Comparison of efavirenz levels in blood and hair with pharmacy refills as measures of adherence and predictors of viral suppression among people living with HIV in Nigeria.

BACKGROUND: Strategies to support adherence are constrained by the lack of tools to objectively monitor medication intake in low-resource settings. Pharmacologic measures are objective, but pharmacy refill data is more accessible and cost-efficient. This study compared short-term and long-term efavirenz (EFV) drug levels with pharmacy refill adherence data (PRA) and evaluated their ability to predict viral suppression among people living with HIV in Nigeria. METHODS: Paired hair and dried blood spot (DBS) samples were obtained from 91 adults living with HIV receiving 600 mg EFV-based antiretroviral therapy (ART) and EFV concentrations were measured via validated methods using liquid-chromatography-mass-spectrometry. PRA was estimated from pharmacy records, based on the number of days a patient collected medication before or after the scheduled pick-up date. PRA was categorized into ≤ 74%, 75-94% and ≥ 95%, defined as poor, medium and high adherence, respectively. HIV viral loads closest to the hair sampling time (within 6 months) were also abstracted. Receiver Operating Characteristics (ROC) curve analyses compared the ability of adherence metrics to predict viral suppression. RESULTS: Based on PRA, 81% of participants had high adherence while 11% and 8% had medium and poor adherence, respectively. The median (IQR) EFV concentrations were 6.85 ng/mg (4.56-10.93) for hair and 1495.6 ng/ml (1050.7-2365.8) for DBS. Of the three measures of adherence, hair EFV concentration had the highest Area Under Curve (AUC) to predict viral suppression. Correlations between EFV concentrations in DBS and hair with PRA were positive (r = 0.12, P = 0.27 and r = 0.21, P = 0.05, respectively) but not strong. CONCLUSIONS: EFV concentrations in hair were the strongest predictor of viral suppression and only weakly correlated with pharmacy refill adherence data in Nigeria. This study suggests that resource-limited settings may benefit from objective adherence metrics to monitor and support adherence.

Self-reported medication adherence and pharmacy refill adherence among persons with ischemic stroke: a cross-sectional study.

PURPOSE: To describe and compare self-reported medication adherence assessed with the 5-item version of Medication Adherence Report Scale (MARS-5) and pharmacy refill adherence based on data from the Swedish Prescribed Drug Register (SPDR) among persons with ischemic stroke, and to investigate independent predictors associated with respective assessments. METHODS: A study questionnaire was sent to persons with ischemic stroke registered in the Swedish Stroke Register between Dec 2011 and March 2012, and who lived at home 3 months after discharge. The primary outcome was dichotomized to adherent/non-adherent based on MARS-5 and SPDR and analyzed with multivariable logistic regression. Adherence according to MARS-5 was defined as score 23 or higher (out of 25). Adherence according to SPDR was defined as at least one filled statin prescription recorded in SPDR in each 6-month interval during 2 years of follow-up. RESULTS: Of 420 participants, 367 (87%) and 329 (78%) were adherent according to MARS-5 and SPDR, respectively, and 294 (70%) participants were adherent according to both assessments. A significant association was shown between medication adherence according to the two assessments (p = 0.020). Independent predictors associated with medication adherence according to MARS-5 were female sex, while factors associated with SPDR were male sex and being younger. CONCLUSIONS: The majority of participants were classified as adherent, 87% according to MARS-5 and 78% based on data from SPDR. However, only 70% were adherent according to both MARS-5 and SPDR, and different predictors were associated with the different measurements, suggesting that these assessments are measuring different aspects of adherence.

Pitfalls of medication adherence approximation through EHR and pharmacy records: Definitions, data and computation.

