To compare rates of medication adherence for once daily (qd) dosing, twice daily (b.i.d.) dosing and multiple daily dosing (MDD) regimens of antihypertensive drugs.

MEDLINE, EMBASE and International Pharmaceutical Abstracts were searched for studies published in English or French between 1980 and 1998; the keywords were stated. Published abstracts and posters from symposia or colloquia were excluded. The reference lists in identified studies, reviews and textbooks were handsearched.

Study designs of evaluations included in the review

Any type of comparative study was eligible for inclusion. Blinding was not a requirement for inclusion. Prospective and retrospective studies were included; further details of the study design were not reported.

Specific interventions included in the review

Studies that compared different dosing frequencies of a chronically administered (10 weeks or more) antihypertensive drug regimen were eligible for inclusion. The drugs had to be delivered in a solid oral formulation. The included studies used beta-blockers and various other drugs, such as diuretics, potassium supplements and calcium-channel blockers. The drugs were used as monotherapy or in combination with other drugs (some studies did not report this information). Where stated, the duration of the studies ranged from 10 weeks to 2 years.

Participants included in the review

Studies of adults (18 years or more) who were being treated for essential hypertension were eligible for inclusion. Studies that only included very old patients (older than 74 years) were excluded. The mean age of the participants ranged from 50 to 67 years.

Outcomes assessed in the review

Studies that reported adherence rates, which were measured using the same method for both treatment groups, were eligible for inclusion. The studies could use any method to assess adherence. The included studies measured adherence using at least one of the following: pill count, patient interview, electronic compliance monitor, prescription refill data, and medication event monitoring system. The studies used different definitions of adherence, including the percentage of patients taking 75% or more to 100% of doses. The primary outcome in the review was the proportion of patients who had taken 80% or more of the prescribed doses. Where studies did not report this outcome, the reviewers used the main adherence outcome in the study.

How were decisions on the relevance of primary studies made?

One reviewer screened the abstracts and titles and selected potential studies. All identifiers were then removed from potential articles. Two trained reviewers then selected studies for inclusion using a checklist, and resolved any disagreements through discussion with a third reviewer.

Validity was assessed and scored using criteria described by Haynes et al. These criteria included study design, selection and description of the sample, description of the condition, methods used to measure adherence, description of the intervention, and the definition of adherence. The maximum possible score was 17 points. Studies scoring 8.5 or higher were classified as acceptable. Two reviewers assessed validity and resolved any disagreements through discussion with a third reviewer.

Two reviewers extracted the data using a data collection form, and resolved any disagreements through discussion with a third reviewer. The data extracted included study design, drug class, type of therapy, study duration, population characteristics, method used to measure adherence, definition of adherence, and the adherence rate for each dosing regime. For each study, the difference in adherence rates was calculated for qd dosing compared with MDD, for qd compared with b.i.d. dosing, and for b.i.d. compared with more than b.i.d. dosing.

How were the studies combined?

The studies were combined in a meta-analysis. A random-effects model was used to pool differences in adherence rates for qd dosing compared with MDD, for qd compared with b.i.d. dosing, and for b.i.d. compared with more than b.i.d. dosing.

How were differences between studies investigated?

The following subgroups were defined a priori: method used to measure adherence; definition of adherence (80% or more versus 90% or more); study design (prospective versus retrospective); medication class (calcium-channel blockers only); and duration of treatment (3 to 6 months versus 12 to 24 months). Subgroup analyses were conducted to explore the effect of specified variables. Obvious outliers were identified by plotting adherence rates against each other in a L'Abbe plot and using regression analysis. Sensitivity analyses were conducted by reanalysing data after removing apparent outlying studies. Statistical heterogeneity was assessed using the chi-squared statistic. When significant heterogeneity was detected, the analyses were repeated after excluding the study that lay furthest from the trend line of the other studies. The relationship between the year of study and quality score, and between quality scores and difference in adherence rates, was explored using the correlation coefficient (Spearmann's rho).

Eight comparative studies (11,485 patients, doses or other measures) were included. There were six prospective and two retrospective studies.

No outliers were identified, although the Q test showed statistically significant results for some comparisons (the paper did not report which comparisons).

All of the included studies scored more than 9 (range: 9.5 to 16) out of a possible 17 on quality criteria. The correlation between the year of study and quality score (rho = -0.10, P=0.81) and between quality scores and difference in adherence rates (rho = 0.23, P=0.59) was small.

The studies differed in duration, their definition of adherence and the method used to measure adherence.

Once versus multiple daily dosing (7 studies, 4,669 observations): qd dosing significantly increased adherence rates compared with MDD regimens (91.4% versus 83.2%, P<0.001). Similar results were obtained in the subgroup analyses (the results were reported).

Once versus twice daily dosing (5 studies, 2,152 observations): qd dosing significantly increased adherence rates compared with b.i.d. dosing regimens (92.7% versus 87.1%, P=0.026).

Twice versus more than twice daily dosing (4 studies, 8926 observations): there was no statistically significant difference in adherence rates between b.i.d. and more than b.i.d. dosing regimens (90.8% versus 86.3%, P=0.069). When defining adherence as a 90% or greater intake, b.i.d. dosing significantly increased adherence rates compared with more than b.i.d. dosing (76.1% versus 67.0%, P<0.001).

The authors concluded that qd dosing regimens are associated with higher adherence rates than b.i.d. or MDD regimens.

The review question was clear in terms of the intervention, participants and outcomes. The inclusion criteria were broadly defined in terms of study design. Several relevant sources were searched and the search terms were stated. As the authors acknowledged, the inclusion of studies published in one of only two languages raises the possibility of publication and language bias. Two reviewers independently selected the studies, assessed validity and extracted the data, thus reducing the potential for bias and errors. Validity was assessed using established criteria and the quality scores were reported. However, there was no discussion of the methodological limitations of the included studies.

Some relevant information on the included studies was tabulated, but other than classifying studies as prospective or retrospective, no other details of the study design were given; this hampers an assessment of the quality of the evidence. The studies were combined in a meta-analysis and statistical heterogeneity was assessed. The unit used in the meta-analysis was patients, doses or other measures. If adjustments were not made for non-independent observations, the analysis might have overestimated the effectiveness of qd and b.i.d. regimens. It is questionable whether combining studies with different designs was appropriate. The influence of various factors on the results was explored and some of the limitations of the review were discussed. In view of the limitations highlighted, the conclusion should be interpreted with caution.

Practice: The authors did not state any implications for practice.

Research: The authors stated that research is required to determine the influence on adherence rates of the timing of and intervals between drug doses, and to determine the aspects of daily drug dosing that have the greatest effect on adherence.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.