medications for hypertension, diabetes, and dyslipidemia among breast cancer survivors

OBJECTIVE

Comorbidity among breast cancer survivors is prevalent, and adherence to medication management of comorbidities may be important for both chronic disease and cancer-related outcomes. Our objective was to determine characteristics associated with nonadherence to common chronic medications among breast cancer survivors.

METHODS

We conducted a retrospective cohort study of 4216 women in an integrated care system diagnosed with early-stage breast cancer between 1990 and 2008 and alive without recurrence or second primary breast cancer in the second-year following diagnosis. Adherence to antihypertensives, oral diabetes medications, and statins was measured during the second-year post-breast cancer diagnosis using medication possession ratios (MPR). Nonadherence was defined as MPR <0.80. We estimated odds ratios (OR) and 95% confidence intervals (CI) for nonadherence to antihypertensives, oral diabetes medications, and statins by various characteristics using multivariable logistic regression.

RESULTS

Among 2308 users of antihypertensives (n = 1779), diabetes medications (n = 499) and/or statins (n = 1072), 37% were nonadherent to antihypertensives; 75% were nonadherent to diabetes medications; 39% were nonadherent to statins. In adjusted models, younger age was associated with nonadherence to all three therapeutic classes. Certain cancer treatments were associated with nonadherence to antihypertensives (radiation: OR 1.21, 95% CI 1.01-1.47; endocrine therapy: OR 1.27, 95% CI 1.03-1.52) and diabetes medications (chemotherapy: OR 1.69, 95% CI 1.17-2.21). Less frequent primary care provider visits were associated with higher odds of nonadherence to antihypertensives (OR 1.70, 95% CI 1.19-2.22); and trends showed higher Charlson comorbidity scores associated with greater adherence to diabetes medications (P < 0.001) and statins (P = 0.032).

CONCLUSIONS

Nonadherence to medications is high among breast cancer survivors, particularly diabetes medications, and is associated with cancer treatments and other patient characteristics. Additional research is warranted to understand the needs of cancer patients taking chronic medications and explore patient and provider factors influencing adherence.

Inhaled Corticosteroid Adherence Patterns in a Longitudinal Asthma Cohort

BACKGROUND

Electronic prescribing records can enable exploration of medication adherence, but analysis decisions may influence estimates and require alignment to new consensus-based definitions.

OBJECTIVE

To compare different computations of inhaled corticosteroid (ICS) implementation in a primary care asthma population initiating ICS therapy when assessed within episodes of persistent use, and examine longitudinal variation in implementation.

METHODS

A historical cohort study was conducted on UK's Optimum Patient Care Research Database. Eligible patients had physician-diagnosed asthma, initiated ICS therapy, and had 3 or more years of continuous registration. ICS treatment episodes were constructed on the basis of 3 definitions, permitting 30-, 90-, and 182-day gaps between prescriptions. Implementation was estimated using 2 continuous medication availability (CMA I and II) definitions to explore effects of carryover of previous prescriptions in 4 observation windows: 6, 8, 12, and 24 months. Impact of methodology was assessed by descriptive statistics, linear mixed models, and measures of agreement.

RESULTS

A total of 13,922 eligible patients (mean age, 39.9 years; 48.7% men) were identified. For CMA I, permitting a 90-day gap, mean ICS implementation for the 2-year period was 89.3% (±16.0%; range, 14.4%-100%). Sensitivity analyses with 30- and 182-day gaps resulted in increased (97.0% ± 7.2%) and decreased (81.1% ± 21.6%) estimates. CMA II produced estimates with varying concordance (0.69-0.87). Substantial variance was found between and within patients (intraclass coefficient, 0.30-0.36).

CONCLUSIONS

Different analysis choices resulted in substantial variation in implementation estimates, highlighting the need for transparent and clinically relevant methododology. Distinguishing between (non)persistence and implementation is important in clinical practice, and may require different interventions in routine consultations.