Mobile phone communication for improving uptake of antiretroviral therapy in HIV-infected pregnant women: updated systematic review and meta-analysis

1. Background

In 2020, there were 37.7 million HIV-positive people worldwide, with women and girls accounting for 53% of all HIV-positive people, and children accounting for 1.7 million [1]. While progress has been made in these priority countries, much more effort is required to meet the Global Plan’s target of 90 percent reduction in new infections among children [2]. Every week, approximately 5000 young women between the ages of 15 and 24 become infected with HIV [1]. Six out of every seven new HIV infections among adolescents aged 15–19 years in Sub-Saharan Africa are among girls [3], young women aged 15–24 years are twice as likely as men to be HIV positive. In 2020, approximately 4200 adolescent girls and young women aged 15–24 years were infected with HIV every week [3]. The most important factor in preventing perinatal transmission in HIV-positive women is antiretroviral therapy (ART) during pregnancy [4]. Without treatment, the risk of perinatal transmission is 25%, which drops to 1.8 percent with zidovudine alone and 0.5% with ART treatment regimens [4]. Individuals living with HIV should begin ART and achieve viral suppression prior to pregnancy, regardless of method of conception [5]. Full viral suppression for at least three months prior to conception is recommended, or at least two viral load measurements below the level of detection at least one month apart [6]. Several large studies have found that suppressing the plasma HIV viral load reduces the risk of sexual transmission [6–8]. Indeed, new maternal HIV infection among pregnant women significantly contributes to mother–to–child HIV transmission (MTCT) [9–11]. According to UNAIDS 2020, new maternal HIV infection during pregnancy was the second leading cause of perinatal HIV transmission in 2019 [3].

Pregnancy is a high-risk period for HIV acquisition [12]. Thomson et al. found that the per-act probability of HIV acquisition increased by 2.8 during pregnancy compared to non-pregnancy [12]. The pooled HIV incidence among pregnant women in Sub–Saharan African and other countries was reported to be 2.1 per 100–person years in a meta–analysis of studies conducted between 2014 and 2018 [13]. The majority of HIV-infected children are infected through MTCT, which can occur during pregnancy, labour and delivery, or breastfeeding. In the absence of any intervention, the risk of such transmission in non-breastfeeding populations is 15–30% [14]. Breastfeeding by an infected mother raises the risk by 5–20%, for a total of 20–45% [14]. Preterm birth, prolonged rupture of membranes, concurrent syphilis, chorioamnionitis, maternal viral load >100 000 copies/mL, and obstetrical interventions, including elective caesarean delivery, are all risk factors for perinatal transmission [14, 15]. However, among women with undetectable viral loads, the length of time between membrane rupture and birth may not be associated with an increased risk of perinatal transmission [16]. When compared to women with chronic and well-managed HIV, acquiring HIV during pregnancy or lactation increases the risk of perinatal transmission by 15 and 4-fold, respectively [17–20]. In newly acquired maternal HIV infection, starting ART as soon as possible after infection can reduce maternal HIV viral load and the risk of MTCT [21]. ART reduces the risk of HIV transmission from mother to child [22–24]. Increased access to ART during pregnancy and breastfeeding has had the greatest impact on reducing perinatal HIV-1 transmission. According to UNAIDS, active management of pregnant women living with HIV-1 has reduced new HIV-1 infections among children by half, from 310 000 in 2010 to 150 000 in 2019 [1]. Despite the evidence showing pregnant women having the highest level of adherence to ART [25], a low level of adherence has been reported in some settings among pregnant women [26]. In fact, the adherence rate among pregnant women varies across different settings (within and across countries) from 35% to 93.5% [27]. A study reported low adherence levels among pregnant women in rural Kwazulu Natal, South Africa [26]; however, a high level of adherence (90%) was reported in Kenya [28]. An adherence level of 73% during pregnancy was reported in Malawi, but dropped to 66% by three month postpartum [26].