BACKGROUND AND PURPOSE: Patients' adherence to medication is a complex, multidimensional phenomenon. Dispensation data and electronic health records are used to approximate medication-taking through refill adherence. In-depth discussions on the adverse effects of data quality and computational differences are rare. The purpose of this article is to evaluate the impact of common pitfalls when computing medication adherence using electronic health records. PROCEDURES: We point out common pitfalls associated with the data and operationalization of adherence measures. We provide operational definitions of refill adherence and conduct experiments to determine the effect of the pitfalls on adherence estimations. We performed statistical significance testing on the impact of common pitfalls using a baseline scenario as reference. FINDINGS: Slight changes in definition can significantly skew refill adherence estimates. Pickup patterns cause significant disagreement between measures and the commonly used proportion of days covered. Common data related issues had a small but statistically significant (p < 0.05) impact on population-level and significant effect on individual cases. CONCLUSION: Data-related issues encountered in real-world administrative databases, which affect various operational definitions of refill adherence differently, can significantly skew refill adherence values, leading to false conclusions about adherence, particularly when estimating adherence for individuals.

Prediction and pattern analysis of medication refill adherence through electronic health records and dispensation data.

BACKGROUND AND PURPOSE: Low adherence to medication in chronic disease patients leads to increased morbidity, mortality, and healthcare costs. The widespread adoption of electronic prescription and dispensation records allows a more comprehensive overview of medication utilization. In combination with electronic health records (EHR), such data provides new opportunities for identifying patients at risk of nonadherence and provide more targeted and effective interventions. The purpose of this article is to study the predictability of medication adherence for a cohort of hypertensive patients, focusing on healthcare utilization factors under various predictive scenarios. Furthermore, we discover common proportion of days covered patterns (PDC-patterns) for patients with index prescriptions and simulate medication-taking behaviours that might explain observed patterns. PROCEDURES: We predict refill adherence focusing on factors of healthcare utilization, such as visits, prescription information and demographics of patient and prescriber. We train models with machine learning algorithms, using four different data splits: stratified random, patient, temporal forward prediction with and without index patients. We extract frequent, two-year long PDC-patterns using K-means clustering and investigate five simple models of medication-taking that can generate such PDC-patterns. FINDINGS: Model performance varies between data splits (AUC test set: 0.77-0.89). Including historical information increases the performance slightly in most cases (approx. 1-2% absolute AUC uplift). Models show low predictive performance (AUC test set: 0.56-0.66) on index-prescriptions and patients with sudden drops in PDC (Recall: 0.58-0.63). We find 21 distinct two-year PDC-patterns, ranging from good adherence to intermittent gaps and early discontinuation in the first or second year. Simulations show that observed PDC-patterns can only be explained by specific medication consumption behaviours. CONCLUSIONS: Prediction models developed using EHR exhibit bias towards patients with high healthcare utilization. Even though actual medication-taking is not observable, consumption patterns may not be as arbitrary, provided that medication refilling and consumption is linked.

Impact of Social Determinants of Health and Demographics on Refill Requests by Medicare Patients Using a Conversational Artificial Intelligence Text Messaging Solution: Cross-Sectional Study.

