Effect of tailored information of vaccination schedule on vaccine uptake in northern Nigeria

ABSTRACT

Introduction: A vaccination schedule is complex and dynamic that requires repetitive and timely clinic visits by children and their caretakers. This study investigates whether providing caregivers with one-time tailored information on the vaccination schedule improves the vaccine uptake among children.

Methods: The study participants were 534 women each with a child age 8 months or less; they were from 11 settlements in Jada, a local government area in Adamawa state, Nigeria. The study was conducted on September 2019 to June 2020. Women were randomly selected to be assigned to a treatment group to be provided with one-time tailored information on their children’s current vaccination status and the next schedule for vaccination, while women in the control group were provided with generic information on the vaccination schedule. We employed the ordinary least squares (OLS) and logistic regression depending on the type of dependent variables to analyze the treatment effect.

Results: After the women received tailored information on the vaccination schedule, the number of clinic visits for vaccination increased. However, the tailored information did not improve the vaccine uptake among children at a particular age nor the full vaccination rate.

Conclusion: One-time tailored information has an immediate but no sustainable effect. It might be important for women to be constantly reminded about the vaccination schedule.

KEYWORDS:

• Tailored information
• vaccination schedule
• vaccine uptake
• Nigeria

Introduction

Despite the proven benefit of vaccines, the child vaccination rate has been stagnant worldwide,¹ including in the African region.² Apart from African, some other countries in Asia, Latin America and Europe also have similar trend.^3–5 Moreover, the vaccination rate in African countries lags behind other regions in the world.⁶ For example, the global coverage of three doses of diphtheria–tetanus–pertussis (DPT3) has been stagnant at 85% since 2015, while the DPT3 coverage in Africa remains at 72% since 2010.⁷ The distribution of unvaccinated children is unequal; out of about 20 million children who are not fully vaccinated, more than 20% reside in three countries, including Nigeria, the study site. Furthermore, Nigeria is one of the few countries where the coverage of DPT3 coverage has decreased over time: from 54% in 2010 to 42% in 2017.⁷

Child vaccination is not a one-time behavior; rather, it is a repeated behavior within a specific time frame. For example, in Nigeria, children are supposed to receive nine different types of vaccines at five different times within the first year since birth.⁸ Each type and dose of vaccine are important in terms of the type of preventable disease and the achievable efficacy level. For example, the oral poliovirus vaccine (OPV) is to protect children from contracting polio, a highly contagious disease. At present, Nigeria is one of the three countries in the world that have not eliminated polio.⁹ Children are scheduled to receive OPV four times: at birth and at 6, 10, and 14 weeks. The efficacy of the first dose of OPV is 82%,¹⁰ while the second and third doses are efficacious at 90% and 99% or more, respectively.¹¹ However, the vaccine schedule changes over time. For example, the new vaccine, inactivated polio vaccine (IPV), was introduced in 2015 in Nigeria; it is to be received at 14 weeks after birth. Rotavirus vaccine and meningococcal A vaccine were also introduced in 2019.

African countries suffer not only from the low vaccination coverage but also from the high dropout rate and the delayed vaccination uptake.^12–14 Although it is difficult to evaluate the objective reasons of this suboptimal vaccination coverage, one of the main reasons could be the complex vaccination schedule. When the Nigeria Multiple Indicator Cluster Survey (MICS) conducted in 2016/2017¹⁵ asked the caregivers of children with incomplete vaccination records why their children were not completely vaccinated, the common response was that the caregivers thought the children had already been fully immunized [44.8%). The caregivers’ misconception of “complete vaccination” can be attributed to the complicated vaccination schedule.

The literature identifies other potential barriers to complete vaccination. For example, 16,conducted a systematic review and found that there were multiple barriers to vaccination in African countries, such as the lack of knowledge of vaccination, the distance to the health facility, financial hardship, the lack of spousal support, and distrust in vaccines or vaccination programs. One limitation of Bangura et al.’s study was its inability to identify the causality of potential barriers to the vaccination, as most studies identified and reviewed in their work were based on observational data.

Among the various potential barriers to vaccination, our study focused on one – the lack of information on the vaccination schedule – and investigated whether providing caregivers one-time tailored information on the vaccination schedule improves the vaccine uptake among children. This study is important in causally identifying if caretakers of children can be provided with the tailored information to improve their vaccination behavior.

