Using Short Message Service Alerts to Increase Antenatal Care and Malaria Prevention: Findings from Implementation Research Pilot in Guinea

Abstract.

Intermittent preventive treatment with sulfadoxine–pyrimethamine (SP) is recommended to prevent malaria in pregnancy. Treatment coverage, particularly for three or more doses, is dependent on pregnant women attending antenatal care (ANC) services as scheduled. The StopPalu project pilot tested short message services (SMSs) to remind women of upcoming ANC visits in the Conakry and Kindia regions of Guinea. Health facilities were selected as pilot and comparison facilities. All women who attended an initial ANC visit at a selected facility during the pilot period and had access to a mobile telephone were enrolled. The pilot group was sent an SMS before each appointment. Percentage of attendance and SP distribution were calculated. A log-binomial regression model determined odds ratios. Pregnant women receiving SMS were 48 times more likely to attend all visits and were 12 times more likely to receive all SP doses during pregnancy.

The WHO recommends four antenatal care (ANC) visits to prevent maternal, prenatal, and neonatal deaths and pregnancy complications, where women receive diagnosis and management of complications, preventive treatment for malaria, information to identify and manage infection, information to promote the use of skilled attendance at birth, and education on postnatal healthy behaviors.¹ Preventive treatment for malaria in pregnant women is of particular importance in Guinea, where 100% of the population is at risk of the disease,² and where national malaria parasitemia prevalence is 15%.³

Malaria infection in pregnancy is associated with maternal anemia, low birth weight, and perinatal mortality.^4,5 The 2012 WHO policy brief recommends reducing malaria during pregnancy in moderate and high malaria transmission areas in Africa through intermittent preventive treatment in pregnancy (IPTp).⁶ The recommended approach involves providing sulfadoxine–pyrimethamine (SP) at every scheduled ANC visit after the first trimester. IPTp with SP clears or suppresses existing malaria infections in the mother’s placental and peripheral blood and provides a prophylactic effect that prevents new infections for several weeks.⁷ As a clear entry point for the implementation of IPTp, it is crucial for malaria control efforts to encourage and facilitate ANC attendance.

It is estimated that 84.6% of pregnant women in Guinea attend a minimum of one ANC visit, whereas only 50.6% attend the recommended four visits.³ This suggests that half of the pregnant women in Guinea are not fully benefiting from IPTp. We speculated that one of the likely reasons for low attendance is the lack of awareness on the part of pregnant women for follow-up visits, which could be addressed through systematically reminding women of their upcoming ANC visits.

The efficacy of short message service (SMS) in reminding patients about upcoming appointments has been well established in areas where mobile technology is readily available.^8–16 Although the literature suggests there is sufficient evidence to promote the use of SMSs to remind patients of appointments, most of these studies were conducted in resource-rich areas. It is unclear whether the same benefits are seen in resource-constrained environments as well.

We report the findings of an implementation research study that pilot tested the use of SMSs as a notification system in the Conakry and Kindia regions of Guinea to remind women of upcoming ANC visits. The aim of the study was to determine whether pregnant women who received SMS reminders were more likely to attend all four ANC visits and whether they were more likely to receive at least three doses of SP via IPTp (IPTp3), compared with women who did not receive SMS reminders.

The study was conducted within a larger operational program, StopPalu, supporting the Guinean Ministry of Health’s National Malaria Control Program. StopPalu was implemented in 19 of Guinea’s 38 districts and was a program led by RTI International and funded by the U.S. Agency for International Development under the U.S. President’s Malaria Initiative. Eight health facilities within the Conakry region and seven health facilities within the Kindia region were randomly selected as pilot facilities for the study. Eight facilities in Conakry and six facilities in Kindia were purposefully selected as comparison health facilities; selection was performed through an anecdotal evaluation of the facilities in discussion with StopPalu staff to mimic a similar sociodemographic and programmatic context to the pilot facilities.

All pregnant women who attended an initial ANC visit at one of the study’s selected health facilities from March 27 to May 2, 2017, on a weekday (i.e., Monday–Friday) and had access to a mobile telephone were enrolled in this evaluation. Access to a mobile phone was defined as the ability to use either her own or a family member’s mobile telephone. Participant mobile telephone numbers were entered into the open-source SMS notification platform, RapidPro (https://community.rapidpro.io/). All women in the pilot group were sent an SMS reminder to their mobile telephone one day before the next scheduled appointment. The reminder (translated to English) read, “You need to go to the health facility tomorrow for your prenatal care visit.” De-identified attendance status and distribution of IPTp for each participant were collected from the health facilities via SMS and entered into the USAID StopPalu project database. The comparison group data were collected from the health facility record.

Percentage of ANC attendance along with IPTp distribution was calculated for the pilot and comparison groups. A t-test was conducted to determine significance of difference between the two groups. A log-binomial regression model was then run to determine odds ratios, where the primary outcome of attendance at all four visits is binary. All analyses were carried out in R version 3.0.1 (R: A Language Environment for Statistical Computing, Vienna, Austria).

RTI’s Institutional Review Board determined that this project was not research with human subjects as defined by the U.S. Code of Federal Regulations (45 CFR 46.102)—specifically these activities would not be considered “research” as defined by that code. Written consent was obtained from participants.