BACKGROUND: Nonadherence among patients with chronic disease continues to be a significant concern, and the use of text message refill reminders has been effective in improving adherence. However, questions remain about how differences in patient characteristics and demographics might influence the likelihood of refill using this channel. OBJECTIVE: The aim of this study was to evaluate the efficacy of an SMS-based refill reminder solution using conversational artificial intelligence (AI; an automated system that mimics human conversations) with a large Medicare patient population and to explore the association and impact of patient demographics (age, gender, race/ethnicity, language) and social determinants of health on successful engagement with the solution to improve refill adherence. METHODS: The study targeted 99,217 patients with chronic disease, median age of 71 years, for medication refill using the mPulse Mobile interactive SMS text messaging solution from December 2016 to February 2019. All patients were partially adherent or nonadherent Medicare Part D members of Kaiser Permanente, Southern California, a large integrated health plan. Patients received SMS reminders in English or Spanish and used simple numeric or text responses to validate their identity, view their medication, and complete a refill request. The refill requests were processed by Kaiser Permanente pharmacists and support staff, and refills were picked up at the pharmacy or mailed to patients. Descriptive statistics and predictive analytics were used to examine the patient population and their refill behavior. Qualitative text analysis was used to evaluate quality of conversational AI. RESULTS: Over the course of the study, 273,356 refill reminders requests were sent to 99,217 patients, resulting in 47,552 refill requests (17.40%). This was consistent with earlier pilot study findings. Of those who requested a refill, 54.81% (26,062/47,552) did so within 2 hours of the reminder. There was a strong inverse relationship (r10=-0.93) between social determinants of health and refill requests. Spanish speakers (5149/48,156, 10.69%) had significantly lower refill request rates compared with English speakers (42,389/225,060, 18.83%; X21 [n=273,216]=1829.2; P<.001). There were also significantly different rates of refill requests by age band (X26 [n=268,793]=1460.3; P<.001), with younger patients requesting refills at a higher rate. Finally, the vast majority (284,598/307,484, 92.23%) of patient responses were handled using conversational AI. CONCLUSIONS: Multiple factors impacted refill request rates, including a strong association between social determinants of health and refill rates. The findings suggest that higher refill requests are linked to language, race/ethnicity, age, and social determinants of health, and that English speakers, whites, those younger than 75 years, and those with lower social determinants of health barriers are significantly more likely to request a refill via SMS. A neural network-based predictive model with an accuracy level of 78% was used to identify patients who might benefit from additional outreach to narrow identified gaps based on demographic and socioeconomic factors.

Risk of cardiovascular event and mortality in relation to refill and guideline adherence to lipid-lowering medications among patients with type 2 diabetes mellitus in Sweden.

Objective: To analyze the risk of cardiovascular (CV) events and mortality in relation to adherence to lipid-lowering medications by healthcare centers and patients with type 2 diabetes mellitus (T2DM). Research design and methods: We included 121 914 patients (12% secondary prevention) with T2DM reported by 1363 healthcare centers. Patients initiated lipid-lowering medications between July 2006 and December 2012 and were followed from cessation of the first filled supply until multidose dispensed medications, migration, CV events, death or December 2016. The study period was divided into 4-month intervals through 2014, followed by annual intervals through 2016. Adherence measures were assessed for each interval. Patients' (refill) adherence was measured using the medication possession ratio (MPR). Healthcare centers' (guideline) adherence represented the prescription prevalence of lipid-lowering medications according to guidelines. The risk of CV events and mortality was analyzed for each interval using Cox proportional hazard regression and Kaplan-Meier. Results: Compared with high-adherent patients (MPR >80%), low-adherent primary prevention patients (MPR ≤80%) showed higher risk of all outcomes: 44%-51 % for CV events, doubled for all-cause mortality and 79%-90% for CV mortality. Corresponding risks for low-adherent secondary prevention patients were 17%-19% for CV events, 88%-97% for all-cause and 66%-79% for CV mortality. Primary prevention patients treated by low-adherent healthcare centers (guideline adherence <48%) had a higher risk of CV events and CV mortality. Otherwise, no difference in the risk of CV events or mortality was observed by guideline adherence level. Conclusions: Our results demonstrate the importance of high refill adherence and thus the value of individualized care among patients with T2DM.

Association between refill adherence to lipid-lowering medications and the risk of cardiovascular disease and mortality in Swedish patients with type 2 diabetes mellitus: a nationwide cohort study.

