Concise Notes on Evidence - Shared Copy
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

View only
Study, Author, YearEffectSample size (Control + Experiment)Study NameLocationExperiment typeInterventionBaseline immunization rateCosts Per Message Sent (USD)Other Costs in USD (name)Time of Reminder(s) in Days Before AppointmentControl Group with Health InformationTime Between Last Vaccination and Final Data CollectionStrength (1-3) KatherineMessage TextAlternative LinkSpam texts common in country?Number of dosesUpdate based on late vaccines?Notes
Schlumberger et al 201517.6% point increase523
Positive impact on the Expanded Program on Immunization when sending call-back SMS through a Computerized Immunization Register, Bobo Dioulasso (Burkina Faso)
Burkina-FasoRCTSMS reminder55.7%N/AN/AN/ANo
1 to 8 months (inferred from sign up dates and data collection date)
Translated notes
In French. Ask Katherine if have questions. It found ~17.5% point increase for 2nd, 3rd, and 4th vaccination. Followed up with those who received SMS but didn't get vaccinated. 66% of them had gone to different clinic, 19% couldn't be contacted, 18% were on a long voyage, 13% forgotten, 3% said message wasn't sent, 2% children passed away. ~20% of numbers were fake or disactivated over time. 96% of women understood messages even though 42% couldn't read it on screen themselves (unclear whether illiteracy or not owning phone themselves)
Haji et al 2016
13% point increase (or 8.45% point if take into account people who took child to different clinic)
Reducing routine vaccination dropout rates: evaluating two interventions in three Kenyan districts, 2014
RCT (cluster-randomized, 9 groups, 3 in SMS arm, 3 in sticker arm)
SMS reminder83%0.02
0.27 (full cost per child reminded up to 6 months)
2,0No0.5 months3N/AlinkNo3
Yes, but no follow-up reminder if never received second dose.
Compared text messages to stickers to control. This has a CEA; prior vaccination rates (and dropout rates) given for each district in trial; "Dropout was defined as any child who failed to return for the third dose of pentavalent vaccine [which takes place at 14 weeks] two weeks or more after the scheduled date."; majority (77%) unemployed; "If a care giver took the child to another facility for second or third pentavalent dose, the system considered the child unvaccinated,leading to misclassification, however, a sensitivity analysis that assumed that these children were actually vaccinated had no effect on the general observed difference between the inteventions." Only included regions with >10% dropout rates. Women without phones weren't included (0.9%). Followed up with people who didn't show up >2 weeks after appointment to find out why didn't come. Called and asked why didn't go, said child taken to another facility 39 (35 %); travelled out of town 33 (30 %); forgot 17 (15 %); child was sick 16 (15 %); or child died 2 (2 %)
Each district, and randomly allocated to send text messages or provide stickers reminding parents to bring their children for second and third dose of pentavalent vaccine, or to the control group
Eze et al 20158.7% point increase
1001 initial sample, 9.6% lost to follow up so only analyzed 905
Enhancing Routine Immunization Performance using Innovative Technology in an Urban Area of Nigeria.
SMS reminder, recall as well
0.15 (full cost per additional child immunized, estimate, originally in Naira 27.47)
1 (with additional messages 1 day before next appointment for anyone who missed initial appointment)
1 month pass reccommended DPT3 dosage (14 weeks) or 1.5 (?) months past third recommended vaccination (p6)
Example message: "Dear client, your child is due for his/her next dose of vaccines tomorrow Tuesday 20/7/10. Kindly bring your child to Hospital X for vaccination at 8am. Please come with immunization card. Thank you."
Includes CEA. "Some respondants" were initially randomized in control group but later swapped with members of the control group who did have phones. No figure is given for how many (but very likely less than 50, since 95% of people had phones). This could obviously result in selection bias.; Data collection and analysis was blinded to treatment/control. Acknowledged limitations: no known delivery success rate, 9.6% dropout rate, inconsistent DPT vaccine supply, no knowledge of literacy rates of participants, no adjustment for those lost to follow-up getting vaccinated anyway.
Bangure et al 2015
16.3% point increase (95% CI 12.5-28%)
Effectiveness of short message services reminder on childhood immunization programme in Kadoma, Zimbabwe - a randomized controlled trial, 2013
ZimbabweRCTSMS reminder75.0%0.08
0.99 (full cost per child up to 18 months, with 3 messages per visit, estimate), 0.33 ("Capturing of data per message", estimate)
Unclear. Minimum 17 days past day due because quartiles referenced )
The translated messages were as follows; A week before appointment date: − “Immunization protects your child against killer diseases such as polio, whooping cough, diphtheria, measles, pneumonia and tuberculosis. You are reminded that the vaccination appointment will be due in 7 days time from today.” Three days before appointment: − “You are reminded that the vaccination appointment will be due in 3 days from today.” A day before appointment: − “Your vaccination appointment is due tomorrow, visit the nearest clinic”.
