Table 1.

Randomised Trials examining the effectiveness of SMS interventions

Author/year	Population	n	Intervention group	Comparator group	Outcome measure	Results	Limitations	Quality rating (Good, fair, poor)/assessment of bias
Low-Middle Income
Bangure et al. (2015) [20]	Zimbabwe: mothers recruited following delivery of newborn	304	SMS reminders at 6, 10 and 14 wks	Routine education	Vaccination coverage and timeliness at 6, 10 and 14 wks	Coverage and timeliness increased at all time points in the intervention group (p < 0.001)	Unsure if blinding occurred to control performance bias	Fair
Chandir et al. (2022) [21]	Pakistan: parents of children < 2 years presenting for first vaccines	11,197	5 arms of differing mobile conditional cash transfers (mCCT) + SMS, SMS only	No SMS reminder	Full vaccination coverage at 12 months	High flat rate incentives + SMS (OR: 1.30, CI 1.11–1.51), High sharp rate incentives + SMS (OR: 1.27, CI 1.09–1.48) & SMS only (OR: 1.16, CI 1.00–1.34) superior compared to control	Some participants could not receive mCCTs due to mobile phone incompatibility	Good
Domek et al. (2019) [22]	Guatemala: parents of children between 6wks to 6mo who presented for their first vaccines	720	One SMS reminder sent 3 days, 2 days and 1 day prior to 2nd and 3rd vaccine visits	Routine care	Vaccination timeliness at 2, 4 and 6 months	Coverage similar across groups. Intervention group received vaccines on scheduled dates of visit 2 (42.2% vs 30.7%, p-0.001) and 3 (34% vs 27%, p = 0.05) and within 7 days of visit 2 (71% vs 63.5%, p = 0.03) compared to control	Vaccine shortages present in Guatemala during study so used attendance as proxy for status. Some errors in SMS system where not all participants were sent SMS	Fair
Domek et al. (2016) [23]	Guatemala: caregivers of children presenting for 1st vaccines at 8–14 wks	321	SMS reminders 1 wk prior to 2nd/3rd vaccines	Routine care	Vaccination coverage: completion of the primary immunisation series	Both intervention/control groups had high rates of vaccine completion (visit 2: 95 vs. 90% and visit 3: 84 vs. 81% respectively	Pilot study. Service interruptions	Fair
Eze et al. (2015) [24]	Nigeria: caregivers attending health clinics	905	SMS reminders to everyone in intervention group and additional recall SMS to parents who did not attend appointment	Routine care	Vaccination coverage and vaccination timeliness of receipt of DPT3 prior to 18th week	Intervention group DPT3 8.7% higher coverage and received DPT3 1.5 × earlier than controls (OR 1.47, CI 1.1–2.0, p = 0.009)	Inconsistent DPT product supply. Randomisation procedures did not account for mobile phone ownership	Poor
Gibson et al. (2017) [6]	Kenya: parents of newborns residing in rural villages	2,018	1) SMS only (3 and 1 day prior to scheduled vaccines at 6, 10 and 14 weeks & 9 months); 2) SMS + low monetary incentive; 3) SMS + higher monetary incentive	Routine care	Vaccination coverage: Proportion of fully immunised children at 12 months (including BCG). and Hepatitis B), measles and BCG vaccines. Vaccine coverage & vaccination timeliness (within 2 weeks) for pentavalent, polio and measles vaccines	SMS + higher monetary incentive group more likely to achieve primary outcome (RR 1.09, (1.02–1.16), p = 0.014). No difference between SMS only or SMS + low monetary incentive group. 20 outcomes: Improved timeliness of measles vaccine seen in all 3 intervention groups but highest in SMS + high incentive group (RR 1.42 (1.23–1/65), p < 0.0001)	Incomplete information about whether reminders were received/read. Randomisation assignment cluster randomised by village at public ceremony. High baseline immunisation rates	Fair
Haji et al. (2016) [25]	Kenya: parents of children residing in low pentavalent coverage districts	1,116	1) SMS reminder; 2) sticker reminders	Routine care	Vaccination coverage: dropout rate (missing 2nd/3rd pentavalent vaccine doses) 2 weeks after scheduled visit for 3rd vaccine	SMS group 20% less likely to drop out compared to control (OR 0.2, CI 0.04–0.8)	Randomised at clinic level. Randomisation methods not detailed. No specification of intention to treat analysis	Poor
