Actual versus prescribed timing of lovastatin doses assessed by electronic compliance monitoring

Effect of timing of administration on lipid-lowering efficacy of statins-meta-analysis

BACKGROUND

To investigate the effect of timing of statin administration on lipid-lowering efficacy.

METHODS

Computer searches of Pubmed, Embase, Cochrane Library, and Web of Science databases from 1986 to 2023. The impact of administration time on the lipid-lowering efficacy of statin drugs was investigated. Following a series of screenings, a funnel plot was constructed to assess its symmetry, and Egger and Beggar tests were conducted using StataMP-64 to evaluate publication bias. Meta-analysis was performed using RevMan 5.3 to combine MD values.

RESULTS

Fifteen papers (1352 participants) met and included the criteria. The results of the meta-analysis showed that the effect of morning and evening administration time on plasma triglycerides (TG) (P > 0.05) and plasma high-density lipoprotein cholesterol (HDL-C) (P > 0.05) was not statistically significant. There were significant reductions in total cholesterol (TC) (MD: 0.15 mmol/L, 95% CI: 0.06-0.23, P < 0.01) and low-density lipoprotein cholesterol (LDL-C) (MD: 0.10 mmol/L, 95% CI: - 0.00-0.20, P < 0.01) in the night group. According to the analysis results of the half-life of statins, only short half-life statins showed that nocturnal administration reduced LDL-C (MD: 0.21 mmol/L, 95% CI: 0.09-0.33, P < 0.01) and TC (MD: 0.32 mmol/L, 95% CI: 0.18-0.46, P < 0.01) levels and was better than morning administration. Long half-life statins did not show significant differences. In addition, the administration time of short half-life statins also showed that night administration tended to reduce TG (MD: 0.16 mmol/L, 95% CI: 0.02-0.30, P < 0.05) levels. In subgroup analysis according to clinical factors in patients aged < 55 years, there was no significant difference in the timing of administration between the two groups; the efficacy of statins in lowering lipids in patients aged ≥ 55 years was significantly different in the TC group (P < 0.01) and LDL-C group (P < 0.01). The administration time of the TC group (P < 0.05) and LDL-C group (P < 0.05) in the Americas, Europe, and Asian groups was significantly different for statins. In addition, the American group also showed that the administration time of the two groups was significantly different from the TG group (P < 0.05).

CONCLUSION

The efficacy of administering short half-life statin drugs at night in reducing plasma levels of TC, LDL-C, and TG surpasses that of morning administration. However, this study did not determine the impact of timing of statin administration in patients taking long half-life statins on the efficacy of the medication. Therefore, it is recommended to consider patient adherence when. The study was registered on PROSPERO (International Prospective Register of Systematic Reviews) as CRD42022372105 (available at https://www.crd.york.ac.uk/prospero/ ).

Statin Dosing Instructions, Medication Adherence, and Low-Density Lipoprotein Cholesterol: a Cohort Study of Incident Statin Users

BACKGROUND

Robust evidence is lacking on optimal timing of statin administration and its impact on patient outcomes.

OBJECTIVE

This study aims to evaluate among incident statin users the relationship between those prescribed evening vs. daily dosing instructions, medication adherence, and changes in low-density lipoprotein cholesterol (LDL-c).

DESIGN

This is an observational cohort study at Sutter Health, a community-based healthcare system, 2010-2016.

PARTICIPANTS

Patients were ≥ 35 years of age as of the first statin prescription (baseline), with 12 to 36 months of electronic health record activity before and after baseline. Incident use was defined as no statin prescription in 12 months prior to baseline.

MAIN MEASURES

Differences in medication adherence (proportion of days covered ≥ 0.80) over 12 months from baseline and mean change in LDL-c between 12 and 24 months from baseline were measured using regression modeling, adjusting for baseline demographics and clinical, prescriber, and statin characteristics.

