Study design

TEXTMEDS is a single-blind, multicentre, randomised controlled trial delivering a 12-month secondary prevention programme via text messages to patients with ACS (figure 1). The recruitment sites include 15 public hospitals in urban and rural Australia that serve ethnically, culturally and socioeconomically diverse communities. The target sample size is 1400 patients with a confirmed diagnosis of ACS. Participants are randomly allocated to either control or intervention group. The control group is assigned to receive standard care as determined by their treating physician. The intervention group is allocated to receive, in addition to standard care, multiple weekly text messages that provide information, motivation and support for medication adherence, smoking cessation (if relevant), recommendations for healthy diet and exercise and other cardiac prevention advice. Trained research personnel blinded to the group allocation conduct assessments at baseline and 12 months during a face-to-face interview and at 6 months over phone. Further data will be obtained via extraction of linked data available with consent from government repositories. The study will be augmented by formal economic and process evaluations to assess acceptability, utility and cost-effectiveness of the intervention. This study will be monitored and managed centrally with periodic site monitoring visits.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow chart of study design. Patients with acute coronary syndromes are considered for inclusion in the study and randomisation is single blind, with allocation determined by computerised programme and initiated by computerised programme after discharge from hospital. ACS, acute coronary syndrome; CV, cardiovascular; LDL, low-density lipoprotein; TEXTMEDS, TEXT messages to improve MEDication adherence and Secondary prevention.

Randomisation and blinding

After obtaining written informed consent, the data are entered by the site coordinator into a web-based case record form (CRF) and then randomisation occurs via a centralised, computerised randomisation programme in a uniform 1:1 allocation ratio. The text message programme is initiated after the patient is discharged from the hospital and the patients (but not their care providers, research personnel or investigators) are informed of their allocation through a single text message. Study coordinators and research assistants conducting the assessments and statisticians will be blinded. This randomisation programme is electronically linked to the software programme that delivers the text message intervention, thereby minimising need for human interference. Key participant characteristics that determine text message customisation and personalisation are also automatically imported into software administering the intervention.

Study population

Patients with a confirmed diagnosis of ACS are eligible for the study. ACS is defined based on one of the two following criteria:

1. Acute myocardial infarction consistent with the Third Universal definition of myocardial ischaemia (MI)26: Detection of rise and/or fall of cardiac troponin with at least one value above the 99th percentile of the upper reference limit together with evidence of MI with at least one of the following: (1) symptoms of ischaemia; (2) ECG changes indicative of new ischaemia; (3) development of pathological Q waves on ECG or (4) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality.

2. Patients admitted to hospital acutely with symptoms typical of cardiac ischaemia not meeting the above criteria and one of the following indicators of coronary artery disease: (1) new or known coronary angiographic evidence of coronary disease (coronary artery narrowing of ≥50% of at least one major vessel) or (2) prior coronary artery bypass graft surgery or (3) prior percutaneous coronary intervention.

Patients must also meet the following inclusion criteria: (1) ability to provide informed consent, (2) own an operational mobile telephone and (3) can read text messages in English on a mobile phone and (4) life expectancy more than 6 months. For those people who are ineligible or decline to participate, a ‘screening log’ of basic demographic information and reasons for non-participation is maintained.

Control group

The control group will receive standard care post-ACS as determined by their usual doctors. This generally includes secondary prevention cardiovascular (CV) medications and referral to a centre-based cardiac rehabilitation or other secondary prevention programme.

Intervention

The intervention group will receive standard care (as described above), plus an additional support package that includes a series of text messages focusing on medication adherence and lifestyle modification as well as an opportunity to communicate with an English-speaking health counsellor over a 12-month period as detailed below.

Data collection and study outcomes

The follow-up study assessments will occur at 6 months over phone and 12 months in-person. The primary outcome of medication adherence to cardioprotective messages will be the difference between intervention and control groups in the percentage of patients adherent to all five classes of recommended cardioprotective medications (unless contraindicated/previously documented intolerance). The five cardioprotective classes of medications are: (1) angiotensin-converting enzyme inhibitor or angiotensin II receptor blockers, (2) beta-blocker, (3) statin lipid-lowering, (4) aspirin and (5) ADP receptor antagonist. Medication adherence will be measured by asking participants on how many days in the past 30 days they missed a particular class of medication for each of the five classes of medications indicated and this is asked at both 6-month and 12-month visits. Patients will be defined as adherent if they report that they have taken their indicated medications on >80% of days (>24/30 days) at both 6-month and 12-month visits.19 Indicated medications will be all five classes of guideline recommended medications listed above unless a contraindication is documented in their case report forms.

The key secondary outcomes are the difference between intervention and control groups at 12 months for LDL-cholesterol level and systolic BP (raised BP). Other secondary outcomes include: total cholesterol, diastolic BP, the proportion of participants adherent to each medication class, other CV risk factors including behavioural risk factors, psychosocial factors, major CV events, hospital readmissions and all-cause mortality and composites of risk factor control and major adverse CV events (table 1). A formal endpoint adjudication process will be implemented and includes a review of original documents for major CV events, hospital readmissions and deaths by two independent adjudicators. The adjudication process is overseen by KS. Serious adverse events will be collected by the investigator/research coordinator from randomisation to 30 days after the end-of-study visit and reported to the central coordinating centre as well as to ethics as per local guidelines. To optimise follow-up, multiple attempts will be made to contact participants including contacting their referring doctor and at a minimum we will aim to record vital status for all participants.

Data management is via a secure clinical trial data management system managed by The George Institute (TGI) with data entry through a password protected, web-based interface by registered staff, real-time data query generation for values entered outside of preset valid ranges and consistency checking. The steering committee (chaired by CC) has the overall responsibility for execution of the study.

