Missed Opportunities for Childhood Immunization in Pakistan: Failure to Vaccinate Despite Health System Contact

Study design and Data source

A secondary data analysis was conducted using the Pakistan Social and Living Standards Measurement Survey (PSLM) 2019-20. Data on sociodemographic factors were extracted to contextualize the reasons for MOVs. The PSLM survey provides biannual national-level statistics that help governmental agencies, stakeholders, and non-governmental organizations monitor health progress using key indicators. Based on the variation in population characteristics in the distribution of urban and rural populations, the availability of resources, cost, and disability coverage. A sample of 195,000 households from 6500 sample areas was taken, and 103,464 children were eligible for immunization. Final regression analysis was conducted on approximately 2500 above mothers who visited healthcare facility. Ethical approval was not required for the study.

Sample and Participants

The analysis included children under 12 months who were excluded due to the standard administration of the measles vaccine at 12 months and the absence of precise birthdates, ensuring the inclusion of only eligible participants. Children with zero-dose status were excluded from the primary analysis to focus specifically on MOV occurring during identifiable healthcare encounters, consistent with the WHO MOV framework. Although Pakistan’s immunization system includes outreach modalities such as door-to-door campaigns, mobile teams, and community health workers, the PSLM dataset dose not capture the timing, or nature of these contacts. Consequently, it is not possible to ascertain whether zero-dose children experience a missed opportunity during a documented health interaction. Their exclusion therefore reflects a methodological constraints of the dataset rather than an assumption of no health system contact, which is acknowledged as a study limitation.

Outcome variable

MOV was defined as failure to receive an age-appropriate dose of either the pentavalent or measles vaccine despite the mother’s visit to a healthcare facility. Defining MOV as a binary variable involved assessing whether children aged 1-5 missed at least one dose of either pentavalent or measles vaccine despite maternal contact with any healthcare facility. The variable was coded as 1 if the dose was unadministered despite contact and 0 otherwise, based on recall or documented vaccination records. Separate variables were created for each vaccine dose (penta-1, penta-2, penta-3, measles-1, and measles-2). Table 1 shows Pakistan’s expanded program on immunization (EPI) schedule for pentavalent and measles vaccines.

Predictor variable

We used reports of contact with healthcare facilities through maternal visits as a predictor variable. The private sector included clinics, dispensaries, or hospitals, and public sector facilities such as public dispensaries, hospitals, Basic Health Units (BHU), Rural Health Centers (RHC), visits to Lady Health Visitors (LHV), and Lady Health Workers (LHW). Visits were coded as a binary variable: 1 for the private sector and 0 otherwise. Our evaluation used maternal healthcare access data due to popular evidence of infrequent vaccinations during sick child visits, the influence of false contraindications, and providers’ widespread failure to screen for vaccinations.²³ Based on prior literature and their availability in the PSLM 19-20 database, the regression model incorporated several determinant variables, including socio-demographic variables (region, child age, child sex, total household members, household head sex, maternal age, maternal formal education, maternal employment status, and wealth index).

Analysis

The prevalence of MOVs for each dose of pentavalent and measles vaccines was calculated in children aged 1–5 years. Among children identified with MOV, the proportion of mothers who consulted healthcare facilities due to sickness or injury within the last two weeks was marked. These proportions were further stratified by the types of healthcare facilities consulted and determinant variables. To visually represent these findings, maps showing the proportions of missed vaccination opportunities for each dose of pentavalent and measles vaccines across Pakistan were generated using ArcGIS software (V 10.8.2).

Those who had received at least one vaccination dose but encountered MOVs during subsequent opportunities were the focus of the study, whereas children with zero dose status were excluded. Potential confounders were selected a priori based on existing literature on childhood immunization and the WHO conceptual framework for MOV. Variables were included if they were plausibly associated with both healthcare-seeking behavior and vaccination outcomes, and were available in the PSLM dataset. These included child characteristic (age, sex), household characteristics (household size, sex of household head and wealth index), maternal characteristics (age, education, employment status), and context factors (urban-rural residence and province). All regression models adjusted for the same set of covariates to ensure comparability across vaccine doses. No data-driven stepwise procedures were used. For multivariate analysis, weighted, multivariable binomial logistic regression was used to measure the association between maternal visits to private healthcare facilities and missed opportunities for pentavalent (Penta-1, Penta-2, and Penta-3) and measles (Measles-1 and Measles-2) vaccinations. The model was adjusted for potential confounders to yield adjusted odds ratio estimates. A sub-analysis was conducted for each province to assess disparities among the provinces. Measures of association are presented as adjusted odds ratios (aORs) with 95% confidence intervals (CI). Mother’s reported healthcare facility issues (i.e. staff cooperation, waiting time, cleanliness, medicine availability, doctors availability, treatment cost, treatment failure) were included as binary covariates in the national-level regression model and adjusted for alongside sociodemographic factors. All analyses were performed using STATA version 17 (College Station, Texas, USA).

Missed vaccinations

We found that 26% of children aged 1-5 missed at least one dose of measles, and 13% had missed a pentavalent vaccine ( Figure 1).

Figure 1. Flow-Chart Analyzing Distribution of Children Aged 1-5 for Assessment of Missed Opportunities of Vaccination, 2019-2020.