KEY RESULTS

Among 31,252 patients with valid statin prescriptions between 2010 and 2016, 5099 eligible incident statin users (mean age, 63 years) were identified, of whom 53% were prescribed evening and 47% daily dosing instructions. No difference in likelihood of statin adherence over 12 months was observed for evening vs. daily dosing (adjusted odds ratio [OR] 0.90; 95% CI 0.75, 1.08). No differences were observed in mean change in LDL-c (adjusted mean difference 1.42 mg/dL; 95% CI - 1.02, 3.89) or likelihood of attaining LDL-c < 70 mg/dL (adjusted OR 0.83; 95% CI 0.67, 1.04) for evening vs. daily dosing over a mean of 19 months follow-up.

CONCLUSIONS

Among incident statin users from a real-world clinical setting, those with daily and evening dosing instructions had similar adherence rates and mean changes in LDL-c. Given potential clinical equipoise for evening and daily dosing, clinicians should consider patient-tailored statin dosing instructions to reduce potentially unnecessary regimen complexity.

A systematic literature review comparing methods for the measurement of patient persistence and adherence

OBJECTIVES

A systematic literature review was conducted comparing different approaches estimating persistence and adherence in chronic diseases with polypharmacy of oral and subcutaneous treatments.

METHODS

This work followed published guidance on performing systematic reviews. Twelve electronic databases and grey literature sources were used to identify studies and guidelines for persistence and adherence of oral and subcutaneous therapies in hypercholesterolemia, type 2 diabetes, hypertension, osteoporosis and rheumatoid arthritis. Outcomes of interest of each persistence and adherence data collection and calculation method included pros: accurate, easy to use, inexpensive; and cons: inaccurate, difficult to use, expensive.

RESULTS

A total of 4158 records were retrieved up to March 2017. We included 16 observational studies, 5 systematic reviews and 7 guidelines, in patients with hypercholesterolemia (n = 8), type 2 diabetes (n = 4), hypertension (n = 2), rheumatoid arthritis (n = 1) and mixed patient populations (n = 13). Pharmacy and medical records offer an accurate, easy and inexpensive data collection method. Pill count, medication event monitoring systems (MEMs), self-report questionnaires and observer report are easy to use. MEMS and biochemical monitoring tests can be expensive. Proportion of days covered (PDC) was recommended as a gold standard calculation method for long-term treatments. PDC avoids use of days' supply in calculation, hence is more accurate compared to medication possession ratio (MPR) to assess adherence to treatments in chronic diseases.

CONCLUSIONS

Decisions on what method to use should be based on considerations of the route of medication administration, the resources available, setting and aim of the assessment. Combining different methods may provide wider insights into adherence and persistence, including patient behavior.

The optimal time of day for statin administration: a review of current evidence

PURPOSE OF REVIEW

In humans, cholesterol biosynthesis varies diurnally, reaching its peak at night. Therefore, choosing the time of statin administration is critical because of their different half-lives. Dose timing becomes more important in patients with polypharmacy because it might affect their adherence to the statin therapy.

RECENT FINDINGS

Herein, we narratively summarized the available clinical studies (n = 17) and meta-analyses (n = 2) that compared the morning with the evening dose of statins in terms of safety and efficacy, with special focus on their low-density lipoprotein-lowering effects. We also explained the difference in efficacy results in case of short-acting compared with the long-acting statins and highlighted how flexibility in choosing the time of statin administration is important for better adherence.

SUMMARY

The current limited evidence suggests that short-acting statins should be given in the evening whereas long-acting statins could be given at any time of the day with allowing more patient-based choice (of timing) for better adherence. Lager RCTs with longer durations are recommended to extend and confirm the current evidence.

