Background

Vaccinating children prevents an estimated 2.5 million deaths each year [1] and ensuring that vaccine coverage remains high is an important public health priority [2]. Despite this, global vaccine coverage was static over the last decade, and fell from 86% in 2019 to 83% in 2020 in the context of the COVID-19 pandemic, leaving an estimated 23 million infants under-vaccinated [3]. The reasons for under-vaccination are complex and multifactorial. Lack of the five ‘A’s—access, affordability, awareness, acceptance and activation—have been proposed as a taxonomy for the core contributing factors across a range of socio-geographical-cultural contexts [4]. Across the world, immunisation is largely coordinated at a population level, and typically as either national or state/provincial level programs [5]. Immunisation programs typically implement a fixed schedule of vaccination at specific age-based timepoints, although vaccines may also be scheduled to align with other events such as school or college entry and pregnancy.

Text messaging by short message service (SMS) via mobile (cellular) phones, has been used to deliver reminders to promote health behaviours, including for vaccination. The SMS content may assist to target specific barriers to vaccination like poor awareness, acceptance or access [6]. Compared to other communication channels, SMS is cheap, instantaneous, and less confrontational [7], and allows the recipient to attend to the message when convenient. Mobile phone coverage is now extensive in both developed and developing settings [8] across income levels [9], enabling broad capture of the population [10]. Although mobile network connectivity has rapidly expanded globally, uptake of health interventions driven through mobile phone technology (mHealth) have been slower in low-middle income settings compared to high income settings, likely due to limited availability of technical support and infrastructure investment to support scaling [11].

Three recent systematic reviews summarised research assessing the effect of SMS-based interventions on childhood vaccine coverage in low-income [12], low-middle income [13] and both high and low-income settings [14]. We sought to update these reviews with newly published research, including studies of adolescents due for vaccination, and including data relating to the effectiveness of SMS-based interventions on vaccine timeliness. A growing number of vaccines are now targeted toward adolescents, and they are a distinct demographic from children and adults. Adolescents are likely to fall somewhere between children and adults with respect to both the achieved uptake of recommended vaccines, and the extent to which they, versus their parents, are responsible for their healthcare decision-making. Furthermore, this group may interact with technology, and hence SMS reminders, differently from other groups.

PICO statement

The aim of this systematic review was to examine: for parents of children or adolescents (< 18 years) eligible for a routine vaccination (P), what is the impact of SMS reminders (I) on vaccine coverage and timeliness (O), compared to standard care or other reminder methods (C).

Methods

Search strategy

This PROSPERO registered systematic review (CRD42016048290) was conducted in accordance with PRISMA guidelines [15]. We searched PubMed, Medline, Embase, Cochrane, Cumulative Index to Nursing and Allied Health Literature (CINHAL), PsycINFO and Web of Science for studies published through to July 2022 using the following search terms in a Boolean strategy: vaccination, immunization, immunisation, immunis*, immuniz*, immunis*, SMS, smartphone, telemedicine, mHealth, mobile health, short message service, cell phone, text messaging, text reminder and mobile phones (see Appendix 1). The search was limited to full-text studies written in English involving adolescents or parents of children less than 18 years old. Additional papers were identified through reference searching of peer reviewed manuscripts and grey literature.

Eligibility criteria

We included randomised studies examining (i) the impact of SMS-based interventions on coverage and/or timeliness of child vaccines. We included studies that compared alternative SMS-based reminder strategies without a non-SMS control group (e.g. postcard reminders). We excluded studies where adjunctive interventions were also used (e.g. flyers or education) that i) did not report the effects of SMS-based reminders only or ii) where the control group did not receive the same adjunctive intervention as the SMS-based reminder group. We excluded randomised studies that did not randomise to a control arm. We excluded non-randomised studies (i.e. original observational studies) due to the availability of higher quality randomised studies, especially considering most SMS evaluations compare before-versus-after designs, and non-randomised studies introduce a high risk of confounding by temporal factors.

Study definitions

Vaccine coverage was defined as the proportion of vaccine-eligible children within a study group who received all specified vaccine(s) within a defined time-period. Vaccine timeliness was defined as a measure of vaccine administration relative to the due date, either (i) the proportion vaccinated within a set period after the scheduled date or (ii) the time to vaccination after the scheduled date. Low-middle and high-income countries were categorised according to World Bank definitions [16] and analysed separately. The impact of SMS-based interventions on special interest groups or vaccines and whether interventions were issued as pre-call (prior to the due date), or recall (after the due date) were also analysed separately for vaccine coverage.

