Does personalised text messaging influence patients' caries risk?

Abstract

Design A single-blinded, randomised controlled trial. The experimental group received 24 personalised text messages each week, disseminating tailored preventive advice using the multifactorial model for individual caries risk assessment: Cariogram. The same frequency of text messaging was delivered to the control group; however, these were non-personalised messages and did not factor the Cariogram.

Sample selection One hundred and ninety-one participants were assessed for eligibility by eight calibrated, volunteer dental practitioners in County Cork, Ireland. Six different inclusion criteria were detailed, including the requirement for medical card holders, serving as an indicator for economically underprivileged status. Other criteria included: aged between 19-70 years; competent with text messaging services; have a minimum of 20 teeth present; not pregnant; and prepared to give consent. Failure to return a baseline food diary or possess a mobile phone saw the exclusion of 20 participants. Following stratified and blocked randomisation, 85 and 86 participants were allocated in the test and control group, respectively. The study received ethical approval by the Clinical Research Ethics Committee of the Cork Teaching Hospitals.

Data analysis One hundred and eleven participants attended the follow-up examination, 26 weeks after randomisation, where the 'chance of avoiding new cavities' was determined as a numerical index for caries risk. The secondary aim was to measure individual changes to seven Cariogram risk factors between the baseline assessment and the re-examination. The ANCOVA intention-to-threat (ITT) protocol and the per-protocol method were adhered to for analyses of outcome measures. Statistical analysis was performed using SAS 9.4, in adherence to a pre-defined significance level of 5% (two-sided).

Results Both analytic techniques confirmed no statistically significant difference (p >0.05) between the groups regarding the 'chance of avoiding new cavities'. Of the risk parameters assessed, only saliva secretion demonstrated a positive effect in the intervention group (p = 0.036, OR = 0.3, 95% CI = 0.1, 0.9). Predictive modelling techniques were not reliable due to the limited sample size of per-protocol analysis.

Conclusions The failure to conclude statistical significance between the groups validates the null hypothesis. Accordingly, no difference can be established between the personalised nor non-personalised mobile text messaging intervention on the caries risk of underprivileged adults in Ireland. The authors address the value of further studies exploring the potential for caries risk reduction through mobile phone communications.