Between March 27 and May 2, 2017, 785 pregnant women were enrolled in the study (541 pregnant women in pilot facilities and 244 pregnant women in comparison facilities). The analysis demonstrates that although attendance at the second and third ANC visits is high in both the pilot and comparison groups, the pilot group performed better than the comparison group for all visits (P-value < 0.001) (Table 1, Figure 1). The pilot group also maintained consistently high attendance, with 98% attendance at all four visits, whereas attendance for the comparison group dropped with each follow-up time point. Attendance at all four visits among the comparison group was 52.5%, which is 45.7% lower than that of the pilot group (P-value < 0.001). The data show that IPTp was consistently delivered during ANC visits among both groups (range: 94.4–100%). Low attendance after the second ANC visit among the comparison group resulted in overall low IPTp delivery in this group.

Table 1.

Outputs from t-test determining difference in percent of women who attended each ANC visit and received IPTp

ANC visits	Attendance				Received IPTp at visit*		Received IPTp†
	Pilot (n = 541)%	Comparison (n = 244)%	Difference %	P-value	Pilot %	Comparison %	Pilot (n = 541)%	Comparison (n = 244)%	Difference %	P-value
ANC2	99.1	94.3	4.8‡	< 0.001	95.3	100	94.5	94.3	0.2	0.9138
ANC3	99.6	81.6	18.0‡	< 0.001	98.3	94.9	97.6	77.5	20.1‡	< 0.001
ANC4	99.4	52.9	46.5‡	< 0.001	99.4	96.1	98.9	52.5	46.4‡	< 0.001
All	98.2	52.5	45.7‡	< 0.001	94.4	96.9	92.6	50.8	41.8‡	< 0.001

ANC = antenatal care; IPTp = intermittent preventive treatment.

* These values are calculated as received IPTp/attended.

† These values are calculated as received IPTp/enrolled.

‡ Denotes statistically significant difference at the 95% level.

Figure 1.

Percent of antenatal care attendance and intermittent preventive treatment receipt among the pilot and comparison groups.

A regression model with attendance at all four ANC visits as the dependent variable and pilot/comparison as the independent variable estimates that women who received SMS reminders were 48 times more likely to attend all ANC visits than those who did not receive reminders (Table 2). A regression model with a receipt of at least three doses of IPTp (IPTp3) as the dependent variable and pilot/comparison as the independent variable estimates that women who received SMS reminders were 12 times more likely to receive all SP doses during their pregnancy, compared with women who did not receive SMS reminders. A third model was run showing that women who received SMS reminders were three times more likely to receive IPTp2 during their pregnancy.

Table 2.

Outputs from regression models

Model	Variable	Odds ratio [95% CIs]	P-value
Attended all antenatal care visits	Pilot	48.1 [25.7–100.4]	< 0.001
Received all IPTp3	Pilot	12.1 [8.1–18.4]	< 0.001
Received IPtp2 (i.e., dosed at the 3rd visit)	Pilot	3.1 [1.2–8.0]	0.015

IPTp = intermittent preventive treatment.

According to our findings, during every ANC visit, nearly all women who attended were administered SP in both the pilot and comparison groups. Clearly, distribution of SP during ANC visits is a priority in programmatic settings such as Guinea, and so increasing attendance rates should provide important malaria protection among this vulnerable population.

Among both the pilot and comparison groups, attendance at the second ANC visit was high (range: 94.3–99.1%), with more than 94% of women receiving at least one dose of SP. However, attendance after the second ANC visit declined significantly in the comparison group—this resulted in lower IPTp coverage among pregnant women sampled in comparison facilities, with SP distribution being 20.1% and 46.4% less in the comparison group than in the pilot group at the third and fourth ANC visits (P-value < 0.001). The low attendance at the fourth scheduled ANC visit (ANC4) seen in the comparison group is not surprising; the 52.5% ANC4 coverage seen in the comparison group is comparable to the 2016 Multiple Indicator Cluster Survey estimate of 50.6% ANC4 coverage.³ Impressively, the pilot group maintained high attendance throughout the evaluation period; women in this group were three times more likely to receive IPTp2 and 12 times more likely to receive IPTp3, when compared with women in the comparison group.

Although our findings demonstrate that the use of SMS should certainly be considered in program planning, it is important to acknowledge that women not having access to mobile phones would not benefit from this intervention. More innovative interventions should be considered for reaching this section of the population. Our study has a number of caveats. First, pilot and comparison facilities were not randomized: we intentionally selected health facilities of similar socioeconomic and programmatic contexts. It is possible that the significant difference in attendance behavior among the two groups (45.7% difference in ANC4 attendance) may be attributed to more factors than just the SMS reminder. Second, pregnant women sampled from study health facilities may have inherent differences, such as gestational age and parity, and our analyses did not control for potential confounders such as gestation age or parity at the first ANC visit. Third, there may be inherent socioeconomic differences between women in the pilot and comparison groups that could confound the results. Future studies should include additional covariates in the analysis and track gestational age and parity of the participants at their first recorded ANC visit.

Our findings show the relatively minimal investment of resources required for implementing SMS notification systems is well worth the potentially beneficial outcomes of instituting such a system for ANC visits; clearly, the implementation of the approach at large scale would require additional thought and investments, including staff training, consistent management, national buy-in by all stakeholders, individual opt-out options, and mobile network operator agreements. Key lessons learned from this study are as follows: 1) weekly supervision is needed at the beginning of implementation to ensure quality data recording, 2) women with access to mobile phones are extremely responsive to SMS reminders, and 3) women who attend ANC visits nearly always receive IPTp. We found that even in a resource-limited setting such as Guinea, those who have mobile phones reacted to the reminders and were significantly more likely to attend ANC services and receive life-saving IPTp during pregnancy.

Disclaimer: The opinions expressed herein are those of the authors and do not necessarily reflect the views of the employing organizations, USAID, or the U.S. government.