OBJECTIVES: To analyse the association between refill adherence to lipid-lowering medications, and the risk of cardiovascular disease (CVD) and mortality in patients with type 2 diabetes mellitus. DESIGN: Cohort study. SETTING: National population-based cohort of Swedish patients with type 2 diabetes mellitus. PARTICIPANTS: 86 568 patients aged ≥18 years, registered with type 2 diabetes mellitus in the Swedish National Diabetes Register, who filled at least one prescription for lipid-lowering medication use during 2007-2010, 87% for primary prevention. EXPOSURE AND OUTCOME MEASURES: Refill adherence of implementation was assessed using the medication possession ratio (MPR), representing the proportion of days with medications on hand during an 18-month exposure period. MPR was categorised by five levels (≤20%, 21%-40%, 41%-60%, 61%-80% and >80%). Patients without medications on hand for ≥180 days were defined as non-persistent. Risk of CVD (myocardial infarction, ischaemic heart disease, stroke and unstable angina) and mortality by level of MPR and persistence was analysed after the exposure period using Cox proportional hazards regression and Kaplan-Meier, adjusted for demographics, socioeconomic status, concurrent medications and clinical characteristics. RESULTS: The hazard ratios for CVD ranged 1.33-2.36 in primary prevention patients and 1.19-1.58 in secondary prevention patients, for those with MPR ≤80% (p<0.0001). The mortality risk was similar regardless of MPR level. The CVD risk was 74% higher in primary prevention patients and 33% higher in secondary prevention patients, for those who were non-persistent (p<0.0001). The mortality risk was 6% higher in primary prevention patients and 18% higher in secondary prevention patients, for non-persistent patients (p<0.0001). CONCLUSIONS: Higher refill adherence to lipid-lowering medications was associated with lower risk of CVD in primary and secondary prevention patients with type 2 diabetes mellitus.

Refill adherence and persistence to lipid-lowering medicines in patients with type 2 diabetes: A nation-wide register-based study.

PURPOSE: This study aimed to describe and compare refill adherence and persistence to lipid-lowering medicines in patients with type 2 diabetes by previous cardiovascular disease (CVD). METHODS: We followed 97 595 patients (58% men; 23% with previous CVD) who were 18 years of age or older when initiating lipid-lowering medicines in 2007-2010 until first fill of multi-dose dispensed medicines, death, or 3 years. Using personal identity numbers, we linked individuals' data from the Swedish Prescribed Drug Register, the Swedish National Diabetes Register, the National Patient Register, the Cause of Death Register, and the Longitudinal Integration Database for Health Insurance and Labour Market Studies. We assessed refill adherence using the medication possession ratio (MPR) and the maximum gap method, and measured persistence from initiation to discontinuation of treatment or until 3 years after initiation. We analyzed differences in refill adherence and persistence by previous CVD in multiple regression models, adjusted for socioeconomic status, concurrent medicines, and clinical characteristics. RESULTS: The mean age of the study population was 64 years, 80% were born in Sweden, and 56% filled prescriptions for diabetes medicines. Mean MPR was 71%, 39% were adherent according to the maximum gap method, and mean persistence was 758 days. Patients with previous CVD showed higher MPR (3%) and lower risk for discontinuing treatment (12%) compared with patients without previous CVD (P < 0.0001). CONCLUSIONS: Patients with previous CVD were more likely to be adherent to treatment and had lower risk for discontinuation compared with patients without previous CVD.

Non-adherence to inhaled corticosteroids and the risk of asthma exacerbations in children.

BACKGROUND: Non-adherence to inhaled corticosteroids (ICSs) is a major risk factor for poor asthma control in children. However, little is known about the effect of adherence to ICS on the incidence of asthma exacerbations. The objective of this study was to examine the effect of poor adherence to ICS on the risk of exacerbations in children with asthma. METHODS: In this nested case-control study using data from the Dutch PHARMO Record Linkage System, children aged 5-12 years who had an asthma exacerbation needing oral corticosteroids or hospital admission were matched to patients without exacerbations. Refill adherence was calculated as medication possession ratio from ICS-dispensing records. Data were analyzed using a multivariable multiplicative intensity regression model. RESULTS: A total of 646 children were included, of whom 36 had one or more asthma exacerbations. The medication possession ratio was 67.9% (standard deviation [SD] 30.2%) in children with an exacerbation versus 54.2% (SD 35.6%) in the control group. In children using long-acting beta-agonist, good adherence to ICS was associated with a higher risk of asthma exacerbations: relative risk 4.34 (95% confidence interval: 1.20-15.64). CONCLUSION: In children with persistent asthma needing long-acting beta-agonist, good adherence to ICS was associated with an increased risk of asthma exacerbations. Possible explanations include better motivation for adherence to ICS in children with more severe asthma, and reduced susceptibility to the consequences of non-adherence to ICS due to overprescription of ICS to children who are in clinical remission. Further study into the background of the complex interaction between asthma and medication adherence is needed.