Coverage estimated by following up with people over the phone, potential source of over-estimtation, 100% preferred single language which could have kept costs down; majority unemployed in control and intervention group. While the sample size is fairly small, Katherine did a power calculation and it would've need only half of the used sample size to detect the effect it did. 64% prefer SMS 1 day before appointment, 27% prefer 3 days before, practically nobody wants further out reminders. Nobody lost to followup which is pretty impressive.
Gibson 2017
4% point better (not stat sig, 95% CI 0.97-1.12 RR)
The readiness, need for, and effect of mhealth interventions to improve immunization timeliness and coverage in rural western Kenya
RCT (cluster-randomized, 4 clusters)
SMS reminder, SMS reminder and cash transfer
No (but intial sign up included sms with "The greatest wealth is health")
1 month to 3 months, inferred from timeline fig 3.1 and 9 month latest vaccination
"Tell Mama<Baby FName> that Penta-1 vaccine is due this week. <Motivational Message>" (chart of messages for all arms with motivational messages on pg 95-96)
Underpowered to detect an effect size of 4% at stat sig. Compared SMS only, SMS plus small cash transfer, SMS plus larger cash transfer, and control. Lots of detail about study logistics and specifics available; located in setting with detailed demographic information on births, deaths and pregancies. Some vaccine stock-outs in area (pg 69-70).; in 3 languages based on parental preference (English, Kiswahili or Dholuo) (74); program adjusted timing of reminders based on when child recieved previous vaccination (74); Was looking for full vaccination vs not full vaccination. 152 cluster (village) randomized controlled trial that enrolled caregivers of infants less than 35 days old who had not initiated the pentavalent vaccine series.
Domek 2016 (pilot, not finished yet)
Not statistically significant, visit 2 +4.9% at p = .12, visit 3 + 3.7 at p =.69.
SMS text message reminders to improve infant vaccination coverage in Guatemala: A pilot randomized controlled trial
GuatemalaRCTSMS reminder80.70%N/AN/A6,4,2No2 months2
"Your child [autopopulate child's name] is due on [autopopulate date] at [autopopulate clinic name] for vaccines."
"Not adequately powered to assess efficacy". If baseline vaccination rate is 80% it would have to have a sample size of 150 to detect 15% point change with 80% probability. To detect a 8% point increase, the lowest increase of the good quality studies, it would've had to have a 656 sample size. The baseline rate was at a pretty high starting point, so is likely to have a smaller effect because the people still not getting vaccinated might have something else stopping them. To participate at least one parent had to be "literate and able to use SMS technology", and be 18 or older.; No data on messages being delivered. SMS system problems inlcuding "power outages and delays in recharging the server that either resulted in missed messages or repeat messages being delivered upon reactivation of the delivery platform."; Usual care included a written reminder of next appointment date at each clinic. Parents of infants between the ages of 8 and 14 weeks presenting for the first dose of the 3-dose infant primary immunization series were eligible if they owned a mobile phone with SMS text messaging capability. Last relevant vaccine is at 6 months old. 17.4% people lost to follow-up.
Uddin et al 201613.1%-30.5% point increase (DID)4158
Use of mobile phones for improving vaccination coverage among children living in rural hard-to-reach areas and urban streets of Bangladesh
Quasi-experimental, pre-post (clustered, 160 clusters)
Mobile registration and SMS reminder and remind health workers and provide info to health worker supervisors
1, 0 (opening of clinic), 0 (2 hours before closing of clinic)
No1 month1N/AUnknownYes
Differences in differences; very, very weak; used original software for CHW; sampled different street children at endline and baseline and didn't restrict to just those registered with original software; "Upon registration, mothers were assigned a unique code and taught how to send SMS text messages from a regular mobile phone to mTika after childbirth". Included children 0-11 months old at beginning of study. Sent reminders to mothers and health workers, as well as monitoring for supervisors of health workers. Symbols taught to mothers at registration were used in SMS messages for mothers who were illiterate. Did based on maternal recall
Kazi et al 2018
5.3% point increase, not stat sig but very underpowered
Effect of Mobile Phone Text Messages Reminders on Uptake of Routine Immunization among Children in Pakistan-Study Protocol for a Randomized Clinical Trial
PakistanRCTSMS reminder59%N/AN/A4 times in week child was due
One-time standard verbal counseling at the time of initial visit (enrollment) by the study staff regarding the timing for EPI vaccines at 6, 10, and 14 weeks
1 month2
"Your baby is due for [6, 10, 14] week vaccination [line break?] immediately take your child to the nearest EPI center" (translation from Urdu and Sindhi)
Kazi AM et al 2016 (pre-register)
Study was powered to detect an effect size of 20%, higher than any previously done study. For how we came to the effect size of 5.3% look at the sheet "Caclulations for Kazi". Did this because there wasn't a convenient summary statistic for the entire study.