Chronotherapy versus conventional statins therapy for the treatment of hyperlipidaemia

BACKGROUND

Elevated levels of total cholesterol and low-density lipoprotein play an important role in the development of atheromas and, therefore, in cardiovascular diseases. Cholesterol biosynthesis follows a circadian rhythm and is principally produced at night (between 12:00 am and 6:00 am). The adjustment of hypolipaemic therapy to biologic rhythms is known as chronotherapy. Chronotherapy is based on the idea that medication can have different effects depending on the hour at which it is taken. Statins are one of the most widely used drugs for the prevention of cardiovascular events. In usual clinical practice, statins are administered once per day without specifying the time when they should be taken. It is unknown whether the timing of statin administration is important for clinical outcomes.

OBJECTIVES

To critically evaluate and analyse the evidence available from randomised controlled trials regarding the effects of chronotherapy on the effectiveness and safety of treating hyperlipidaemia with statins.

SEARCH METHODS

We searched the CENTRAL, MEDLINE, Embase, LILACS, ProQuest Health & Medical Complete, OpenSIGLE, Web of Science Conference Proceedings, and various other resources including clinical trials registers up to November 2015. We also searched the reference lists of relevant reviews for eligible studies.

SELECTION CRITERIA

We included randomised controlled trials (RCTs), enrolling people with primary or secondary hyperlipidaemia. To be included, trials must have compared any chronotherapeutic lipid-lowering regimen with statins and any other statin lipid-lowering regimen not based on chronotherapy. We considered any type and dosage of statin as eligible, as long as the control and experimental arms differed only in the timing of the administration of the same statin. Quasi-randomised studies were excluded.

DATA COLLECTION AND ANALYSIS

We used the standard methodological procedures expected by Cochrane. We extracted the key data from studies in relation to participants, interventions, and outcomes for safety and efficacy. We calculated odds ratios (OR) for dichotomous data and mean differences (MD) for continuous data with 95% confidence intervals (CI). Using the GRADE approach, we assessed the quality of the evidence and we used the GRADEpro Guideline Development Tool to import data from Review Manager to create 'Summary of findings' tables.

MAIN RESULTS

This review includes eight RCTs (767 participants analysed in morning and evening arms). The trials used different lipid-lowering regimens with statins (lovastatin: two trials; simvastatin: three trials; fluvastatin: two trials; pravastatin: one trial). All trials compared the effects between morning and evening statin administration. Trial length ranged from four to 14 weeks. We found a high risk of bias in the domain of selective reporting in three trials and in the domain of incomplete outcome data in one trial of the eight trials included. None of the studies included were judged to be at low risk of bias.None of the included RCTs reported data on cardiovascular mortality, cardiovascular morbidity, incidence of cardiovascular events, or deaths from any cause. Pooled results showed no evidence of a difference in total cholesterol (MD 4.33, 95% CI -1.36 to 10.01), 514 participants, five trials, mean follow-up 9 weeks, low-quality evidence), low-density lipoprotein cholesterol (LDL-C) levels (MD 4.85 mg/dL, 95% CI -0.87 to 10.57, 473 participants, five trials, mean follow-up 9 weeks, low-quality evidence), high-density lipoprotein cholesterol (HDL-C) (MD 0.54, 95% CI -1.08 to 2.17, 514 participants, five trials, mean follow-up 9 weeks, low-quality evidence) or triglycerides (MD -8.91, 95% CI -22 to 4.17, 510 participants, five trials, mean follow-up 9 weeks, low-quality evidence) between morning and evening statin administration.With regard to safety outcomes, five trials (556 participants) reported adverse events. Pooled analysis found no differences in statins adverse events between morning and evening intake (OR 0.71, 95% CI 0.44 to 1.15, 556 participants, five trials, mean follow-up 9 weeks, low-quality evidence).

AUTHORS' CONCLUSIONS

Limited and low-quality evidence suggested that there were no differences between chronomodulated treatment with statins in people with hyperlipidaemia as compared to conventional treatment with statins, in terms of clinically relevant outcomes. Studies were short term and therefore did not report on our primary outcomes, cardiovascular clinical events or death. The review did not find differences in adverse events associated with statins between both regimens. Taking statins in the evening does not have an effect on the improvement of lipid levels with respect to morning administration. Further high-quality trials with longer-term follow-up are needed to confirm the results of this review.

