Table 1 Characteristics of included systematic reviews
Study ID (Reference) | Number and type of included studies | Â | Target population | Health problem addressed in review (disease) | Prevention or Management | Location of studies (number) | Conclusions | Effect sizes | Knowledge gaps identified |
|---|---|---|---|---|---|---|---|---|---|
Cole-Lewis 2010 [19] | 12 studies: RCTs (9), Crossover RCTs (2), Quasi-experimental trial (1) | 2408 | Not specified | Behaviour change (weight loss, physical activity, diabetes, asthma, adherence to vitamin C) | Both | Canada (1), Finland (1), New Zealand (2), USA (2), France (1), Korea (2), UK (1), Croatia (1), and Austria (1) | There are short term effects on behavioral or clinical outcome related to disease prevention and management. Text messaging is a useful tool for behavior change interventions. | NA | 1. Methodological limitations in studies |
2. Text message characteristics and combinations should be explored | |||||||||
3. Long term effects should be investigated | |||||||||
De Jongh 2013 [18] | 4 studies: RCTs (4) | 182 | Not specified | Self-management of illness (diabetes, hypertension, asthma) | Management | Scotland (1), Croatia (1), USA (1), Spain (1) | Text messaging may support the self-management of long term conditions but have few direct impacts on health outcomes | NA | 1. Limited evidence of efficacy |
2. Long-term effectiveness unknown | |||||||||
3. Risks and limitations and consumer satisfaction are unknown | |||||||||
Finitsis 2014 [20] | 8 studies: RCTs (8) | 1785 | People living with HIV | Adherence to medication (HIV) | Management | USA (4), Kenya (2), Brazil (1), Cameroon (1) | Researchers should consider the adoption of a less than daily frequency of messaging that is individually timed and tailored and designed to evoke a reply from the recipient. | Odds ratio for adherence =1.39; 95% CI = 1.18-1.64 (8 RCTs) | 1. Comparisons of design and intervention characteristics to obtain optimal effect are needed. |
Gurol-Urganci 2012 [22] | 1 study: RCT (1) | 2785 | Not specified | Communicating results of medical investigations for anxiety (Down’s syndrome prenatal screen) | Management | Taiwan (1) | Unable to draw reliable conclusions dues to low quality of evidence coming from only one study. Positive and negative results delivered by text message may have different effects on anxiety | Mean anxiety score = −2.48; 95% CI-8.79 to 3.84 (1 RCT) | 1. Methodological limitations in studies |
2. Some outcomes of interest are: health-seeking behaviour, patients’ evaluation of the intervention, costs, economic benefits, and potential adverse effects. | |||||||||
Gurol-Urganci 2013 [21] | 8 studies: RCTs (8) | 6615 | Not specified | Attendance at healthcare appointments (not specified) | Both | China (2), UK (2), Malaysia (2), Kenya (1), Australia (1) | Mobile phone text message reminders increase healthcare appointment attendance rates when compared to no reminders and postal reminders. The current findings are insufficient to inform policy decisions | Relative risk for attendance rate at appointment = 1.14; 95% CI 1.03 to 1.26 (7 RCTs) | 1. Methodological limitations in studies |
2. Some outcomes of interest include: health effects, adverse effects and harms, user evaluation of the intervention and user perceptions of its safety. | |||||||||
Horvath 2012 [6] | 2 studies: RCTs (2) | 969 | People living with HIV | Adherence to medication (HIV) | Management | Kenya (2) | Weekly text messages are efficacious in improving adherence to ART in resource limited settings and may be efficacious in suppressing viral load. | Risk ratio for non-adherence at 48–52 weeks = 0.78; 95% CI 0.68 to 0.89 (2 RCTs) | 1. Larger RCTs in adolescent populations, and in persons who care for children and infants with HIV. |
2. Trials in high and middle-income countries are needed. | |||||||||
3. Data on acceptability, and culture-specific issues such as message-content and message-length are needed. | |||||||||
Militello 2012 [23] | 6 studies: RCTs (4), Crossover RCT (1), Quasi-experimental trial (1) | 433 | Pediatric and adolescent populations | Health promotion (diabetes, antirejection medication adherence, physical activity, diet and sedentary behaviour) | Both | USA (3), UK (2), New Zealand (1), Austria (1) | Text messaging should be considered as an add-on to clinic care to improve health behaviours | NA | 1. Methodological limitations in studies |
2. Long term effects and dose response data are of interest | |||||||||
Nglazi 2013 [24] | 4 studies: RCT (1), Observational (3) studies | 565 | Patients with tuberculosis | Adherence to medication (tuberculosis) | Management | Argentina (1), Kenya (1), South Africa (2) | The evidence is inconclusive on text messaging to improve adherence to TB treatment, but there is some potential | Risk ratio for adherence = 1.49; 95% CI 0.90 to 2.42 (one RCT). | 1. Outcome measures for TB cure, successful completion of TB treatment, and development of drug resistance should be standardized. |
Vodopivec-Jamsek 2012 [25] | 4 studies: RCTs (4) | 1933 | Not specified | Preventive health care (antenatal care, smoking, physical activity, diet and sedentary behaviour, adherence to vitamin C) | Prevention | Canada (1), Thailand (1), New Zealand (1), USA (1) | Text messages have the potential to contribute to health behaviour change in the short term alongside other media of health prevention information. . | NA | 1. Long term effects are unknown |
2. Data is needed on costs, and possible risks and harms | |||||||||
3. More information is needed for scale-up |