Introduction

Coronary heart disease and stroke are the leading causes of death and disability worldwide. [1] Combination drug therapy with blood pressure and lipid-lowering treatment is effective in preventing cardiovascular disease events, reducing risk by about 75% with complete adherence to treatment [2],[3].

Failure to take medication is common and reduces its potential benefit, both through stopping medication completely and taking it less frequently than prescribed. It has been estimated that about one third of patients do not adhere to blood pressure or lipid-lowering treatment prescribed following a myocardial infarction or stroke (secondary prevention) and about half do not adhere to preventive treatment prescribed to prevent a first cardiovascular disease event (primary prevention) [4], [5].

Over 90% of the UK and US population own a mobile phone [6], [7], and text messaging is increasingly used in medical practice to remind patients of clinic appointments, to report test results and to adjust treatment dose [8]–[11]. Text messaging may be a useful way of reminding patients to take their medication, identifying who has not done so and the cause for not doing so if there is a reason and in such cases providing appropriate advice.

Methods

The protocol for this trial and CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S1.

Results

Figure 1 shows a flow diagram of the trial. Of the 303 patients enrolled, 2 were lost to follow-up providing 301 patients for analysis. The characteristics of the patients at entry to the trial were similar in the two groups (Table 1), as was the use of lipid-lowering drugs, blood pressure lowering drugs or both. The baseline characteristics of the two randomised groups were similar and no differences were statistically different.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Trial Flow Diagram.

https://doi.org/10.1371/journal.pone.0114268.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of the trial participants.

https://doi.org/10.1371/journal.pone.0114268.t001

Six months after randomisation, 14 out of 150 patients (9%) assigned to the Text group and 38 out of 151 patients (25%) assigned to the No text group stopped their medication or continued to take <80% of the prescribed regimen (took medication on fewer than 22 of the last 28 days of follow-up) – a 16 percentage point improvement in medication adherence (95% CI 7%–24%), p<0.001 (Table 2). The differences between the two trial groups, remained statistically significant when the proportion of patients who stopped medication completely and continued to take <80% of the prescribed regimen were considered separately (9% (3%–14%), p = 0.002 and 7% (0.2%–14%), p = 0.045, respectively. There were no statistically significant differences in blood pressure (132 mmHg versus 137 mmHg systolic, 77 mmHg versus 81 mmHg diastolic) or serum cholesterol (4.20 mmol/L versus 4.21 mmol/L total cholesterol and 2.29 mmol/L versus 2.22 mmol/L LDL cholesterol) between the Text and No text groups respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Use of medication at 6 months according to trial group.

https://doi.org/10.1371/journal.pone.0114268.t002

In the Text group, 98/150 (65%) of patients were reminded to take their medication on at least one occasion and 23/150 (15%) did not take their medication on at least one occasion, because of uncertainty over the need for treatment, concern over side-effects or another medical illness that led to medication being discontinued. Following a telephone consultation 20 or these patients (13%) resumed their medication and 3 stopped completely.

The results were not materially affected by age, sex, smoking, presence or absence of diabetes and history or no history of cardiovascular disease; no statistically significant interactions were found, but the power to detect differential effects was limited.

Discussion

The results of this trial show that in patients prescribed blood pressure and/or lipid-lowering medication for the prevention of cardiovascular disease, text messages improved the use of prescribed medication at 6 months compared with the use in patients who were not sent text messages. Stopping medication completely and taking less than 80% of the prescribed regimen were both reduced, by 9% and 7% respectively – an overall improvement in medication adherence of 16%.

The 80% cut-off (>22 of the last 28 days medication was taken) that classified a person as adhering to their medication was arbitrary, but has been used in previous publications. [4], [11] Had we adopted a lower or higher cut-off, say 75% or 85%, the difference in adherence between the text and no text groups would have been 15% or 20% respectively. The estimate is therefore reasonably robust.

The software used in this trial automated the sending of texts and recording of replies in a way that reminded patients to take their medication (about two-thirds of patients) and identified those who did not (1 in 6 patients), so the reason for not doing so could be determined and advice provided. This approach differs from previous trials which used one-way text messages to encourage the use of sunscreen [12], malaria prophylaxis [13], antibiotics [14] and blood pressure drugs [15] and did not show statistically significant improvements in adherence. Two randomised trials that, like ours, sent a text message and required a response, did show statistically significant improvements in adherence in the use of anti-retroviral treatment among HIV patients, in one trial [16], and vitamin C in healthy people in another [17]. Bidirectional texting may be an important factor in achieving a benefit.

The text messages were customized to the time of day that patients took their medication but the frequency that texts were sent was not automatically adjusted in the light of the response. Future versions of the software could be made more “intelligent” to enhance the customisation. The method was simple, is scalable and could be extended to other disorders for which drugs are taken long-term, such as asthma, epilepsy, HIV infection and multiple sclerosis.

In this trial, medication use was high because the main method of recruitment involved responding to a text invitation (response rate of 9%). The patient sample was likely to be a particularly conscientious group that was likely to adhere to their medication more than patients in general. This limited the trial's power to show statistically significant differences in cholesterol and blood pressure between the randomised groups. Such selection is probably inevitable in trials such as this, which require the consent and cooperation of the participants involved. As a consequence the results of this trial may underestimate the routine use of texting as a means of improving adherence to medication.

Several questions remain. For example, would the benefit of text messaging be improved or diminished in patients who were starting drug treatment, compared to patients who had been taking medication for some years already, as we studied here? Would the text intervention provide similar benefits in patients taking medication for other chronic disorders, such as tuberculosis or HIV infection? Would the improvement in medication adherence, observed at 6 months, be sustained if the text intervention stopped? Further research would be needed to answer these questions.

In conclusion, in this randomised trial, we found that among patients taking blood pressure and/or lipid-lowering treatment for the prevention of cardiovascular disease events, text messaging, improved the extent to which patients took their prescribed medication.

Supporting Information

Acknowledgments

We are grateful to Shahena Butt, Fatima Bangash and Julia Kowalczyk for assisting with the recruitment and follow-up of patients, to Mo Joomun for database support, to Sotiris Antoniou for helping to secure the initial funding from Astra Zeneca and to the participating General Practitioners and staff at their recruiting Health Centres; Dr Nick Brewer at Lower Clapton Health Centre and Sorsby Practices, Dr Ajay Goel at Lea Practice, Dr Salma at Jubillee Street Practice, Dr Loh at Neaman Practice, Dr Jonathan Gore at Lawson Practice and Dr Ellsbury at Paxton Green.