A scoping review of studies comparing the medication event monitoring system (MEMS) with alternative methods for measuring medication adherence

Different methods are available for measuring medication adherence. In this paper, we conducted a scoping review to identify and summarize evidence of all studies comparing the Medication Event Monitoring System (MEMS) with alternative methods for measuring medication adherence. A literature search was performed using the open database www.iAdherence.org that includes all original studies reporting findings from the MEMS. Papers comparing methods for measuring adherence to solid oral formulations were included. Data was extracted using a standardized extraction table. A total of 117 articles fulfilled the inclusion criteria, including 251 comparisons. Most frequent comparisons were against self-report (n = 119) and pill count (n = 59). Similar outcome measures were used in 210 comparisons (84%), among which 78 used dichotomous variables (adherent or not) and 132 used continuous measures (adherence expressed as percentage). Furthermore, 32% of all comparisons did not estimate adherence over the same coverage period and 44% of all comparisons did not use a statistical method or used a suboptimal one. Only eighty-seven (35%) comparisons had similar coverage periods, similar outcome measures and optimal statistical methods. Compared to MEMS, median adherence was grossly overestimated by 17% using self-report, by 8% using pill count and by 6% using rating. In conclusion, among all comparisons of MEMS versus alternative methods for measuring adherence, only a few used adequate comparisons in terms of outcome measures, coverage periods and statistical method. Researchers should therefore use stronger methodological frameworks when comparing measurement methods and be aware that non-electronic measures could lead to overestimation of medication adherence.

The partners for life program: a couples approach to cardiac risk reduction

Morbidity and mortality are reliably lower for the married compared with the unmarried across a variety of illnesses. What is less well understood is how a couple uses their relationship for recommended lifestyle changes associated with decreased risk for illness. Partners for Life compared a patient and partner approach to behavior change with a patient only approach on such factors as exercise, nutrition, and medication adherence. Ninety-three patients and their spouses/partners consented to participate (26% of those eligible) and were randomized into either the individual or couples condition. However, only 80 couples, distributed across conditions, contributed data to the analyses, due to missing data and missing data points. For exercise, there was a significant effect of couples treatment on the increase in activity and a significant effect of couples treatment on the acceleration of treatment over time. In addition, there was an interaction between marital satisfaction and treatment condition such that patients who reported higher levels of marital distress in the individuals condition did not maintain their physical activity gains by the end of treatment, while both distressed and nondistressed patients in the couples treatment exhibited accelerating gains throughout treatment. In terms of medication adherence, patients in the couples treatment exhibited virtually no change in medication adherence over time, while patients in the individuals treatment showed a 9% relative decrease across time. There were no condition or time effects for nutritional outcomes. Finally, there was an interaction between baseline marital satisfaction and treatment condition such that patients in the individuals condition who reported lower levels of initial marital satisfaction showed deterioration in marital satisfaction, while non satisfied patients in the couples treatment showed improvement over time.

Effect of dosing frequency on chronic cardiovascular disease medication adherence

BACKGROUND

Many cardiovascular diseases (CVDs) require patients to take one or more long term medications, often administered multiple times a day. We sought to determine the effect of chronic CVD medication dosing frequency on medication adherence.

METHODS

A search of Medline and Embase from 1986 to December 2011 was performed. Included studies used a prospective design, assessed adults with chronic CVDs, evaluated scheduled oral medications administered one to four times daily, and measured adherence for ≥1 month using an electronic monitoring device. Mixed linear model meta-regression was used to determine how dosing frequency affected adherence using three definitions of increasing strictness: taking, regimen and timing adherence.