Methods

Study site and population

The study was conducted in Jada, a local government area (LGA] in Nigeria’s Adamawa state, where the full vaccination rate was low: 29% (National Immunization Coverage Survey (NICS], 2018). We partnered with Adamawa State Primary Healthcare Development Agency to obtain the logistic support. Within Jada LGA, we employed convenience sampling to identify 11 settlements under the catchment area of a health clinic, Jada I PHCC [Primary Health Care Clinic), for our study site. We selected large settlements to enable us to find as many eligible respondents as possible. In each settlement, one woman from each household was selected. Our interviewers visited all the households in each settlement to identify the eligible women. A mother who had a child age 8 months or younger at the time of the baseline survey was eligible to participate in the study. If one household had more than one mother with such children, interviewers randomly selected one per household.

Sample size

Our initial target sample size was 400 women. According to the power calculation, this sample size can detect the standardized effect size of 0.28 with power 0.8 and significance level 0.05. Assuming conservatively that the standard error is 0.5, we expected the treatment effect to be a 14-percentage-point increase in the vaccine uptake. Our minimum detectable treatment effect of 14 percentage points was reasonable. The meta-analysis conducted by 17, found that educational interventions increased the DPT3 vaccination rate by 18 percentage points in middle- and low-income countries.

Procedures

Baseline survey

After we identified eligible women, our interviewers administered a baseline survey to all eligible, consenting women at their households in September 2019. The baseline survey collected quantitative data relating to vaccination, including knowledge, belief, and attitude toward vaccination as well as sociodemographic characteristics.

Intervention

Following the baseline survey, approximately 40% of the sample (198 of the 515 women] were randomly selected for treatment. Women in the treatment arm were provided with one-time tailored information on their children’s current vaccination status and the schedule for the next vaccination visit. In particular, our trained interviewers first examined the vaccination cards of the children of women in the treatment arm. Then, if they identified any missed vaccinations in the vaccination cards, they informed the women and told them the type of vaccinations that were missed. Then, the women were informed and encouraged to bring their children to the health facility as soon as possible to receive the vaccines that they had missed. If the children of the women in the treatment arm did not miss any vaccination, they were informed about the next vaccination date. Interviewers wrote down the next vaccination date in the vaccination card and gave it to the women. On the other hand, women in the control arm were provided general information on the vaccination schedule verbally regardless of the children’s vaccination status. Women in the control group were not given tailored information, such as whether their children missed any vaccination in the past and when the next vaccination was scheduled. The intervention took place at the respondents’ households. Randomization was done by having the respondents pick paper strips from a bag. If a woman picked a marked strip, then she was assigned to a treatment group. Neither interviewers nor respondents could see inside the bag. An online appendix presents the vaccination sheet we used to explain the vaccination schedule to the women in the treatment arm. We gave this sheet to the women after the information provision.

Follow-up on the vaccination records

We recorded the children’s vaccination records at Jada I PHCC for nine months based on the baseline survey from September 2019 to June 2020. These vaccination records were later merged with the baseline survey information through the unique identifier assigned by our study.

Data analysis

To causally evaluate the effect of information provision on the vaccine uptake, we employed the following regression framework:

(1) ${{\mathrm{y}}_{\mathrm{i}}}_{\mathrm{j}}=\propto +{\beta }_1\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{m}\mathrm{e}\mathrm{n}{\mathrm{t}}_{\mathrm{i}\mathrm{j}}+{\mathrm{v}}_{\mathrm{j}}+{\epsilon }_{\mathrm{i}\mathrm{j}}$(1)

where $\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{m}\mathrm{e}\mathrm{n}{\mathrm{t}}_{\mathrm{i}\mathrm{j}}$ indicates whether the individual ‘i’ in village ‘j’ was assigned to receive the tailored information of vaccination schedule (Treatment) compared to the general information (Control). In this regression specification, we controlled for children’s age and village fixed effects, ${\mathrm{v}}_{\mathrm{j}}$. Village fixed effects control for any unobserved characteristics of village that are time-invariant. The coefficient of interest to measure the impact of tailored information is ${\beta }_1$. We have several main outcome variables (${\mathrm{y}}_{\mathrm{i}\mathrm{j}}$): the number of times the children visited the health clinic for vaccination after the intervention (as well as whether children visited the health clinic), the number of days it took for children to visit the health clinic after the intervention, and the vaccine uptake.