Reviewing the literature, the field of HIV adherence interventions is quite broad. Major reviews of adherence interventions, primarily conducted in developed countries, revealed that the most effective are typically patient-centered, behavioral interventions designed to increase patient knowledge and efficacy through practical medication management skills [29–33]. While some social and behavioral scientists have become heavily involved in developing adherence measures, studying explanatory factors, and developing programs to improve patient adherence, others [33–36] have argued, l’arbre qui cache la forêt’ [33, 36], meaning patient adherence is multidisciplinary and complex. Coomes et al. proposed a conceptual model that integrates short message services (SMS) communication functionality with important psychosocial factors that could mediate the impact of SMS communication on health outcomes [33, 37]. This model is based on limited but growing research that has used SMS to assist patients in self-management or adherence across a variety of disease conditions [33, 37]. The use of mobile phones to improve HIV-related health outcomes is being emphasized in most HIV-affected countries, as emerging evidence suggests that reminder SMS can increase adherence to ART and retention in care, decrease viral load and treatment interruptions, and improve communication with healthcare personnel [38]. The main aim of this systematic review is to provide some evidence on the use of mobile phone communication for improving ART adherence during pregnancy, as well as to investigate whether mobile phone text messaging could improve follow-up in HIV-exposed infants.

2. Methods

This systematic review and meta-analysis was updated following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [39]. The study protocol was registered in Prospero: CRD42016032800. The previous review version was published [33].

3. Results

Figure 1 depicts electronic data sources and the selection process. Initially, CENTRAL (Cochrane Central Register of Controlled Trials), Scopus, MEDLINE via PubMed, Web of Science, and CINAHL (EBSCO) were used to identify a total of 2020 records. After removing 98 duplicate records, the remaining 1922 studies were screened based on titles and abstracts, and the remaining 1894 records were excluded. 72 of the 79 studies’ titles and abstracts were unrelated to the study’s objectives. The remaining seven studies were screened for inclusion in the analysis by critically reading their full-texts. Also, additional two eligible studies were included from the previous systematic review. In the end, a total of 9 studies were included in this meta-analysis.

• Download figure
• Open in new tab

Figure 1:

Flow diagram of included studies. For more information, visit: http://www.prisma-statement.org/

3.1 Characteristics of included studies

Table 1 shows the characteristics of the nine eligible studies included in this systematic review and meta-analysis. All of the included studies were carried out in five Sub-Saharan African countries: Kenya (n = 5) [44–48], Malawi (n = 1) [49], South Africa (n = 1) [50], Botswana (n = 1) [51], and Uganda (n = 1) [52]. The sample size of studies that are eligible ranges from 165 to 2472. All of the included studies took place between 2014 and 2021. The age range ranged from 16 to 29.2 years. The gestational age range is approximately 14 to 36 weeks. The interventions differed between studies. Two studies found that women sent up to 14 text messages during pregnancy and after giving birth. The message content and schedule were the same as in the previous study that demonstrated the efficacy of this intervention [46, 47]. Another study sent an automated SMS every four weeks from the second to the twelfth week [51], and one study assigned 3 to 6 SMS messages each week at a self-selected time of day and in their preferred language [44]. Kinuthia et al. assigned interventions such as 1-way SMS (participants received SMS but were unable to send messages to the study) and 2-way SMS (participants received SMS but were unable to send messages to the study) [45], to studies that did not specify the number of SMS reminders [49, 52]. Another intervention consists of two phone calls during the first week of starting PMTCT services, followed by one call per week until the participant delivers (a maximum of 26 calls), two calls during the first week after delivery, and one call per week for the next 14 weeks (a maximum of 16 calls) [48]. The risk of bias assessment of included 9 studies as showed in the Figure 2. Five studies reported a low risk of random sequence generation [44-46, 48, 51], two studies reported an unclear risk [49, 52], and two study reported a high risk [47, 50]. Two studies had a low risk of allocation concealment [45, 46], six studies had an unclear allocation concealment [44, 47–50, 52], and one study had a high risk [51]. Two studies [45, 51] had low risk of participant and personnel blinding, four studies were unblinded [46, 48, 50, 52], and three studies were unclear [44, 47, 49]. All of the included studies reported a low risk of outcome assessment blinding. Only one study reported a high risk of an incomplete outcome, while the other studies reported a low risk. Reporting bias was minimized in all studies, and four studies reported low risk of other bias [44–46, 48], while other studies reported high risk of other bias (recall bias, different in some baseline characteristics, short intervention duration, blunted the effect of the interventions on retention and null effect of the trial) [47, 49–52].