RESULTS

A total of 29 studies, comprising 41, 29, and 27 dosing frequency arms for the taking, regimen and timing adherence definitions were included. Crude pooled adherence estimates were highest when the lenient taking definition was assessed (range for dosing frequencies: 80.1%-93.1%), and lowest when the strictest timing definition was assessed (range: 57.1%-76.3%). Upon meta-regression, the adjusted weighted mean percentage adherence for twice and three times daily dosing regimens (no studies evaluated four times daily regimens), were 6.9% and 13.7% lower than once daily regimens for the taking, 14.0% and 27.5% lower for the regimen, and 22.9% and 30.4% lower for the timing adherence definition (p < 0.01 for all).

LIMITATIONS

The presence of residual confounding and publication bias cannot be ruled out.

CONCLUSION

Patients appear to be more adherent with once daily dosing compared with more frequently scheduled chronic CVD medication regimens. This finding is magnified when more stringent definitions of adherence are used.

Adherence to medication in patients with dementia: predictors and strategies for improvement

Measures to facilitate patient medication adherence should be considered an integral part of the comprehensive care of older patients with multiple diseases. However, impairment of cognitive functions and dementia, in particular, may substantially compromise adherence behaviour. Therefore, a literature review was performed to identify factors associated with adherence to medication in patients with cognitive impairment or dementia, and to discuss strategies for improvement of non-adherence. Evidence-based information on how to deal with adherence to medication in patients with dementia is scarce because of a lack of specific studies. However, there is increasing knowledge about factors influencing medication adherence behaviour in older age, and emerging insight into the relationships between adherence behaviour and cognitive capacity, memory and executive function, in particular. Nevertheless, understanding elderly persons' strategies for maintaining regular use of even complex drug regimens is still limited. Progress of research in this field is needed. It is notable that measures to improve adherence consist of combinations of educational interventions and cognitive support but assessment of study participants' cognitive function is rare. In clinical practice, awareness of non-adherence as a result of cognitive impairment is relatively low. The most important step is early detection of cognitive impairment when this is impacting negatively on medication management. A practical geriatric screening test is recommended to identify memory problems and further functional impairments associated with cognitive impairment. Performance-based assessments might be useful for screening medication management capacity, in addition to a careful drug history, inspection of all medicines used (including over-the-counter drugs) and proxy information. However, no feasible screening methods have as yet found their way into clinical practice. Patients with impaired executive function, lack of awareness of illness and personality traits such as independency and high self-confidence may be at particular risk of non-adherence. The question is when to switch patient medication self-management to another person's responsibility if cognitive decline progresses. Further research is needed on measures to differentiate cognitive function and the relationships between memory concerns, memory strategy use and medication management. Also, studies evaluating the influence of personal support, health status and depression on the memory strategies used are needed. It is important to assess patients' attitudes toward medication and their relationship with proxies. Strategies for facilitating medication adherence in patients with dementia include prescribing as few medicines as possible, tailoring dose regimens to personal habits, and coordinating all drug dosing schedules as much as possible. When providing medication organizers, it is important to observe the patient's ability to use devices appropriately. In addition, automated computer-based reminding aids, online medication monitoring and telemonitoring may be helpful for patients with mild dementia. The decision as to when assistance with medication self-management is needed has to be made taking into account patient independency and safety aspects. This holds true for medicines with a narrow therapeutic range, in particular. Interactions among the individual patient's cognitive status, mood, level of self-efficacy and particular living situation must also be taken into consideration when searching for the optimal medication adherence strategy. No evidence-based recommendations can be given as yet. However, comprehensive assessment of the individual patient and careful consideration of all potential drug-related problems will probably help facilitate adherence and prevent compromised health outcomes in patients with dementia.