Study selection, data analysis, and bias

Two reviewers (GC and CM) independently performed and screened the search output and reviewed potentially eligible full-text studies after removing duplicates. Studies were summarised by design, study population, intervention and comparator groups, outcomes and limitations. The primary reviewer (GC) performed study quality assessment using the National Heart Lung and Blood Institute (NHLBI) checklist for randomised trials [17]. Ten percent of data extraction and bias assessments were randomly selected and cross-checked for accuracy by the second reviewer (CM). Discrepancies between the primary reviewers were resolved by consensus, or where necessary by a third reviewer (TS). A meta-analysis was not performed owing to the marked heterogeneity of included studies. Findings are therefore described by narrative review.

Results

Search results

A total of 536 publications were identified after removing duplicates; after screening abstracts, 44 papers were selected for full text review (See Fig. 1 for flow diagram). Of these, 30 met the inclusion criteria and were included in the final review (Table 1). Two trials were excluded from review as they did not assign participants to a control group, instead comparing recipients of different SMS reminders to no control [18] and non-enrolled parents in the study [19].

Fig. 1
figure 1

PRISMA flow diagram of the search results

Full size image
Table 1 Randomised Trials examining the effectiveness of SMS interventions
Full size table

Study setting and participants

Of the selected trials, 19 were conducted in high-income countries and 11 were conducted in low-middle income countries. Sixteen of 19 trials conducted in high-income countries were limited to the United States and targeted parents of children from low-income or ethnic minority groups, and children attending tertiary-affiliated, private paediatric clinics or local hospitals (Table 1). One study targeted parents attending a local baby exhibition event [36]. Trials in low-middle income countries were conducted in Nigeria (3), Kenya (2), India (2) Guatemala (2), Pakistan (1), Zimbabwe (1). In the trials from high-income countries, the SMS-based intervention recipients were predominantly English-speaking and female. Maternal education levels were more commonly reported among trials conducted in low-middle income countries.

Interventions and comparator

The SMS-based interventions were compared against a range of comparators ranging from routine care (no SMS or reminder) (16), written reminders (7), telephone reminders (either from practice staff or automated calls) (3), sham or health-related SMS-based reminders unrelated to vaccination (3) health education (1). Two RCTs compared the effectiveness of SMS-based reminders when combined with monetary or phone credit incentives compared to SMS-based reminders alone, or other strategies [6, 28].

Study quality

Please see Table 1 for individual study quality and risk of bias assessment. Seventy percent (21/30) of the trials were deemed to be of fair to good quality. The most frequently identified sources of bias were related to poor or poorly documented randomisation procedures, or a lack of adequate detail regarding allocation concealment or blinding of practice staff.

Effect of SMS-based interventions on vaccine coverage

Low-middle income countries (LMIC)

Eight of ten trials conducted in LMICs reported higher vaccine coverage among children of parents who received SMS-based reminders compared to non-SMS interventions or routine care (see Table 1) [20, 21, 23,24,25,26,27, 29]. Two of ten trials found no evidence of an effect of SMS-based reminders alone on vaccine coverage compared to no SMS, but found evidence of a small effect when SMS-based reminders were combined with a monetary incentive [6] or phone credit incentive [28]. One trial found evidence that SMS-based reminders were more effective than control (no SMS reminder), and that the effectiveness of SMS-based reminders was greater when combined with incentives in the form of high phone credits [21].

High income countries (HIC)

Ten of 17 trials conducted in HICs [34, 35, 37, 38, 42,43,44,45,46] reported small or modest improvements in vaccine coverage among children of parents receiving SMS-based reminders compared to those who received no SMS or alternative non-SMS strategies; the remaining 7 trials [30, 32, 33, 36, 39, 41, 47] found no evidence of an effect of SMS-based intervention compared to non-SMS control (appointment cards, alternative health messages or no reminder).