Outcome variables

1. Clinic visit

The first outcome is whether children visited the health clinic for the vaccination after the intervention. We hypothesized that the tailored information minimizes women’s misunderstanding about the vaccination schedule, which promotes the clinic visits. We recorded women’s visits for their children’s vaccination (only for vaccination purposes) at the health facility.

• (ii) The number of times children visited the health clinic for vaccination after the intervention

Because visiting the health facility multiple times, depending on children’s age, is necessary to complete the vaccination schedule, we measure the number of times children visited the health clinic for vaccination, given their age. Through this outcome variable, we can test whether the tailored information can increase the number of clinic visits, which leads to the increased number of vaccination.

• (iii) The number of days it took for children to visit the health clinic after the intervention

The literature suggests that the delayed vaccination uptake is the issue, so it is important to determine whether tailored information promotes timely vaccination among children. We measured the number of days from the intervention to the date of the clinic visit among those who visited the clinic for vaccination after controlling for the age of children.

• (iv) The vaccine uptake

The ultimate outcome is the vaccine uptake. We measured whether the children received each type of vaccination for their age.

For dummy variables (i and iv), we employed logistic regression. Meanwhile, we used the ordinary least squares (OLS) regression for other nonbinary variables (ii and iii).

Ethical approval

Ethical approval was obtained from the Adamawa State Ministry of Health (S/MoH/1131/I).

Results

Our final sample included 534 women from 11 settlements. Out of these 534 women, the analysis sample was reduced to 515 women with no missing information. These 515 women were divided into two groups: 317 (61.6%) were in the control arm, while the remaining 198 women (38.4%) were in the treatment arm. Table 1 presents the baseline characteristics of women (caregivers) and children and the balance. Almost 30% of the caregivers were between 21 and 25 years old, 25% were between 26 and 30, 13% were between 31 and 35, and 7% were over 35. Over 43% did not have any form of education, while 23% had primary education and the remaining 34% had secondary education. On average, the number of participants’ household members was 6.8, and 42% of women had paid jobs. The children’s average age was 4.5 months. The vaccination rate among these children was as follows: 69.9% at birth, 54.6% at 6 weeks, 43.1% at 10 weeks, 33.0% at 14 weeks, and less than 2% at 9 months (1.4%). Most of the variables were balanced between the control and treatment groups. One exception is the vaccination rate at 10 weeks. The rate was 46.4% in the control group, while it was 38.4% in the treatment group.

Table 1. Baseline characteristics and balance

	Mean
	Control (n = 317)	Treatment (n = 198)	Difference
Caregivers’ age (% women)
21–25	0.29	0.29	−0.00
26–30	0.26	0.23	−0.03
31–35	0.13	0.12	−0.01
over 35	0.06	0.08	0.01
Education level (% women)
Primary	0.21	0.26	0.05
Secondary and more	0.36	0.32	−0.04
Number of household member	6.74	7.04	0.30
Caregiver has paid job (% women)	0.44	0.40	−0.03
Children’s age (in months)	4.52	4.50	−0.02
Children’s vaccination (% children)
All vaccines completed at birth	0.73	0.67	−0.06
All vaccines completed at 6 weeks	0.56	0.54	−0.02
All vaccines completed at 10 weeks	0.46	0.38	−0.08*
All vaccines completed at 14 weeks	0.35	0.30	−0.05
All vaccines completed at 9 months	0.02	0.01	−0.00

* significant at 10%

Table 2 presents the effect of the treatment – the provision of tailored information on the vaccination schedule – on the number of visits to health facilities for vaccination and the timing of the clinic visits. We measured the timing of the clinic visit with the number of days it took for women to take their children to the clinic for vaccination since the date of the baseline survey. If women received the tailored vaccination schedule information, the number of times they visited health facilities is 0.18 times higher than in the case of the women in the control group who received only the generic vaccination schedule information. The treatment did not change the timing of the first health clinic visit, however.