Repeated measures analyses of dose timing of antiretroviral medication and its relationship to HIV virologic outcomes

Medication adherence is a critical predictor of the effectiveness of antiretroviral medications in the treatment of HIV/AIDS. Studies of adherence, however, have focused primarily on the per cent of prescribed doses taken (per cent adherence). In the Adherence and Efficacy of Protease Inhibitor Therapy study, we collected detailed adherence data including dose timing information as well as data regarding patients' virologic responses. For 48 weeks, adherence data and virologic outcomes were collected every 4 weeks, and demographics and other measures were collected at baseline and at weeks 8, 24, and 48. We constructed eight different dose timing error (DTE) measures and evaluated their associations with virologic outcomes using longitudinal analyses. Repeated measures mixed effect models were fitted to evaluate the predicting power of each of the DTE measures. Among 52 036 electronically measured doses obtained from 122 patients, DTE measures significantly predicted virologic outcomes. Of the eight different DTE measures, the six DTE measures were significantly predictive of virologic outcomes even after controlling for per cent adherence. In conclusion, we identified several measures of DTE that explain HIV virologic outcomes not captured by traditional adherence measures. Investigations of adherence to antiretrovirals would benefit from measuring not only per cent adherence but dose timing adherence.

Association between adherence measurements of metoprolol and health care utilization in older patients with heart failure

OBJECTIVE

Data from electronic dosing monitors and published pharmacokinetic parameters were used to derive medication adherence measures for immediate-release metoprolol and examine their association with health care utilization of outpatients aged 50 years or older with heart failure.

METHODS

We used a 1-compartment model and published population pharmacokinetic parameters to estimate mean plasma metoprolol concentrations for patients treated for 6 to 12 months. In the absence of directly measured plasma concentrations, we calculated the intended mean plasma concentration (Cp'(ave)) under the assumption of perfect adherence to the prescribed dose and frequency of administration. Projected mean plasma concentrations (Cp(ave)) were estimated by use of data from recorded dosing times. In addition to taking adherence (percentage of dose taken) and scheduling adherence (percentage of doses taken on schedule), we calculated the deviation from the intended exposure (DeltaCp(ave) = Cp'(ave) - Cp(ave)) and the proportion of intended exposure achieved by the patient (Cp(ave) /Cp'(ave)). We assessed the association between the adherence measures and the numbers of emergency department visits and hospital admissions experienced by the patients.

RESULTS

Patients (N = 80) were aged 62 +/- 8 years. Mean DeltaCp(ave) and Cp(ave)/Cp'(ave) were 7.9 ng/mL (SD, 10.7) and 0.6 (SD, 0.3), respectively. Log-linear models adjusted for patient functional status indicated that greater deviation from the intended metoprolol exposure (DeltaCp(ave)) was associated with increased numbers of emergency department visits ( P < .0001) and hospital admissions (P < .0001). A higher proportion of intended exposure (Cp(ave) /Cp'(ave)) corresponded to a reduced number of emergency department visits (P = .0204) and hospital admissions (P = .0093). Taking adherence was univariately associated with both emergency department visits and hospital visits (P < .0001 and P = .0010, respectively). Scheduling adherence was associated with the number of emergency department visits (P = .0181) but not with the number of hospital admissions (P = .1602). Model selection procedures consistently chose the proposed measures over taking adherence and scheduling adherence.

CONCLUSION

Deviation from the intended exposure and proportion of intended exposure achieved by the patient are valid adherence measures for immediate-release metoprolol and are associated with health care utilization. The potential utility of these measures for other beta-adrenergic antagonists and perhaps other cardiovascular drugs should be investigated.

Some economic consequences of noncompliance

The twin problems of poor compliance and poor persistence with prescribed antihypertensive drug regimens appear to be responsible for much of the huge shortfall in the proportion of hypertensives whose treatment brings their blood pressure down to satisfactory levels. A further problem is the confounding of nonresponse and poor compliance in patients with "drug-resistant hypertension," in that about half of such patients are poor compliers, whose response to simple regimens usually proves satisfactory once their compliance with prescribed regimens is corrected. Electronic means for compiling ambulatory patients' drug dosing histories have now made it both technically and economically feasible to distinguish clearly between noncompliers and nonresponders, which clinical judgment cannot do because it is no better at making this crucial distinction than a coin toss. With the advent of reliable, economical measurements of patient compliance with prescribed drug dosing regimens, we can probably eliminate most of the compliance problems. The problem awaiting us after that is poor persistence with prescribed regimens for antihypertensive and other cardiovascular medicines that are meant for long-term or life-long use. A recent study has shown that median persistence with fully reimbursed drugs of the statin class is only 6 months, which is about one fortieth of the length it should be to realize full benefits of such therapy.

