Reducing routine vaccination dropout rates: Evaluating two interventions in three Kenyan districts, 2014

Background: Globally, vaccine preventable diseases are responsible for nearly 20 % of deaths annually among children <5 years old. Worldwide, many children dropout from the vaccination program, are vaccinated late, or incompletely vaccinated. We evaluated the impact of text messaging and sticker reminders to reduce dropouts from the vaccination program. Methods: The evaluation was conducted in three selected districts in Kenya: Machakos, Langata and Njoro. Three health facilities were selected in 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 (routine reminder) with next appointment date indicated on the well-child booklet. Children aged <12 months presenting for their first dose of pentavalent vaccine were enrolled. A dropout was defined as not returning for vaccination ≥2 weeks after scheduled date for third dose of pentavalent vaccine. We calculated dropout rate as a percentage of the difference between first and third pentavalent dose. Results: We enrolled 1,116 children; 372 in each intervention and 372 controls between February and October 2014. Median age was 45 days old (range: 31-99 days), and 574 (51 %) were male. There were 136 (12 %) dropouts. Thirteen (4 %) children dropped out among those who received text messages, 60 (16 %) among who received sticker reminders, and 63 (17 %) among the controls. Having a caregiver with below secondary education [Odds Ratio (OR) 1.8, 95 % Confidence Interval (CI) 1.1-3.2], and residing >5 km from health facility (OR 1.6, CI 1.0-2.7) were associated with higher odds of dropping out. Those who received text messages were less likely to drop out compared to controls (OR 0.2, CI 0.04-0.8). There was no statistical difference between those who received stickers and controls (OR 0.9, CI 0.5-1.6). Conclusion: Text message reminders can reduce vaccination dropout rates in Kenya. We recommend the extended implementation of text message reminders in routine vaccination services. We conducted an evaluation study in three selected districts in Kenya. District selection was based on pentavalent vaccine coverage for 2012. Districts with more than 10 % dropout rates for the third pentavalent dose, which is considered above acceptable limits in the expanded programme on immunization, were considered for inclusion in the study. In 2012, 34 districts with dropout rates more than 10 % were identified. Among these, districts with very high coverage rates (third dose pentavalent coverage ≥90 %) were excluded, as were districts that were geographically hard-to-reach or with security concerns. Six districts with dropout rates ranging from 13-27 % with both rural and urban settings were identified, and subjected to simple random sampling to select three districts. These districts included Machakos, Langata and Njoro (Fig. 1). Map of Kenya showing study sites Machakos District is in Machakos County, covers an area of 6,281.4 kmsq, most of which is semi-arid with a projected population of 211,404 from 2009 census [26] and 13,271 live births in 2012. There are 56 health facilities, on average the distance to health facilities is 5 km, in 2012 the district achieved coverage of 88 % for the first dose of pentavalent, 79 % for the third dose, and 76 % full vaccination coverage. The dropout rate among children for the third dose of pentavalent vaccine was 13 % . Langata district is one of the districts in Nairobi County, the district is an area with glaring contrast in living standards, ranging from the plush homes of Karen and Langata to the sprawling Kibera slums, which are characterized by poor living standards with a projected population of 397,238 from 2009 census [26], live births of 12,402 and 97 Health facilities. In 2012 the district achieved coverage of 96 % for the first dose of pentavalent, 84 % for the third dose, and 82 % full vaccination coverage. The dropout rate among children for the third dose of pentavalent vaccine was 13 %. Njoro district is in Nakuru county, had an estimated population 100,000,7,904 [26], live births and 37 health facilities. The main economic activity agribusiness with a large proportion of the population having to travel for more than five kilometres to access the nearest health facility. In 2012, the district achieved coverage of 86 % for the first dose of pentavalent, 75 % for the third dose, and 74 % full vaccination coverage. The dropout rate among children for the third dose of pentavalent vaccine was 13 %. Vaccination outreach services are conducted in all the three district to reach hard to reach areas and engagement of mass media to pass information to promote timely vaccination of young children. Children <12 months of age who were brought to the selected vaccinating health facilities in the three districts for their first dose of pentavalent vaccine were recruited on a first come basis until the strategy-level target sample sizes was reached. Children whose mothers did not have a telephone number were excluded from the study. Dropout was defined as any child who failed to return for the third dose of pentavalent vaccine two weeks or more after the scheduled date. We selected three health facilities in each district, and randomly allocated each facility to one of the two interventions to provide short text messages or stickers reminding caretakers to return for second and third dose of pentavalent vaccines, or to serve as the control group, receiving no extra reminder messages and continue providing the next appointment date in the well-child booklet. Participants were conveniently enrolled in the selected health facilities until the strategy-level target sample sizes were reached. Caretakers of participants in the SMS intervention group received two text reminders via SMS. Reminders were dispatched from an automated web based system two days before and on the day of the scheduled vaccination due date for the second and third dose of pentavalent vaccine. The first message reminded the parent of the next due date for the vaccination and which health facility to attend for vaccination. The second message reminded the caretakers that the actual due date was that day. The text messages were sent in Kiswahili and English. The sticker intervention group received two stickers at the time of enrollment which noted the day of the scheduled vaccination due date and the name of the health facility. Caretakers were instructed to place one sticker on the child’s health booklet, and the other sticker in a visible area of the main household or within the bedroom. Placement of the sticker within the home was verified during subsequent visits by asking the parent where they placed the sticker. The control group received no reminders, but the scheduled vaccination due date was indicated on the child’s health booklet as per routine procedures. All the groups received routine health education and advice on vaccination. Any caretaker who failed to return the child for vaccinations two weeks or more after the expected completion of third pentavalent dose was contacted by the investigator to establish reasons for missed vaccinations. Data were collected by study nurse and principal investigator during routine working hours at the maternal child health clinic on a daily basis. Caretakers were interviewed face to face using a pretested standard questionnaire. The questionnaire collected information on socio-demographic, knowledge and source of information on vaccination, and recorded details of vaccines received during each visit. Data were entered and analyzed using Epi info software. The primary outcome measure was receipt of scheduled vaccines at 10 and 14 weeks. The secondary outcome measures were dropout in vaccination and factors associated with missed vaccinations. We conducted data analysis using Epi Info version 7.1.4 and excel analysis software. Proportions and means were calculated for categorical and continuous variables respectively and summarized into tables and figures for univariate analysis. Bivariate and Multivariate analysis using unconditional logistic regression using facility clusters were conducted to identify independent predictors of missed vaccinations. Odds and Adjusted Odds Ratio (OR & AOR) and 95 % Confidence Interval (CI) were used to estimate the strength of association between independent variables and the dependent variable. The threshold for statistical significance was set at p < 0.05. We calculated dropout rate as a percentage of the difference between first and third pentavalent dose. Sample size calculation was done using Casagrande et al. [27] formula for comparing two proportions to detect a 15 % decrease in the drop-out vaccination rate for each of the three intervention groups, assuming a dropout rate for the third dose of pentavalent of 15.6 % [28], study power of 80 %, and confidence level of 95 %. The minimum sample size was 372 participants per intervention arm.