Ibraheem et al. (2021) [26]	Nigeria: mother-infant pairs present for first vaccination	560	1) SMS reminders; 2) Educational SMS; 3) Phone call reminders	Routine care (no reminder)	Vaccination coverage and vaccination timeliness	All intervention groups had higher completion rates compared to control. Timeliness of plain SMS (AOR 2.56, 1.96–3.35) and educational SMS (AOR: 2.44, 1.87–3.18) similar odds and superior compared to control. Calls superior to SMS	Reduced generalisability. Randomisation methods not described in detail. Demographics table not broken into randomised groups. No specification of intention to treat analysis	Fair
Kawakatsu et al. (2020) [27]	Nigeria: parents of children attending primary health centres	9,368	SMS reminder	Usual care (verbal and written reminders)	Vaccination coverage: Antenatal and family planning appointment attendance	SMS group had higher vaccine uptake compared to control (4.8–6% higher at all time points, p < 0.001) and more likely to receive vaccines (AOR: 1.17, 95% CI: 1.05–1.31)	Randomisation did not account for the appointment type so required further statistical adjustments	Good
Seth et al. (2018) [28]	India: pregnant women and parents of children < 24mo in rural India	608	1) SMS reminders + phone credit incentives; 2) SMS reminders only	Written reminder	Vaccination coverage for all required vaccines at study end Vaccination timeliness within 14 days of each vaccine	SMS + phone credit incentives group had higher vaccination coverage (RR 1.09, CI1.002–1.18, p = 0.04) and more timely vaccine receipt (40%) compared to SMS only and control	Low literacy level of study population. Study field staff not blinded to random allocation. No intention to treat analysis. No power calculations provided	Poor
Shinde et al. (2018) [29]	India: mothers of children 0–3 weeks old attending the maternity ward	125	SMS reminders	Immunisation card	Vaccination coverage at 10 wks	SMS reminder group had higher coverage at 10w compared to control (95% vs 77%, p = 0.011)	Low sample size. Randomisation concealment not described	Fair
High-Income
Ahlers-Schmidt et al. (TRICKS) (2012) [30]	USA: parents of newborns discharged from hospital in Kansas	90	SMS reminder + appointment card	Appointment card	Vaccination coverage and timeliness of vaccines at 2, 4, 6 months	Greater numbers in intervention group received vaccines and on time, but not statistically significant	Pilot; small sample size. Selection bias (differed in income status) and attrition bias due to loss of phone service. Problematic software	Poor
Coleman et al. [31]	USA: Parents of preschool children in New York	57	SMS reminder	Written reminder	Influenza vaccine timeliness	Time to vaccination shorter in SMS group (42 days vs. 62 days; p < 0.05)	Small sample size. Performance and selection bias (randomised per patient preference & 8% had sibling in control group)	Poor
Gurfinkel et al. (2021) [32]	USA: parents of children due for initial or subsequent HPV doses in New York	37,003	1) SMS; 2) Autodial reminder	Usual care (no reminder)	HPV vaccination coverage for 1, 2 or 3rd dose. Timeliness to completion	No significant difference between groups for uptake or timeliness	Randomisation procedures not described in detail. Unknown if randomisation concealment occurred. Baseline characteristics of participants per arm not provided	Fair
Hofstetter et al. (2015) [33]	USA: parents of children due for 12-month check-up (9.5mo-10mo) in New York	2054	1) SMS appointment reminder (2 days prior to scheduled 1 year appointment); 2) × 3 scheduling reminders to book vaccination appointment AND SMS appointment reminder as above	Usual care (routine telephone reminder)	Vaccination coverage: 12-month appointment attendance Vaccination timeliness of MMR vaccine	No difference in MMR vaccination by 13 months between groups. Only in post-hoc for parents that had not booked an appointment prior to study (attendance rates: 62.1% vs. 54.7%, relative risk ratio 1.14 95% CI 1.04–1.24). More likely to have timely MMR vaccine (61.1% vs. 55.1%, relative risk ratio 1.11 95% CI 1.01–1.21)	Specific low-income, minority, Spanish-speaking families, limiting generalisability to other settings. No mention of intention to treat analysis or randomisation concealment	Fair