A systematic review of the associations between dose regimens and medication compliance

BACKGROUND

Previous reviews of the literature on medication compliance have confirmed the inverse relationship between number of daily doses and rate of compliance. However, compliance in most of these studies was based on patient self-report, blood-level monitoring, prescription refills, or pill count data, none of which are as accurate as electronic monitoring (EM).

OBJECTIVE

In this paper, we review studies in which compliance was measured with an EM device to determine the associations between dose frequency and medication compliance.

METHODS

Articles included in this review were identified through literature searches of MEDLINE, PsychInfo, HealthStar, Health & Psychosocial Instruments, and the Cochrane Library using the search terms patient compliance, patient adherence, electronic monitoring, and MEMS (medication event monitoring systems). The review was limited to studies reporting compliance measured by EM devices, the most accurate compliance assessment method to date. Because EM was introduced only in 1986, the literature search was restricted to the years 1986 to 2000. In the identified studies, data were pooled to calculate mean compliance with once-daily, twice-daily, 3-times-daily, and 4-times-daily dosing regimens. Because of heterogeneity in definitions of compliance, 2 major categories of compliance rates were defined: dose-taking (taking the prescribed number of pills each day) and dose-timing (taking pills within the prescribed time frame).

RESULTS

A total of 76 studies were identified. Mean dose-taking compliance was 71% +/- 17% (range, 34%-97%) and declined as the number of daily doses increased: 1 dose = 79% +/- 14%, 2 doses = 69% +/- 15%, 3 doses = 65% +/- 16%, 4 doses = 51% +/- 20% (P < 0.001 among dose schedules). Compliance was significantly higher for once-daily versus 3-times-daily (P = 0.008), once-daily versus 4-times-daily (P < 0.001), and twice-daily versus 4-times-daily regimens (P = 0.001); however, there were no significant differences in compliance between once-daily and twice-daily regimens or between twice-daily and 3-times-daily regimens. In the subset of 14 studies that reported dose-timing results, mean dose-timing compliance was 59% +/- 24%; more frequent dosing was associated with lower compliance rates.

CONCLUSIONS

A review of studies that measured compliance using EM confirmed that the prescribed number of doses per day is inversely related to compliance. Simpler, less frequent dosing regimens resulted in better compliance across a variety of therapeutic classes.

Electronic monitoring of compliance to lipid-lowering therapy in clinical practice

Nonadherence to treatment is a common problem in the clinical management of hypercholesterolemic patients. This study was carried out with the aim of monitoring the daily compliance to a 6-month course of lipid-lowering therapy, using a microelectronic device, the Medication Event Monitoring System (MEMS), versus pill count. Forty men with primary hypercholesterolemia were prescribed fluvastatin 1 x 40 mg daily, provided in a MEMS package to record the date and time of each opening of the pillbox. Thirty-nine of 40 patients (98%) completed the study. Total cholesterol and LDL cholesterol levels decreased significantly (18% and 25%, p < 0.001) during the 6-month therapy period. A high mean rate of compliance was achieved by MEMS using the following three indexes--compliance to total prescribed dose (88.8% +/- 13.5%), compliance to prescribed days (82.4% +/- 19.5%), and compliance to prescribed time of day (81.86% +/- 19.5%)--and by pill count (93.4% +/- 9.5%). In addition, the MEMS provided some patterns of nonadherence to medication, undetectable by pill count alone, such as a drug holiday in 38% of cases, a drug omission for more than 7 consecutive days in 9% of cases, and, conversely, use of more than the one prescribed daily dose in 47% of cases. A significant correlation between the rate of compliance and the decrease in LDL cholesterol was observed only when the compliance was assessed by MEMS. The results indicate that MEMS is a useful tool for monitoring compliance in clinical practice and may possibly increase adherence to long-term lipid-lowering therapy.