• Download figure
• Open in new tab

Figure 2:

risk of bias summary

View this table:

• View inline
• View popup

Table 1:

Characteristics of included studies in the review

3.3. Pooled odds and relative ratios estimate

We included nine studies to explore the impact of mobile phone communication in improving the couple HIV-infected pregnant women and children outcomes. The pooled RR estimate (RR = 1.02, 95%CI: 0.95 – 1.10, I² = 65.65%, P = 0.565) showed that the odds of antenatal visits in HIV-infected pregnant women receiving mobile phone interventions was not statistically significant compared to non-mobile phone group (Fig 3).

• Download figure
• Open in new tab

Figure 3:

Forest plot of ANC visits among HIV-infected pregnant women: mobile phone vs control groups

Pooled odds of Maternal post-partum retention was 2.20 times higher among mobile phone group (OR 2.20, 95%CI: 1.55 – 3.13, I² = 53.20%, P < 0.001) compared to HIV-infected pregnant women in the control group (Fig 4). Pooled odds of ART uptake in HIV-infected pregnant women assigned with mobile phone interventions was 1.5 times higher than in the control group (OR 1.5, 95%CI: 1.07-2.11, I² =0%, P = 0.020) (Fig 5). Evidence has shown statistically significant impact of mobile phone communication on HIV testing at 6 weeks and above among HIV exposed children (OR 1.89, 95%CI: 1.04 – 3. 48, I² =88.04%, P = 0.032) (Fig 6). In the outcome Infant HIV infection among HIV-infected pregnant women, we found no statistically significant difference between the mobile phone communication and control groups (OR 0.81, 95%CI: 0.14 – 4.88, I² = 62.18%, P = 0.821) (Fig 7). The meta-analysis of maternal ART adherence greater than 95% included only one study. The difference between the two groups was not statistically significant (OR 1.03, 95%CI: 0.07 – 1.50, P = 0.864) (Fig 8). Lastly, the pooled results of maternal viral load non-suppression included only one study. The difference between the mobile phone and control groups was not statistically significant (OR 0.88, 95%CI: 0.48 – 1.62, P = 0.692) (Fig 9).

• Download figure
• Open in new tab

Figure 4:

Forest plot of post-partum retention among HIV-infected pregnant women: mobile vs control groups

• Download figure
• Open in new tab

Figure 5:

Forest plot of ART uptake among HIV-infected pregnant women: mobile phone vs control groups.

• Download figure
• Open in new tab

Figure 6:

HIV testing at 6 weeks and above among HIV exposed children: mobile phone vs control groups.

• Download figure
• Open in new tab

Figure 7:

Infant HIV infection among HIV-infected pregnant women: mobile phone vs control groups.

• Download figure
• Open in new tab

Figure 8:

Maternal ART adherence >=95%: mobile phone vs control groups.

• Download figure
• Open in new tab

Figure 9:

Maternal viral load non-suppression: mobile phone vs control groups.

• Download figure
• Open in new tab

Figure 10:

Sensitivity analysis of HIV testing at 6 weeks in children

4. Discussion

We conducted a systematic review and meta-analysis to determine the impact of mobile phone on 6, 830 HIV-infected pregnant women and 4, 111 HIV-exposed children. Our review found that mobile phone had a minor impact on ANC visits among HIV-infected pregnant women (RR = 1.02, 95 percent CI: 0.95 – 1.10, 1, 778 participants, 5 studies) [45, 48–50, 52]. In comparison to our previous review, which found that mobile phone communication was ineffective in improving ANC visits (RR 1.66, 95%CI 1.02 to 2.70, 381 participants, 1 study) [33], this current review confirmed the previous findings on ANC visits. The pooled post-partum retention study, on the other hand, included four studies [46–48, 52]. Post-partum retention was found to be 2.20 times higher in the mobile phone group than in the control group (OR 2.20, 95%CI: 1.55 – 3.13, P 0.001). Our post-partum retention findings are highly statistically significant and may be preferable to other post-partum retention interventions. The trial discovered that integration of care can lead to better transitions between prevention of mother-to-child transmission (PMTCT) and ART, as measured by a higher percentage of ART-eligible women starting ART within 12 months of testing HIV-positive in ANC in integrated care clinics versus non-integrated care clinics (40% vs. 17%; OR 3.22; 95%CI: 1.81–5.72) [53]. Previous reviews have found that text messaging is effective in increasing adherence to ART [53–56]. Finitsis et al. [57] reported a pooled OR of 1.48 (1.09 to 2.01) on any HIV outcome, but objective and subjective outcomes were pooled across all types of intervention as long as they included some text messaging. Similarly, in a randomized controlled trial that examined the use of monetary incentives to increase postpartum ART retention, the effect of conditional cash transfers on retention in PMTCT at six weeks postpartum was 12.5 percent higher probability of retention at six weeks postpartum for women in the intervention arm compared to the control arm (81% vs. 72%; RR ratio: 1.11; 95%CI: 1.00–1.24) [56]. Another individual randomized control trial (IRT) discovered that task-shifting to nurses, as well as home visits by peer counsellors to defaulted patients, did not result in a change in post-partum retention in the first 12 months after ART initiation [58].