Medication oversupply in patients with diabetes.

BACKGROUND: Studies in integrated health systems suggest that patients often accumulate oversupplies of prescribed medications, which is associated with higher costs and hospitalization risk. However, predictors of oversupply are poorly understood, with no studies in Medicare Part D. OBJECTIVE: The aim of this study was to describe prevalence and predictors of oversupply of antidiabetic, antihypertensive, and antihyperlipidemic medications in adults with diabetes managed by a large, multidisciplinary, academic physician group and enrolled in Medicare Part D or a local private health plan. METHODS: This was a retrospective cohort study. Electronic health record data were linked to medical and pharmacy claims and enrollment data from Medicare and a local private payer for 2006-2008 to construct a patient-quarter dataset for patients managed by the physician group. Patients' quarterly refill adherence was calculated using ReComp, a continuous, multiple-interval measure of medication acquisition (CMA), and categorized as <0.80 = Undersupply, 0.80-1.20 = Appropriate Supply, >1.20 = Oversupply. We examined associations of baseline and time-varying predisposing, enabling, and medical need factors to quarterly supply using multinomial logistic regression. RESULTS: The sample included 2519 adults with diabetes. Relative to patients with private insurance, higher odds of oversupply were observed in patients aged <65 in Medicare (OR = 3.36, 95% CI = 1.61-6.99), patients 65+ in Medicare (OR = 2.51, 95% CI = 1.37-4.60), patients <65 in Medicare/Medicaid (OR = 4.55, 95% CI = 2.33-8.92), and patients 65+ in Medicare/Medicaid (OR = 5.73, 95% CI = 2.89-11.33). Other factors associated with higher odds of oversupply included any 90-day refills during the quarter, psychotic disorder diagnosis, and moderate versus tight glycemic control. CONCLUSIONS: Oversupply was less prevalent than in previous studies of integrated systems, but Medicare Part D enrollees had greater odds of oversupply than privately insured individuals. Future research should examine utilization management practices of Part D versus private health plans that may affect oversupply.

Refill adherence and self-reported adverse drug reactions and sub-therapeutic effects: a population-based study.

PURPOSE: To assess refill adherence to dispensed oral long-term medications among the adult population and to investigate whether the percentages of self-reported adverse drug reactions (ADRs) and sub-therapeutic effects (STEs) differed for medications with adequate refill adherence, oversupply, and undersupply. METHOD: Survey responses on self-reported ADRs and STEs were linked to the Swedish Prescribed Drug Register in a cross-sectional population-based study. Refill adherence to antihypertensive, lipid-lowering, and oral anti-diabetic medications was measured using the continuous measure of medication acquisition (CMA). The percentages of self-reported ADRs and STEs were compared between medications with adequate refill adherence (CMA 0.8-1.2), oversupply (CMA > 1.2), and undersupply (CMA < 0.8). RESULTS: The study included 1827 persons, and the refill adherence was measured for 3014 antihypertensive, 839 lipid lowering, and 253 oral anti-diabetic medications. Overall, 65.7% of the medications had adequate refill adherence, 21.9% oversupply, and 12.4% undersupply. The percentages of self-reported ADRs and STEs were respectively 2.6%, 2.7%, and 2.1% (p > 0.5) for ADRs and 1.1%, 1.6%, and 1.5% (p > 0.5) for STEs. CONCLUSIONS: Adequate refill adherence was found in two thirds of the medication therapies. ADRs and STEs were unexpectedly equally commonly reported for medications with adequate refill adherence, oversupply, and undersupply. These results suggest that a better understanding of patients' refill behaviors and their perceived medication adverse outcomes is needed and should be considered in improving medication management. The impact of individual and healthcare factors that may influence the association between refill adherence and reported medication adverse outcomes should be investigated in future studies.