Key points

  • The available evidence suggests that interventions delivered by mobile phones may be effective in improving adherence to oral hygiene advice.

  • Argues there is a need to design mobile phone interventions that are grounded in behaviour change theory to explore this concept further.

  • Suggests, given the rapid proliferation of apps and other online information targeted at patients, there is a need to assess quality and effectiveness of these resources and navigate patients towards the most appropriate ones.

Introduction

Dental caries is almost entirely preventable; however, globally it affects 60-90% of school-aged children.1 In 2013, a national survey (England, Wales and Northern Ireland) reported a 28% prevalence of dental caries among five-year-old children.2

The management of extensive decay in young children is often under general anaesthesia and dental caries is now the most common reason for admission to NHS hospitals in England for 5-9-year-olds.3 Repeat episodes of dental general anaesthetic are reported to be between 4.2% and 17.0%.4,5 Furthermore, general anaesthesia carries risks to health and is costly, often necessitating time off school for children and time away from work for their parents. Recent research has shown that children who have received a dental general anaesthetic are over 2.5 times more likely to be dentally anxious in their late teens than those who have not.6 There are then implications associated with this, in that dental anxiety often leads to avoidance of dental care and allows for dental disease to progress, causing irreversible damage.

The Royal College of Surgeons of England has identified dental caries and dental general anaesthesia as major health care challenges. The need for Public Health England to invest in programmes to improve children's oral health was also at the forefront of recommendations made by the college to the Chief Dental Officer.7 Given that diet and oral hygiene are key components in the aetiology of dental caries, approaches to effect a change in diet and oral hygiene-related behaviours are essential to address this global challenge. Traditionally, approaches to improve oral health behaviours have aimed to increase patient knowledge. However, at present, there is weak evidence that improvements in knowledge lead to improved oral health behaviour.8 Conversely, there is evidence supporting the use of interventions developed using psychological behaviour change models to improve oral health.8 Although many models of behaviour change exist, a contemporary and widely accepted framework is the Behaviour Change Wheel (BCW). Developed by Michie et al.,9 the BCW is a theoretical framework based on multiple models of behaviour change. The COM-B model forms the core of this and proposes that individuals require capability (C), opportunity (O) and motivation (M) to perform or adapt a particular behaviour (B). Available evidence shows that interventions based on behaviour change theory and those with more behaviour change techniques (BCTs) are more effective than those that are not based on theory and with fewer BCTs.10 Behaviour change techniques are defined as 'the smallest identifiable components that in themselves have the potential to change behaviour'.11 Ninety-three BCTs have been identified and categorised in the BCT taxonomy V1.12

Mobile phones may be invaluable tools in delivering interventions developed using behaviour change theory. This technology allows for several approaches to be utilised simultaneously in order to address an individual's capability, opportunity and motivation in a cost-effective manner. Mobile phones are readily available, with some sources reporting 100% penetration in Western Europe.13 Moreover, they are very versatile; for example, they can also be utilised to provide personalised treatment information, such as appointment and toothbrushing reminders, at times which are convenient to the patient. A scoping review of the literature revealed that a number of randomised controlled trials assessing the effectiveness of mobile phones in improving adherence to treatment advice had been reported. Notably, evidence is emerging to suggest that apps and mobile phone-based reminders are effective in improving oral health.14,15

The aim of this systematic review was therefore to assess the effectiveness of interventions delivered by mobile phones in improving adherence to oral hygiene advice for children and adolescents.

Objective

A systematic review of randomised controlled trials to determine the effectiveness of interventions delivered by mobile phones versus other interventions not using mobile phones in improving adherence to oral hygiene advice for children and/or adolescents.

Methods

The methodology for this systematic review including criteria for considering studies eligible for inclusion, the outcomes assessed, settings, information sources, data management, analysis and proposed synthesis was registered online on the PROSPERO database in November 2017: CRD42017078414.

Protocol changes

The registered protocol initially included 'children aged 10 to 17 years (inclusive)', however, an initial screening of the results highlighted that a number of studies included patients up to the age of 18 years. To maximise the potential studies for inclusion it was decided to amend the inclusion criteria to allow the inclusion of individuals aged 10 to 18 years (inclusive).