Two of the 17 trials reported improvements that were limited to specific timepoints or in specific recipient groups, but not all [33, 35]. One of the 2 trials found evidence that SMS-based reminders were effective compared to control (no SMS reminder) for vaccines scheduled at 12 months-old only, with the effect slightly greater when SMS-based reminders were combined with a personalised calendar; a post hoc analysis found evidence of a greater effect among children who had been late for any previous vaccine [35]. The second trial reported no overall difference between groups (SMS reminders and SMS vaccination appointment reminder, SMS vaccination appointment reminder only & control) for receipt of MMR vaccination, but a sub-group analysis reported a difference for parents who did not have an appointment prebooked in the SMS reminder and appointment reminder arm compared to SMS only and control [33].

Regarding message content, six of the 17 trials compared SMS-based reminders with embedded educational/persuasive content; of these 5 found evidence of increased vaccine coverage compared to plain SMS-based reminders without these features [34, 37, 42, 43, 46].

Of the 17 studies, one trial reported improved vaccine coverage among parents receiving reminders through interactive messaging (ability to exchange bidirectional messages or receive further information) compared to no SMS reminder [38], and one reported interactive messaging in combination with educational SMS reminders resulted in higher coverage compared to educational SMS only or telephone reminders [34].

SMS-based interventions: pre-call and recall vaccine reminders

In 16 trials, SMS-based reminders were issued prior to vaccine due-dates; of these 12 found evidence that vaccine coverage was higher in the SMS-based intervention group than the comparator group [6, 20, 21, 23, 25,26,27, 29, 34, 43, 45, 46]. Among the 4 trials that found no evidence of a difference in coverage [30, 33, 41, 47], two trials reported significant implementation problems in the intervention group including a high rate of failed SMS delivery [30, 33].

In 6 trials [24, 34, 38, 43, 44], SMS-based reminders were issued to parents whose children were already overdue for receipt of a recommended vaccine; all found evidence that vaccine coverage was higher in the SMS-based intervention group compared to control.

In 3 trials both pre-call and recall SMS-based reminders were used [27, 35, 37]; two trials reported improved vaccine coverage in the intervention groups compared to control [27, 35], and one reported that receipt of an SMS-based reminder was only effective if it contained an educational message [37]. No trials directly compared pre-call to recall message strategies.

Special interest groups/vaccinations

Four trials [32, 38, 39, 44] examined the effect of SMS-based reminders for adolescent vaccines, including HPV and meningococcal vaccines; two trials [38, 44] reported evidence of higher vaccine coverage among SMS-based reminder recipients (parents in 4 studies, and either parent or adolescent in 1 study) compared to no SMS, other comparator groups, or historical control.

SMS-based interventions and timeliness of vaccination

Low-income settings

All five trials [6, 20, 22, 26, 28] conducted in low-income countries found evidence of improved vaccine timeliness in children of parents receiving SMS-based reminders compared to control. One trial reported that compared to a control group, there was improved vaccine timeliness in groups who received an SMS-based reminder either with or without a monetary incentive [6]; the other trial did not find evidence that SMS-based reminders alone improved timeliness, but found evidence that an SMS-based reminder plus a phone credit incentive did improve timeliness compared to control [28]. One trial found evidence that standard SMS and educational SMS-based reminders had a similar and superior effect on vaccine timeliness compared to control (no reminder), but phone call reminders appeared to be more effective than either SMS-based intervention [26].

High-income settings

Of 7 trials that reported on vaccine timeliness, 5 found evidence that SMS-based reminders improved vaccine timeliness compared to standard care [31, 33, 34, 40, 42]. One trial reported that compared to a standard SMS-based reminder or non-SMS control, a higher proportion of children whose parents received an educational SMS-reminder received a timely second dose of influenza vaccine; there was no difference in timeliness between the standard SMS and control [42].

Discussion

Compared to alternative strategies to try to improve vaccine coverage and timeliness, SMS-based strategies are instantaneous, convenient, scalable, have potential for automation, and are relatively low cost [48]. We found evidence that they can be effective in both low-middle- and high-income country settings, but where effect was observed, it was usually small to moderate in size, with the greatest observed effect for vaccine coverage being a risk ratio of vaccination of 1.36 (see Table 1).

The SMS-based interventions evaluated varied in several respects; some included educational content, some were combined with incentives, and some were delivered as recall rather than as pre-call reminders. The apparent effectiveness of these strategies varied across settings; for example, one of the more robust LMICs studies reported that SMS combined with airtime incentives were most effective for parents in Pakistan compared to SMS only and no reminder [21]. The three largest trials examining the effectiveness of SMS-based reminders on vaccine uptake, found no evidence of an effect compared to control [32, 41, 47]. Baseline vaccine coverage was low in these trials, and we note that none of these studies used SMS reminders with educational or persuasive content. We speculate that plain SMS-based reminders might only be effective where population acceptance of vaccination is already high. No trials were identified which directly compared pre-call to recall SMS-based reminders.

