Table 2.
Characteristics of the intervention studies included in the review and the relevant findings.
| Study/ (Focus of evaluation) | Study design | Target population or size (Setting) | Form of mHealth | mHealth function | Relevant findings |
| 2014, Munro et al [28] (Maternal) |
Nonrandomized design (preintervention and postintervention evaluation) | 99 TBAsa trained; 63 retained 1-year posttraining for complete evaluation (Rural Bong county in Liberia) |
Utilization of mobile phone functions, coded SMS text messaging | Data collection and transmission (locating pregnant women; take data on age and referring them for antenatal care) | Participants demonstrated an increase in the mean number of skills that they were able to perform between pretest and both the immediate posttest and 1-year posttest. Mean number of skills that the participants were able to complete did decrease slightly between immediate and posttest. Individual skills verified a significant retention of knowledge between the pretest and posttest. Many TBAs continued to have trouble with the more complex skills of adding credit to a mobile phone (11% was able to do this at posttest). 70% of TBAs relied primarily on others with higher education to assist them with mobile functions; 24% used their phone to communicate with the certified midwife; 14.3 % wanted to communicate, but had poor reception. |
| 2014, Pathfinder (Grey Literature) [30] (Maternal) |
Nonrandomized design (preintervention and postintervention evaluation) | 258 participants in an infant feeding health education program (Abuja and Nasawara in Nigeria) |
150 CHWc in 10 primary health clinics | Education and communication (ANCd protocols, and client follow-up) | CHWs increased HIVe testing from 68% to 82%. Blood pressure measurement increased from 87% to 97%. The quality of care score from client interviews increased from 13.33 at baseline to 17.15 at end line, with the most significant improvements related to health counseling. Individual and group health counseling sessions became structured. |
| 2013, Little et al [26] (Maternal) |
Nonrandomized design (preintervention and postintervention evaluation | 20 HEWf, 12 midwives, 5 supervisors (Kilte and Awelalo districts in the Tigray region of Ethiopia) | Mobile phone app using open source components | Data collection (using appropriate technologies to meet needs of HEW and midwives | GPRSg connection was available in 35 health posts and centers (74%) of the study districts. There were very few instances of the mobile data network being unavailable for a substantial period of time. 34 of the 36 phones were retained; 2 were stolen with one later recovered; 3 phones had issues with insensitive screens and were replaced. Most health workers rapidly learned how to use and became comfortable with the touch screen devices so only limited technical support was needed. |
| 2012, Zhang et al [22] (Maternal and Neonatal) |
Randomized controlled trial | 10 students of the Hebei Union School of Public Health (Zhaozhou Township, Hebei Province, China) |
Mobile phone data collection | Data collection (Use of mobile phones for data collection on infant feeding practices compared with use of pen and paper) | In 120 copies of pen-and-paper questionnaires, 55 questionnaires contained errors. 65% of the pen-and-paper records did not match and needed to be checked. There was no significant difference between duration of pen-and-paper method versus mobile phone method. The mean cost per questionnaire was higher for the mobile phone questionnaire (US $23) than for the pen-and-paper questionnaire (US $13). The mobile phone method was acceptable to interviewers, with only minor problems that did not result in data loss. |
| 2012, Lori et al [27] (Maternal) |
Nonrandomized design (preintervention and postintervention evaluation) | 99 TBAs (Rural Liberia) |
SMS text messaging | Data collection (using a pregnancy reporting protocol) | Mean increase in mobile phone knowledge scores was 3.67 (95% CI 3.39-3.95). Data collectors also demonstrated a significant increase in their ability to perform each individual mobile phone task. Participants with a mobile phone in the family did significantly better on 3 of the 7 tasks in pretest. |
| 2012, Seidenberg et al [31] (Infant) |
Nonrandomized design (preintervention and postintervention evaluation) | At least 2 health workers from each facility (2 districts in Southern Zambia) |
SMS text messaging | Data collection and transmission (to reduce the time between blood sampling for the detection of infant HIV infection and notification of the test results to the relevant point-of-care health facility by using SMS-based system | Mean turnaround time for delivery of a test result to the relevant health facility fell from 44.2 days (SDh:28) preimplementation to 26.7 days (SD:31.8) postimplementation. Mean turnaround times for delivery of a test result to a caregiver of the tested infant were 66.8 (SD: 38.8) preimplementation and 35 days (SD: 31.2) postimplementation. |
| 2012, Lemay et al [25] (Maternal) |
Nonrandomized controlled trial (staged design) | Health surveillance assistants and community health workers. 95 SMS users in Salima. 95 nonusers in Salima. 95 nonusers in Kasungu (Salima, Nkhotakota, and Kasungu Districts in Malawi) |
SMS text messaging | Communication (reducing communication gaps between health workers and their district teams; increasing access to information and improve quality of services) | SMS used to report stock-outs, asking general information, reporting emergencies, confirming meetings, and requesting technical support. Among respondents who received phones, the most common form of communication was SMS (100%), phone calls (94%), public transport to travel (8%). Among respondents who did not receive phones: 92% used transport and only 6% used phone calls. None used SMS for communication. SMS participants needed an average of 9 minutes to report issues and receive feedback at an average of USD 0.61$. Health workers with no access to SMS spend an average of 1445 minutes (24 hours) to report and receive feedback at an average of USD 2.70$. In control district, it took 1681 minutes to report and receive feedback at an average of USD 4.56$. |
| 2012, Ngabo et al [29] (Maternal) |
Nonrandomized design (pre intervention and postintervention evaluation) | 432 community health workers and the rest of the health system (Musanze, Rwanda) | SMS Text messaging (Rapid SMS-MCHi system) | Communication (SMS-based platform, enabling effective and real time 2-way communication for action, between CHWs at community level, and the rest of the health system. Used to improve access to antenatal, postnatal care, institutional delivery, and emergency obstetric care) | 5734 SMS were sent. 11,502 pregnancies (81% of the 14,200 estimated annual pregnancies in district) were monitored. A 20% increase in facility-based delivery from 72% 12 months before to 92% at the end of pilot phase. |
| 2011, Andreatta et al [23] (Maternal) |
Nonrandomized design (posttraining evaluation) | 8 TBA, 2 professional nurse midwives (Sene District in Ghana) | SMS text messaging | Data collection and communication (reporting postpartum hemorrhage occurrence, management, and outcome) | Both professionals and TBAs were able to use the specified reporting and text messaging protocols to report clinical outcomes. 425 births were reported during study period, with 13 cases of PPHjoccurring (3.1%) cases. |
| 2010, Kwaewkungwal et al [24] (Maternal and child, including neonatal) |
Nonrandomized design (preintervention and postintervention evaluation) | Health workers in charge of ANC or EPIk services (sample size not indicated in paper) (Phung district, Thai-Myanmar) | SMS text messaging and Web-based apps | Data collection, automated generation of list, and update information regarding the antenatal care and child's immunization status on mobile phone when performing ANC or EPI activities off health care clinic | 59% come on time as per scheduled dates after implementation compared with 44% before implementation. 44% of children who came to receive scheduled vaccines on time on the preset monthly immunization date after intervention compared with 35% before. Updating immunization data on mobile phone increased odds of EPI on time by 2.04. |
aTBA: traditional birth attendant.
bSMS: short message service.
cCHW: community health worker.
dANC: antenatal care visit.
eHIV: human immunodeficiency virus.
fHEW: health extension worker.
gGPRS: general packet radio service.
hSD: standard deviation.
iMCH: maternal and child health.
jPPH: postpartum hemorrhage.
kEPI: expanded program on immunization.