Results

The search of databases (up to 18 January 2018) retrieved 524 titles and abstracts; after removing duplicates, 516 were eligible for screening (Supplmentary online appendix). The titles and abstracts were screened independently by MOS and SJC and categorised as: 'include', 'exclude' or 'uncertain'. A weighted kappa score demonstrated the overall level of agreement to be 'good' (Κ = 0.664). There was 100% agreement for the records for 'inclusion', the full texts of these studies and those studies categorised as 'uncertain' were obtained for further assessment. After assessing nine full texts, two studies were eligible for inclusion and seven were excluded.

No additional studies were identified on the ClinicalTrials.gov or the World Health Organisation International Clinical Trials Registry Platform, the reference list screening of included studies, communication with experts in the field or communication with contact authors. Figure 1 presents a flow diagram for the review. The searches were updated on 18 December 2018, and no additional studies were identified.

Fig. 1
figure 1

Study flow diagram

Included studies

Two studies were included in this review.16,17 One study explored the use of text messages16 and the other explored the use of an 'app'.17 Both of these studies exclusively recruited orthodontic treatment patients. Table 1 presents characteristics of the included studies and summarises details of the design, methods, participants, interventions, comparisons and outcome measures.

Table 1 Characteristics of included studies

Characteristics of the trial settings and investigators

The Bowen et al. trial16 was conducted in the Seton Hill University Centre for Orthodontics, USA, but the providers of care were not stated. The corresponding author was contacted by email to obtain clarification, but no response has been received to date.

The setting and care providers were not stated in the Zotti et al. paper.17 However, communication with the contact author confirmed that the study was performed in a dental hospital setting with second- and third-year orthodontic postgraduates, supervised by clinical instructors, providing patient care.

Characteristics of trial participants

The total number of participants across the included studies was 130. One hundred and twenty participants completed all follow-up assessments. The mean age of participants in the Bowen et al.16 and Zotti et al.17 trials was 15.1 and 13.9 years, respectively. More females were recruited in each of the studies and both study samples comprised 58% females and 42% males.

There was some heterogeneity between the included trials; Bowen et al.16 included participants aged 10-18 years of age, whereas Zotti et al.17 included participants aged 12-17 years of age. Bowen et al.16 stated that participants were included if they had maxillary fixed appliances and had at least six months of treatment remaining, which suggests that participants were in active treatment before enrolment in the study. However, Zotti et al.17 recruited participants before commencing treatment.

Characteristics of interventions

The interventions and follow-up periods varied between the two included studies. Bowen et al.16 provided participants in the intervention group with automated text messages two to three times a week for four weeks and followed participants up for three months. Zotti et al.17 provided participants in the intervention group with access to smartphone-specific video tutorials and a chat room, as outlined in . Participants were followed up for 12 months.

In the Bowen et al.16 trial, all participants watched an audio-visual presentation on how to brush with a conventional toothbrush (using the Bass technique). In the Zotti et al.17 trial, all participants received standardised oral hygiene instructions along with toothpaste, toothbrush, mouthwash, interproximal brush, dental floss and plaque-disclosing tablets.

None of the interventions were reported to have been developed based on a specific theory of behaviour change.

Characteristics of outcome measures

Both studies reported plaque scores, however there was heterogeneity as the method of plaque assessment differed. Bowen et al.16 utilised planimetry which provides the percentage of plaque coverage on each tooth; whereas Zotti et al.17 utilised the plaque index, scoring zero to three for each surface, and subsequently calculated the overall mean. It was therefore not possible to combine the data in a meta-analysis. Zotti et al.17 also reported bleeding scores and caries. Neither of the included studies reported adverse events, cost-effectiveness or patient preferences.

Excluded studies

Seven studies were excluded and the reasons for exclusion are as follows:18,19,20,21,22,23,24

  • Patients were not the focus of the intervention

  • Mobile phones were not used to deliver the intervention

  • Patients over the age of 18 years were included, the authors were contacted to determine whether data were available for adolescents only but, to date, no response has been received

  • Inadequate follow-up period.

Ongoing studies

Two potentially relevant studies are currently ongoing and were identified by contact with experts in the field. However, no data are available as of yet. The protocol for one of these studies has been published.25 The results of this study may be appropriate for inclusion when they become available.