Hofstetter et al. (2015) [34]	USA: Low-income, minority parents of under-vaccinated children (6mo-17y) in New York	5462	1) Educational + interactive SMS; 2) educational only SMS	Telephone reminder	Influenza vaccination coverage Influenza vaccination timeliness	More children in educational & interactive group vaccinated than other two groups (38.5% vs. 35.3% vs. 34.8%; RRR: 1.09, 95% CI 1.002–1.19) & more timely vaccination than educational only (AHR = 0.90, 95% CI = 0.81–1.00) and standard care (adjusted hazards ratio = 0.88, 95% CI = 0.79–0.98)	Urban low-income participants, limiting generalisability to other settings. No mention of randomisation concealment	Good
Menzies et al. (2020) [35]	Australia: Parents of children due for routine childhood vaccines	1,594	(1) SMS reminder only; (2) Calendar only; (3) SMS reminder + calendar	No reminder	Vaccination coverage: 30 days within due date of 2,4,6,12 and 18 month vaccines	SMS reminders alone (RR 1.09, 95% CI 1.01–1.18) or in combination with a personalised calendar (1.11, CI 1.03–1.20) higher compared to control at 12 m endpoint only	Low sample size in 4 m timepoint. High compliance in control group compared to national statistics. No mention of randomisation concealment or blinded assessment. Major policy change required unplanned subgroup analysis	Good
Niederhauser et al. (2015) [36]	USA: mothers/neonates in Hawaii	57	SMS reminders sent 4 and 2 weeks prior to the 2, 4 and 6 mo. Vaccines	Sham SMS of age-appropriate newborn health messages sent at equivalent time point	Vaccination coverage: compliance with vaccinations + 7 and + 14 days post due-dates. Barriers SHOTS survey	At all assessment points (bar 1 time point), the control group had higher rates of vaccine compliance, although higher barriers in intervention group	Pilot; small sample size. Selection and attrition bias due to high drop out rate. Randomisation procedures not described in detail. Groups not equal at baseline. No intention to treat analysis performed	Poor
O’Grady et al. (2022) [37]	Australia: mothers of children attending primary care clinics	196	1) SMS reminder; 2) Educational SMS reminder	No reminder	Vaccination coverage at 7 months for 2, 4 and 6 month vaccines Vaccination timeliness in days	Improved vaccine coverage at all timepoints in educational SMS reminder compared to control (7mo ARR: 2.28 95% CI 1.05—4.94). There was no difference between simple SMS reminders and control. Timeliness: no differences between either SMS group and control	Low sample size. Groups not similar at baseline; authors state this did not impact results in their regression analysis	Good
O’Leary et al. (2015) [38]	USA: parents of adolescents attending 5 private, 2 public clinics in Colorado	4587	SMS bidirectional message (1) Clinic will call to schedule (2) Parent will call (3) STOP	No reminder	Vaccination coverage: receipt of all vaccinations and any vaccine	Intervention group more likely to receive all vaccinations (RR 1.29, 95% CI 1.12–1.5) and any vaccine (RR 1.36; 95% CI 1.2–1.54)	Didn’t directly compare unidirectional and bidirectional messaging	Good
Rand et al. (2015) [39]	USA: Adolescents 11–16 years with no prior HPV at 39 practices in New York	3812	SMS reminders (up to 4) to parents of adolescents for HPV	Sham SMS of general health messages	Vaccination coverage: receipt of HPV1, receipt of HPV2 & 3	No differences between groups for any dose. Post-hoc analysis for those able to receive message showed 30% HPV1 (HR 1.3, 95% CI 1–1.6)	Only half participants had a phone with SMS capability despite being randomised. Single centre. Planned stratified analysis limited by insufficient sample size. Unclear if lack of randomisation concealment impacted performance bias	Poor
Rand et al. (2017) [40]	USA: Parents of adolescents 11–16 years in 3 urban primary care clinics presenting for HPV 1 or 2 in New York	749	1) SMS reminder; 2) telephone calls	No reminder	Vaccination coverage as receipt of all 3 doses. Vaccination timelines: time to receipt of HPV vaccination	49% SMS vs. 40% controls received 3 HPV doses (p = 0.001). Time to receipt HPV3 greater in SMS group (HR 2.34, p < 0.001) and phone group (HR = 1.91, p = 0.007) who enrolled at time HPV1 vs. controls	Limited sample size. Didn’t directly compare SMS vs. phone reminders. Randomisation was based on parental preference of receiving SMS or phone	Fair