Measurement of patient compliance

The compliance of 91 diabetic patients using oral antidiabetics was studied. Patient compliance was measured using four different methods. Patients received their medication in a Medication Event Monitoring System (MEMS)-container. Each time the patient went back to the pharmacy for refill prescriptions, the number of tablets left in the container were counted. Pharmacy records were used to study the number of days of delay in getting the next refill. At the end of the study, a questionnaire was sent to every patient. Using MEMS as a standard, the results show that pill count and refill data overestimate the compliance of this group of patients. The MEMS data also show that the compliance data using only the number of tablets may be biased, because of possible overconsumption. Pill count does not show a correlation with compliance as measured by MEMS. The relation between compliance as measured with MEMS and refill compliance is weak.

Medication compliance as a feature in drug development

Well-designed clinical trials maximize the information that can be obtained regarding the clinical pharmacology of a drug and, in turn, can streamline and enhance the drug development process. Until recently, little emphasis has been placed on integrating the role of variability in individual patterns of drug-taking into the drug development process. With the use of electronic monitoring, the temporal relationship between an individual's pattern of dosing and the prescribed regimen may be examined, and individual drug exposure may be estimated based on the actual history of dosing. As a result, accurate estimation of exposure-response relationships (or surrogate markers of response) can be obtained. Considerations in the design of clinical trials must therefore be expanded to include appropriate methods to measure compliance, sufficient frequency of monitoring to allow the time course of response to be mapped, and the use of statistically valid methods of data analysis.

Role of patient compliance in clinical pharmacokinetics. A review of recent research

Until 1986 to 1987, the estimation of patient compliance with prescribed drug regimens in ambulatory care relied on methods that were biased either by their subjectivity or by the improvement in compliance that commonly occurs during the day or two prior to a scheduled examination, so called 'white-coat compliance'. In 1986 to 1987, 2 objective methods were developed: electronic monitoring and low-dose, slow-turnover chemical markers (digoxin or phenobarbital [phenobarbitone]) incorporated into dosage forms. While neither method is without limitations, both have enabled major advances in the understanding of patients' compliance with dosage regimens and, thus, the spectrum of drug exposure in ambulatory care. The new methods have also triggered not only a revival of interest in patient compliance and its determinants, but also new statistical approaches to interpreting the clinical correlates of widely variable drug administration, and thus drug exposure, in drug trials. The marker methods prove dose ingestion during the 3 to 7 days prior to blood sampling, but do not reveal the timing of doses. The electronic monitoring methods, i.e. time and date-stamping microcircuitry incorporated into drug packages, provide a continuous record of timing of presumptive doses throughout periods of many months, but do not prove dose ingestion. The electronic record has been judged robust enough to detect certain types of investigator fraud, and to support modelling projections of the complete time course of the plasma drug concentration during a trial. Both marker and electronic methods show that the predominant errors are those of omission, i.e. delays or omissions of scheduled doses. Patient interviews, diaries, and counts of returned, untaken doses have been shown by both marker and electronic monitoring methods to consistently and substantially to overestimate compliance. Monitoring of plasma drug concentrations also overestimates compliance, because white-coat compliance is prevalent, and the pharmacokinetic turnover of most drugs is rapid enough that measured concentrations of drug in plasma reflect only drug administration during the period of white-coat compliance. Thus, compliance is a great deal poorer in clinical trials than has been revealed by the older methods. The long-standing underestimation of poor compliance in drug trials has many implications for the interpretation of drug trials, for optimal dose estimation, for the interpretation of failed drug therapy, and for accurate labelling of prescription drugs.