Risk of bias in included studies

The Cochrane risk of bias tool, Review Manager 5.3,26 was used to aid with presentation of the risk of bias. The assessment for each of the included studies is included in Table 2 and Table 3. The risk of bias graph and summary are presented in Figure 2 and Figure 3.

Table 2 Risk of bias assessment, Bowen et al.16
Table 3 Risk of bias assessment, Zotti et al.17
Fig. 2
figure 2

Risk of bias graph. Judgements about each risk of bias item presented as percentages across all included studies

Fig. 3
figure 3

Risk of bias summary. Judgements about each risk of bias item for each included study

Allocation

Sequence generation and allocation concealment

Random sequence generation and allocation concealment were assessed to be at unclear risk for Bowen et al.,16 as insufficient detail was present to make a clear judgement and it has not been possible to obtain further information. The Zotti et al.17 study was considered to be at low risk of bias, the authors reported using a stratified randomisation list produced by an external office which was contacted by the researchers to determine patient allocation.

Blinding

Blinding of participants was judged to be a low risk for Bowen et al.16 The authors reported that patients were not aware that messages were part of the study. The Zotti et al.17 study was deemed to be at unclear risk of bias. However, it is appreciated that, given the nature of the study, it was not possible to blind subjects.

Blinding of outcome assessment was considered to be an unclear risk for Bowen et al.,16 as the authors did not specify any measures taken to allow for this. The Zotti et al.17 study was deemed to be at low risk of bias in this domain, as the authors reported blinding.

Incomplete outcome data

This domain was judged as an unclear risk for the Bowen et al.16 study, as there were some inconsistencies regarding the flow of patients through this trial, as detailed in Table 1. There were no dropouts reported in the Zotti et al.17 study and therefore this was deemed to be at low risk of bias.

Selective reporting

Selective reporting was considered to be at unclear risk for both Bowen et al.16 and Zotti et al.17

Other sources of bias

Bias from other sources was deemed to be an unclear risk for Bowen et al.16 and as low risk for Zotti et al.17

Overall assessment of bias

All domains had to be assessed as being at low risk of bias for the study to be considered low risk of bias overall, both studies were therefore considered as being at unclear risk of bias overall.

The COM-B components and behaviour change techniques in included studies

In both studies, capability, opportunity and motivation were addressed to some degree and the BCTs used for this varied between studies. The results are summarised in Table 4, and some examples are provided to support the judgements made in the review.

Table 4 The COM-B component and behaviour change techniques addressed in the studies

Effects of interventions

Plaque scores

For both studies, plaque scores were statistically significantly lower in the intervention group when compared with the control group at the final follow-up. However, the final follow-up time point differed between studies. Bowen et al.16 followed patients up for a maximum of three months (T0: baseline, T1: one month and T2: three months) while Zotti et al.17 followed patients up for 12 months (T0: baseline, T1: three months, T2: six months, T3: nine months and T4: 12 months).

Bowen et al.16 reported significantly less plaque accumulation in the intervention group at one month and three months. Interestingly, Zotti et al.17 reported no statistically significant difference in plaque scores between the intervention and control groups at three months; the difference was evident only from six months onwards (p <0.01).

Gingival bleeding scores

Only the Zotti et al.17 study reported gingival bleeding scores. There was no significant difference at baseline or three months between the intervention and control groups. However, at six, nine and 12 months there was significantly less gingival bleeding in the intervention group (p <0.05 for all three time points).

Caries

Only the Zotti et al.17 study reported caries and there was no statistically significant difference in white spot lesions at baseline, three months or six months. However, at nine and 12 months, patients in the intervention group were significantly less likely to have white spots than the control group (p <0.05 for both time points).

Summary

The results of the studies were not pooled as the content and delivery of interventions was different. Both studies reported plaque scores, however the method of plaque assessment differed and therefore it was not possible to combine these scores in a meta-analysis.

Overall strength of evidence

The overall strength of evidence for the effectiveness of mobile phones in reducing plaque scores, as rated by GRADE,27 was considered to be moderate while the effectiveness of mobile phones in reducing bleeding scores was considered to be high. The results of the GRADE assessment are summarised in Table 5.