Table 2. Effect of treatment on clinic visit for vaccination: OLS

Sample:	All	Among those who visited the clinic
	# times to visit	Days before visiting clinic
Dependent variable:	(1)	(2)
Treatment	0.177*	−3.711
	[−0.024,0.377]	[−18.993,11.571]
Age in months	−0.030**	0.652
	[−0.053,-0.007]	[−0.946,2.250]
N	513	177
r2	0.044	0.056

The result is based on the OLS regression with village fixed effect. Column 1 presents the effect of tailored information (Treatment) on the number of times women visited the health clinic for their children’s vaccination among all the respondents. Column 2 presents the effect of tailored information (Treatment) on the number of days it took from the intervention to the actual health clinic visits among respondents who visited the clinic. Children’s age is controlled for. * significant at 10%, ** significant at 5%, *** significant at 1%.

Table 3 presents the effect of the treatment on various important indicators, namely, whether women visited the health clinic at least once after the intervention and the vaccination status at different times. Consistent with the findings shown in Table 2, the treatment increased the odds of visiting the health center for vaccination at least once by 1.5. On the other hand, the treatment did not improve the vaccination rate at any given time and the full vaccination rate.

Table 3. Effect of treatment on vaccine uptake: odds ratio [95% confidence interval]

		Vaccine completed
	Visited health center for vaccination at least once	for 0 month	for 6 weeks	for 10 weeks	for 14 weeks	for 9 months	Full
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treatment	1.470**	0.715	1.045	0.951	0.949	1.090	0.951
	[1.010,2.139]	[0.465,1.099]	[0.700,1.560]	[0.651,1.390]	[0.651,1.384]	[0.720,1.649]	[0.598,1.512]
Age in months	0.967	1.082**	1.199***	1.178***	1.118***	1.044*	1.079***
	[0.919,1.018]	[1.007,1.162]	[1.112,1.292]	[1.099,1.263]	[1.050,1.191]	[0.999,1.091]	[1.026,1.136]
N	513	513	513	513	513	513	513

The result is based on the logistic regression with village fixed effect. * significant at 10%, ** significant at 5%, *** significant at 1%.

Discussion

We found that tailored information increased the likelihood and the number of health clinic visits among children and their caregivers. The potential positive effect of information provision on vaccination uptake is well cited in extant studies, but only from the perspective of the lack of information. Many studies indicated that the lack of knowledge on vaccination is one of the important barriers to vaccination.^18–21 However, no causal study has evaluated the effect of information provision about vaccination schedule on the vaccine uptake. Thus our study is important to test the potential effectiveness of information provision, particularly tailored information, on the vaccination behavior.

Although we found that tailored information positively affected health facility visits among children and their caregivers, it did not improve the vaccination rate for any given age of children, nor did it increase the full vaccination rate. This result implies that one-time tailored information has only an immediate but no sustainable effect. However, because one-time information provision can influence the immediate vaccination behavior, it might be important to constantly remind women about the vaccination schedule in a timely manner, particularly right before when caregivers are expected to take their children to the health facility for vaccination.

In order to remind the caregivers about the vaccination schedule repeatedly, using the reminder system is one proven way to improve the vaccination rate in Africa, including Nigeria.^22–24 As the vaccination schedule is complex and can change over time, providing one-time information is not sufficient for women to secure full vaccination for their children. It is important to provide tailored and real-time information.

Limitations

Our study has some limitations. First, our assignment ratio between the control and treatment groups was not equal, although the equal probability of the treatment assignment was our original design. We do not know the reasons for this unbalanced assignment ratio. However, as Table 1 presents, the characteristics of the respondents between the control and treatment groups are similar. Second, there are many other potential barriers to vaccination, such as vaccine hesitancy, lack of access to health facilities, social or economic instability, incapacity of the health system during a state of emergency such as the COVID-19 pandemic.^25–27 The present study exclusively focused on the effect of tailored information on the vaccination uptake, leaving those factors not assessed. Third, because this study only covered a small geographical area, we cannot generalize this result to other settings.

Conclusion

One-time tailored information on the vaccination schedule increases women’s clinic visits for the vaccination of their children, but it is not sufficient to improve the full vaccination rate.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Supplemental data

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2021.1936861.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.