Process measures

Process evaluations explore the implementation, receipt and setting of an intervention and assist in the interpretation of outcome results.32 Analyses will be informed by the Pawson and Tilley realistic evaluation model,33 which is an approach grounded in realism that considers social, individual, contextual and material contributions to change. We will use mixed methods to investigate why the TEXTMEDS strategy may or may not have been effective and which intervention components were most influential.

The process evaluation will include a semiquantitative survey and focus groups with intervention participants. All participants receiving the intervention will be invited to complete a written questionnaire assessing their perceptions, experience and the degree of engagement with the intervention at the completion of the trial. The questionnaire will comprise a combination of closed 5-point Likert scale and open questions. A sample of the intervention group participants will be invited to participate in focus groups on perceptions of the utility and acceptability of the intervention programme. To obtain a broad range of views we will use a maximum variation sampling method based on patient characteristics such as gender and site. Sampling will continue until thematic saturation is reached; however, we anticipate approximately four focus groups will be needed. An in-depth interview will also be conducted with the health counsellor to understand his/her perceptions of patients’ engagement with the programme. Focus groups and the health counsellor interview will be conducted by a trained qualitative interviewer, digitally recorded and transcribed. Analyses will be thematic and coding based on emergent themes.

Economic analysis

The costs and health outcomes associated with the intervention will be compared in an incremental cost-effectiveness analysis. A health sector perspective will be adopted. The costs of delivering the intervention will be assessed by measuring and valuing the incremental resources used.34 Hospitalisations will be collected at each follow-up. Consent will be sought for access to individual participant Medical Benefits Schedule and Pharmaceutical Benefits Scheme claims usage through Medicare Australia to ascertain use of medical services and prescribed pharmaceuticals. Changes to health benefits will be assessed using quality-adjusted life years (QALYs). These will be assessed using the 12-item Short Form Survey Version 2 (SF-12_V2) Health Survey35 (License Number QM024287). Although we do not expect significant differences in survival or quality of life between treatment groups within trial, these data are needed to provide an estimate of the baseline quality of life in this patient population. Given the likely small numbers of CV events occurring within the trial, the quality of life and costs data collected in trial will need to be modelled using data on quality of life and cost associated with CV events from literature review.36 Long-term costs and QALYs will be modelled using a decision analytic Markov model that will enable us to extrapolate beyond the data collection period, predict within the cohort the occurrence of CV events over a lifetime and thus estimate longer-term costs, potential cost savings and benefits of the intervention.37

Statistical considerations

A 10% improvement in the proportion taking appropriate secondary prevention is considered to be a clinically meaningful increase and associated with a decrease in short-term mortality.38 To test for a 10% improvement in the primary outcome relative risk (RR of 1.10) in the treatment compared with the control arm (assuming ~70% of patients are taking appropriate secondary prevention in control arm, based on contemporary Australian registry data,39 40 and this would rise to 77% in the intervention arm) with 80% power (type I error=5%, two-sided test), we would require a total sample size of 1246, rising to ~1396 accounting for a ~10% loss to follow-up. This sample size of 1246 would enable us to also detect with 80% power a minimum detectable difference in secondary outcomes of (1) 0.15 mmol/L in LDL-cholesterol assuming a SD of 0.92 mmol/L,41 (2) 2.7 mm Hg in systolic BP, assuming an SD of 17 mm Hg41 and a minimum detectable RR of 0.69 in CV events assuming a major CV event rate of 19% or a power of approximately 13% to detect a RR of 0.80 in CV events, assuming a lower CV event rate of 5% at 12 months from contemporary Australian data.

The intention to treat principle will be used and patients will be analysed according to the group to which they are allocated. A detailed statistical analysis plan will be prepared prior to the analysis. The primary outcome will be compared between groups using the χ² test. The primary analysis will be unadjusted. The mean level of BP and cholesterol will be analysed using an analysis of covariance using the baseline values as the covariate with estimation of the mean difference and corresponding 95% CI. For each of the secondary outcome measures, parameters will be compared between the two groups with relevant statistical tests providing mean difference for continuous variables or RRs for dichotomous variables along with the 95% CIs and two-sided P values.

We will also analyse effects across subgroups, such as sex and age, using logistic regression models with treatment group, subgroup and subgroup by treatment interaction as fixed effects. Number of patients with major CV events will be compared between treatments using a χ² test. The criterion for statistical significance will be set at α=0.05. Number needed to treat will also be calculated and its 95% CI for the primary outcome.

Ethics and dissemination

The findings of this study will be disseminated via the usual scientific forums including peer-reviewed publications and presentations at international conferences. The study is sponsored by The George Institute for Global Health at the University of Sydney, and centrally coordinated and managed by staff based mainly at Westmead Hospital and the George Institute. The design and conduct of the study is overseen by a Steering Committee (authors and some site principle investigators). This committee has expertise in large-scale clinical trials and qualitative research, economic analysis, clinical CVD management and healthy policy implementation. This study will adhere to the Australian National Health and Medical Research Council ethical guidelines for human research. Formal ethical approval for this study has been obtained from Western Sydney Local Health District Human Research Ethics Committee (HREC2012/12/4.1 (3648) AU RED HREC/13/WMEAD/15). The current protocol is V.3.0 (31 October 2013). Any changes in the protocol during the trial that may affect the conduct of the trial, safety, and the benefit to patients will require a formal amendment to the protocol. Written and informed consent will be obtained from all participants.

Reporting and dissemination

Results of this study will be presented at national meetings and published in a scientific journal. Participants will not be individually notified regarding the results of this study.