Table 5 GRADE assessment summary table

Discussion

The studies included in this review were exclusively aimed at supporting orthodontic patients; because excellent oral hygiene is a prerequisite for orthodontic treatment, the results obtained cannot necessarily be generalised to the dental population as a whole. Although the interventions in both studies were delivered via a mobile phone, the content varied, with one trial providing text messages16 and the other utilising video tutorials and a chatroom.17 Both studies were deemed to be at unclear risk of bias, overall.

The results indicated that there is some evidence to suggest that the use of mobile phones is effective in improving adherence to oral hygiene advice. These findings are consistent with recently published systematic reviews assessing the effectiveness of reminders (including the use of mobile phones to deliver these) in improving the oral hygiene of orthodontic patients.15,28

Overall completeness, quality and applicability of evidence

The overall strength of evidence for the effectiveness of mobile phones in reducing plaque scores, as rated by GRADE,27 was considered to be moderate. The effectiveness of mobile phones in reducing bleeding scores was considered to be high. However, the generalisability of this review is limited due to the inclusion of only two trials which were focused solely on orthodontic patients and their unclear risk of bias. Only one of the outcomes assessed was the same in both studies (plaque score) and the method of outcome assessment differed; meta-analysis, therefore, was not appropriate. In addition, the duration of follow-up differed in the two studies and neither study followed patients up for the whole duration of their orthodontic treatment.

Behaviour change techniques utilised

Neither of the interventions were reported to have been developed based on a specific theory of behaviour change. This highlights a significant area for future research, given that the available evidence suggests interventions based on behaviour change theory and those with more BCTs are more effective than those that are not based on theory and have fewer BCTs.10,27 Interestingly, from the 93 available, only six BCTs were utilised across the two trials to address psychological capability, physical and social opportunity, and automatic and reflective motivation to some degree.

Implications for practice and future research

Neither of the included studies reported utilising digital interventions that were designed using a 'ground up' approach with patient and professional engagement. However, given that there is now some evidence in support of digital interventions, the next stage should be to develop comprehensive behaviour change interventions based on behaviour change theory. In addition to the BCTs identified in this review, incorporation of others such as 'feedback on outcomes of behaviour' (2.7) and 'self-monitoring' (2.3) would seem logical, given their potential role for influencing adherence.

The maximum follow-up period identified in this review was 12 months. However, the aim of an intervention designed to improve adherence to oral hygiene advice would be to sustain change over much longer periods, preferably a lifetime. Therefore, future research also has a role in exploring the impact of digital interventions in terms of prolonged behaviour change.

This review has highlighted that there is significant heterogeneity in regards to outcome measures and interventions utilised in the current literature. Additionally, the risk of bias in the included studies is unclear. There is an increasing trend in the use of mobile phone technology, more specifically apps, in supporting patients with health care. In July 2018, a screening search of apps relating to oral hygiene on Apple's App Store and the Google Play Store retrieved 1,075 potential apps for inclusion.29 A detailed assessment of 20 apps for each search term utilised in this screening search revealed that the majority were developed after 2015, focused on the provision of oral hygiene information and were frequently free of charge. There was no indication of independent dental or oral health organisation approval or testing of effectiveness and acceptability for any of the apps. Given this availability, there is a need for practitioners to assess the quality and content of information available to patients and to direct patients towards high-quality, effective apps/resources to support them with their oral hygiene practices. A judgement must then be made in regards to recommending or guiding patients towards appropriate information resources to support their oral health. Furthermore, there is a need to assess mobile interventions utilising robust randomised controlled trial methodology including a core outcome set related to oral hygiene. This will help to ensure that the results of future studies may be synthesised in future systematic reviews.

Conclusions

There is some evidence to suggest that mobile phones are effective in improving adherence to oral hygiene advice in orthodontic patients. However, the generalisability of this review is limited, as the included studies were exclusively aimed at supporting orthodontic patients and were associated with an unclear risk of bias.

In the short-term, given the rapid proliferation of apps and other online information aimed at improving oral hygiene, there is a need to assess the quality and effectiveness of these resources, as this will help dental professionals navigate patients towards effective resources. In the medium- to long-term, this review suggests the need to develop mobile phone interventions grounded in behaviour change theory; using core outcomes to allow for meta-analysis and the assessment of cost effectiveness. Future studies should utilise a core outcome set related to oral hygiene and explore outcomes related to patient satisfaction and engagement with the technologies being tested, this may help to identify features of successful digital interventions.