The differential effect of SMS-based interventions across socioeconomic groups within the same setting has not been extensively studied. In many settings, children from low income families have lower rates of vaccine coverage [49]; reduced health literacy and logistical barriers such as poor access to primary healthcare have been reported as potential contributing factors [50, 51]. SMS-based reminders may be effective for families with limited access to other forms of communication, such as email [52]; however, some studies have reported specific barriers to SMS in families with low-socioeconomic status, including unreliable service delivery [23] and changing contact details and service providers [30]. In some settings mobile phone service providers require the SMS recipient to have sufficient credit to receive messages; this may not be relevant to all settings.

We sought to understand whether there would be observed differences in the impact of SMS reminders across different contexts, including childhood and adolescent vaccinations. There was a paucity of evidence assessing impact of reminders on adolescent vaccinations; two of the four studies reported improvements, however only one study was considered good quality [38]. Among the studies that reported higher vaccination uptake, these improvements were broadly comparable to improvements observed in trials in childhood vaccination (up to 30% in coverage). No studies directly compared the effectiveness of SMS reminders delivered to adolescent recipients versus their parents, which would be helpful to ascertain which is most effective, and whether different messaging strategies for each are required.

It can be difficult to know whether an SMS has been received, read, and understood by the intended recipient. Bidirectional messaging, wherein SMS messages are sent back-and-forth between the recipient and the vaccine provider, may be used to confirm receipt of the message and/ or understanding of its content, or to provide supplementary educational material to parents prior to vaccine appointments. While we identified some evidence of the effectiveness of bidirectional messaging in two trials [34, 38], the cost and burden on providers to issue more personalised messaging needs to be considered.

We identified evidence that SMS-based reminders had improved efficacy where the messages included educational content, especially for vaccines that may not be part of a routine vaccine schedule, such as for influenza vaccine. Trust between parents and vaccine providers has been identified as important in preventing vaccine hesitancy [53]. This may indicate that educational or persuasive SMS reminders from providers that have a strong and trusting relationship with families may be a determinant of vaccination behaviour.

SMS-based strategies may represent an opportunity to directly address adverse vaccine beliefs through educational messaging. However, more research is needed to determine what educational content and message framing is most effective (e.g. benefit versus risk-based message framing). Many SMS services impose a message character limit, so achieving a message with sufficient content to motivate action is a challenge [34]. There may be benefit to developing educational content which is based on behavioural theories or frameworks such as the health belief model [54].

Strengths and limitations

This review included trials across a range of contexts, including high- and low-middle income country settings. We also included vaccine timeliness as an outcome of interest as prior research has indicated that important delays in vaccine receipt may exist even in settings with high vaccine coverage [55,56,57]. Timeliness was less frequently reported as an outcome than vaccine coverage. The decision for a narrative review has limited our ability to summarise the effect size of SMS-based interventions. Meta-analysis was not suitable due to the vast heterogeneity of the interventions, contexts of the studies, and the outcomes measured and reported.. We only included trials in this review, although we note that a number of observational studies have reported on the post-implementation impact of SMS-based reminders, and these might provide additional insights into the apparent heterogeneity in effects.

Conclusions

We found evidence that SMS-based reminders can have a beneficial effect on the coverage and timeliness of routine vaccines in childhood across a range of LMIC and HIC settings. We found some weak evidence of the effectiveness of educational versus standard (non-educational) SMS message content, and for an additional effect of monetary or phone credit incentives, although more studies are needed to corroborate these findings. No studies directly assessed the effect of pre-call versus recall timing of messages. As such, neither the optimal message content (i.e. plain versus educational/persuasive) nor optimal timing of SMS-based reminders have been clearly determined. Multi-arm or factorial-design trials evaluating alternative options for SMS content and timing in varying combinations and across different age groups and programmatic contexts could help to address these gaps [58]. Trials should also assess their cost-effectiveness when delivered as vaccine pre-call versus recall, or in the context of targeted and possibly multifaceted strategies which are tailored for specific populations.