Background: Medication organisation devices (MODs) provide compartments for a patient’s medication to be organised into the days of the week and the recommended times the medication should be taken. Aim: To define the optimal trial design for testing the clinical effectiveness and cost-effectiveness of MODs. Design: The feasibility study comprised a systematic review and focus groups to inform a randomised controlled trial (RCT) design. The resulting features were tested on a small scale, using a 2 × 2 factorial design to compare MODs with usual packaging and to compare weekly with monthly supply. The study design was then evaluated. Setting: Potential participants were identified by medical practices. Participants: Aged over 75 years, prescribed at least three solid oral dosage form medications, unintentionally non-adherent and self-medicating. Participants were excluded if deemed by their health-care team to be unsuitable. Interventions: One of three MODs widely used in routine clinical practice supplied either weekly or monthly. Objectives: To identify the most effective method of participant recruitment, to estimate the prevalence of intentional and unintentional non-adherence in an older population, to provide a point estimate of the effect size of MODs relative to usual care and to determine the feasibility and acceptability of trial participation. Methods: The systematic review included MOD studies of any design reporting medication adherence, health and social outcomes, resource utilisation or dispensing or administration errors. Focus groups with patients, carers and health-care professionals supplemented the systematic review to inform the RCT design. The resulting design was implemented and then evaluated through questionnaires and group discussions with participants and health-care professionals involved in trial delivery. Results: Studies on MODs are largely of poor quality. The relationship between adherence and health outcomes is unclear. Of the limited studies reporting health outcomes, some reported a positive relationship while some reported increased hospitalisations associated with MODs. The pre-trial focus groups endorsed the planned study design, but suggested a minimum recruitment age of 50–60 years. A total of 35.4% of patients completing the baseline questionnaire were excluded because they already used a MOD. Active recruitment yielded a higher consent rate, but passive recruitment was more cost-effective. The prevalence of intentional non-adherence was 24.7% [n = 71, 95% confidence interval (CI) 19.7% to 29.6%] of participants. Of the remaining 76 participants, 46.1% (95% CI 34.8% to 57.3%) were unintentionally non-adherent. There was no indication of a difference in adherence between the study arms. Participants reported a high level of satisfaction with the design. Five adverse/serious adverse events were identified in the MOD study arms and none was identified in the control arms. There was no discernible difference in health economic outcomes between the four study arms; the mean intervention cost was £20 per month greater for MOD monthly relative to usual supply monthly. Conclusions: MOD provision to unintentionally non-adherent older people may cause medication-related adverse events. The primary outcome for a definitive MOD trial should be health outcomes. Such a trial should recruit patients by postal invitation and recruit younger patients. Future work: A study examining the association between MOD initiation and adverse effects is necessary and a strategy to safely introduce MODs should be explored. A definitive study testing the clinical effectiveness and cost-effectiveness of MODs is also required.