In terms of the pooled odds of ART uptake in HIV-infected pregnant women, we found that mobile phone increase ART uptake when compared to our previous review, which found no statistically significant difference [33]. Similarly, our evidence has shown a statistically significant impact of mobile phones on HIV testing at 6 weeks and above among HIV-exposed children, which was also confirmed in our previous review [33]. In comparison to the first review, the updated version included a quite large number of included studies. Furthermore, maternal ART adherence, viral load non-suppression, and infant HIV infection were not statistically significant between intervention and control groups, which could be attributed to the small number of studies included in those outcomes. Our review has several clinical implications. Our findings, in particular, highlight the importance of mobile phone communication as an effective intervention for improving outcomes in HIV-infected pregnant women and their children during the postnatal period. However, more research is needed in the prenatal period. Another advantage of this review is that HIV-infected pregnant women are known to be a vulnerable study population, with poor ART outcomes and being understudied. In comparison to the previous review [33], this review included a much larger number of HIV-infected pregnant women. Because of its low cost, mobile phone intervention is a “low-hanging fruit” that PMTCT programs around the world could easily adopt and spread more broadly [47].

These results should be considered in a context of several limitations in included studies. All studies were conducted in sub-Saharan Africa, and the findings are likely to be an application in other similar settings. Due to the wide variation in interventions, study designs, and outcome measures, other forests included only one study in the meta-analysis [59]. The risk of bias is another limitation. Only two studies out of nine had a low risk of allocation concealment [45, 46] and also, two studies had low risk of participant and personnel blinding [45, 51]. Inadequate allocation concealment and participants and personnel blinding are influential RCT methodological factors. This may impact on the review validity. By the way, the overall strength of the evidence has been rated as moderate. Other important limitation of our study include may include the lack of electricity for charging phones in rural areas. A study cited a lack of electricity as a potential barrier to an mHealth intervention [59, 60]. Mobile phone-based interventions pose an added burden on daily battery consumption and policy makers should be aware of this challenge, particularly in rural areas and informal settlements, where electricity scarcity may hinder equitable access to the benefits offered [59]. Our findings may not be generalizable to women who never attended the clinic or do not use phones, and that more research is needed to quantify smartphone penetration in this and other low-resource areas. While phone ownership and usage preferences may differ in other contexts, phone use is rapidly increasing, and we sought to investigate user behaviour in this context to inform future interventions. Despite the fact that cell phone coverage has increased dramatically across the world, 477 million people in Sub-Saharan Africa subscribed to mobile services, accounting for 45% of the population at the end of 2019 [61]. Thus, many women do not have access to cell phones and thus could not benefit mobile phone interventions. Lastly, a more direct measure of adherence, such as therapeutic drug monitoring or antenatal and postnatal viral load monitoring, should be included in future studies.

5. Conclusions

Our findings can help to inform evidence-based decisions about how to improve outcomes for HIV-infected pregnant women and their children. This study found that using a mobile phone improved post-partum retention, ART uptake, and HIV testing at 6 weeks and above in HIV-infected pregnant women. However, no statistically significant results were found for ANC visits, maternal ART adherence, viral load non-suppression, or infant HIV infection. Given the growing importance of elimination of mother-to-child transmission (eMTCT) and the HIV-related impact in pregnant women, our research is critical.

More research is needed to determine the effect of mobile phones on prenatal period, maternal ART adherence, viral load non-suppression, and infant HIV infection. In light of the bias, our evidence has been rated as moderate. Further, the implementation of these findings should also consider the various barriers to phone ownership and coverage in various settings.