Szilagyi et al. (2020) [41]	USA: Parents of children in New York primary care practices	61,931	1) SMS reminder; 2) autodial reminder; 3) mailed reminders	No reminder	Influenza vaccine coverage within 6 months	No difference between SMS reminders and control group (27.6% vs 26.6%)	No mention of randomisation concealment or blinding procedures. No mention of how sample size was calculated	Good
Szilagyi et al. (2020) [41]	USA: Parents of children attending elementary school in New York	15,768	1) SMS reminder + school located vaccination	Usual care (autodial reminder and consent packet) and school located vaccination	Influenza vaccination coverage within 6 months	No difference between SMS reminder group and usual care groups (4.4% vs 4.3%)	High opt-out rate. No mention to randomisation concealment or blinding procedures. No mention of how sample size was calculated	Good
Stockwell et al. (2015) [42]	USA: Low-income, urban, minority (Latino) population presenting for 1st flu vaccine between 6–18 months of age in New York	660	1) SMS scheduling reminder + written reminder; 2) SMS educational + written reminder	Written reminder only	Vaccination coverage: receipt, and Vaccination timeliness of second dose of influenza vaccine	Educational SMS arm more likely to receive 2nd dose (72.7% vs 66.7% vs 57.1%, p = -0.03) and have timely receipt (p < 0.001)	Low-income minority population may limit generalisability	Good
Stockwell et al. (2012) [43]	USA: Low-income, mostly Latino parents in New York	9213	Up to 5 weekly SMS educational & scheduling reminders	Routine care	Vaccination coverage: receipt of influenza vaccine	Higher proportion in intervention group (43.6%, n = 1653 vs. 39.9%, n = 1509; RRR 1.09 95% CI 1.04–1.15, p = 0.001)	Possible selection bias as there were 8% of siblings allocated to opposite group	Good
Stockwell et al. (Text4Health: Adol) (2012) [44]	USA: Low-income parents of under-vaccinated adolescents (11y-18y) in New York	361	SMS reminder	No reminder	Vaccination coverage: meningococcal and Tdap dose uptake	More adolescents in intervention group received meningococcal & Tdap at 24 weeks compared to control (36.4% vs. 18.1%, p < 0.001)	Potential under-reporting of vaccination receipt. Randomly selected intervention and control groups from cohort instead of traditional 1:1 randomisation procedure. Unsure if blinding occurred to negate performance bias	Poor
Stockwell et al. (Text4Health: Paeds) [44]	USA: Low-income parents of under-vaccinated children (7mo-22mo) in New York	174	SMS recall + letter	Letter recall	Vaccination coverage: Hib vaccine uptake	More children in intervention group received Hib vaccine compared to control (21.8% vs. 9.2%, p < 0.05)	Unsure if blinding occurred to negate performance bias. Low sample size	Fair
Tull et al. (2019) [4]	Australia: Parents of adolescents due for HPV vaccine	4,386	1) Motivational (educational) SMS; 2) Self-regulatory SMS	No SMS reminder	Vaccination coverage: HPV vaccine uptake	Both SMS reminder group similarly effective to improve vaccination rates (88.35% vs 89) compared to control (85.7%, p < 0.016)	Sample skewed towards metropolitan schools. Adolescents had to consent to receive a vaccine before being sent a reminder	Good
Wiseman et al. (2016) [45]	USA: Parents of children attending a primary care clinic in Arizona	136	SMS reminder	Sham health-related SMS	Vaccination coverage: influenza vaccine uptake by end of influenza season	More children in SMS reminder group received vaccine compared to sham SMS (83.5% vs 45.4%). OR: 4.46, 1.704–11.706, < 0.001)	Small convenience sample and potential selection bias. Randomisation procedures not adequately described	Poor

Abbreviations: AHR Adjusted hazards ratio, AOR Adjusted odds ration, BCG Bacillus Calmette–Guérin, CI Confidence interval, DPT Diphtheria pertussis tetanus, Hib Haemophilus influenzae type b, HPV Human papilloma virus, HR Hazard ratio, mCCT Mobile conditional cash transfer, mo Month, OR Odds ratio, RR Risk ratio, RRR Relative risk reduction, SMS Short message service, wk Week