At the national level, missed vaccine rates range from 9.7% to 25%, with the highest rates in Balochistan, ranging from 29.5% to 43.1% for specific vaccines, and the lowest in Punjab, where the range is between 3.2% and 18.6%. Missed vaccinations were higher in rural than in urban areas (11.2 to 24.9%), and for children of uneducated mothers (17.1% to 30.9%), and among those visiting private facilities. Although point estimates suggested higher missed opportunities among male children (10.2 to 25.4% for boys vs. 9.3 to 24.5% for girls), and male-led households (10–25.4% vs. 5.8–19.2%), the confidence intervals overlapped, indicating that these differences should be interpreted cautiously.

At the national level, pentavalent vaccines were missed by 13% and measles by 25% ( Figure 2), with considerable interprovincial variation. There are always more children who miss measles than the pentavalent vaccine.

Figure 2. All Missed Vaccinations for Pentavalent and Measles Doses on National & Provinces of Pakistan in 2019-2020.

The number of missed vaccinations was higher in sparsely populated areas. Balochistan has the least dense population and the highest missed rates at 35.4% for pentavalent and 45.4% for measles, followed by Sindh with 26.1% for pentavalent and 38.5% for measles. In contrast, Punjab and KP, with denser populations, had the lowest missed rates. At the district level, in Balochistan, very sparsely populated Khuzdar had 83.8% pentavalent and 88.4% measles missed cases, while Awaran had 80.9% and 89% missed cases, respectively. In Sindh, Tharparkar reported 80.2% for pentavalent and 87.4% for measles, while Kamber Shahdabkot reported 63.2% and 71.5%, respectively ( Figure 3).

Figure 3. Population Density per Square KM of Pakistan at the National, Provincial and District Level from Census 2017 & 2023.

This population-missed vaccination relationship was visualized using heat maps ( Figure 4). The districts of Awaran (Penta-1 = 80.9%, Penta-2 = 80.9%, Penta-3 = 80.9%, Measles-1 = 80.5%, Measles-2 = 88.4%), Barkhan (Penta-1 = 81.2%, Penta-2 = 81.6%, Penta-3 = 81.3%, Measles-1 = 82.4%, Measles-2 = 83.8%), Khuzdar (Penta-1 = 83.8%, Penta-2 = 83.8%, Penta-3 = 83.9%, Measles-1 = 83.8%, Measles-2 = 88.3%), Qilla Saifullah (Penta-1 = .61.6%, Penta-2 = 64.1%, Penta-3 = 67.01%, Measles-1 = 61.9%, Measles-2 = 68.5%), and Qilla Abdullah (Penta-1 = 54.8%, Penta-2 = 54.9%, Penta-3 = 55.63%, Measles-1 = 49.9%, Measles-2 = 52.7%) in Balochistan along with Tharparkar (Penta-1 = 71.1%, Penta-2 = 76.2%, Penta-3 = 82.3%, Measles-1 = 80.3%, Measles-2 = 87.4%) in Sindh, respectively. The spatial distributions of pentavalent and measles were identical in various districts of Pakistan.

Figure 4. Geospatial Analysis of Missed Opportunities for Vaccination (MOVs) for Pentavalent and Measles Doses in Districts of Pakistan in 2019-2020.

Missed opportunities to vaccinate

When a mother presents to a healthcare provider, there is an opportunity to ask for and catch up with missed doses of vaccines for her children. However, if the mother visits facilities that do not offer vaccinations, such opportunities are consistently lost. We found that 10% of mothers had children with a missed dose of the pentavalent vaccine (Penta 1). Of these, 91% had visited a healthcare provider. Nationally, the odds of missing the first dose of the pentavalent vaccine (Penta-1) were higher if a private facility was visited (OR: 1.364, CI: 1.337, 1.391) compared to those visiting public facilities ( Table 3). Missed opportunities vary across the provinces. KP had the most missed opportunities for the third dose of pentavalent vaccine (Penta-3) (OR: 2.232, CI: 2.119, 2.351). In contrast, Sindh (OR: 0.555, CI: 0.539, 0.571) had fewer odds of missing penta-3 if the mother had visited a private facility compared to a public facility. For measles vaccination, Balochistan had the highest burden of Measles-2 in the context of private healthcare setups (OR: 1.360, CI: 1.271, 1.455). For Measles-2, Punjab demonstrated the lowest odds of missed vaccinations compared to the other provinces (OR: 0.697, CI: 0.682, 0.712). Individual provincial-level regression analyses are presented in Annexure Tables 1–4.

Table 4 presents the national-level analysis of MOV, including all socioeconomic and demographic factors among the children whose mothers visited a healthcare facility. Urban residence, smaller households, and those with unemployed mothers were associated with higher odds of missing all doses of vaccines compared to rural areas. Male children had marginally lower odds of missed opportunities for early vaccinations, but higher odds for Measles-2 (OR: 1.149, CI: 1.133,1.165) than female children.

The model also accounted for complaints about healthcare facilities, with uncooperative staff, long waiting times, non-availability of medicine, and lack of cleanliness notably associated with higher odds of missed opportunities for all vaccines. The wealth index did not exhibit a linear trend indicative of lower odds of missing vaccines among wealthier households compared to the poorest quintile.

Limitation and Future Policy Implications

This study relied on existing national survey data, over which the investigators had no control regarding question wording or recall periods. In particular, maternal healthcare visits were captured only for the two weeks preceding the survey, which may have led to an underestimation of MOV. In addition, the analysis could not account for certain clinical and service-level factors, such as illness severity during maternal visits, provider referral practices, or recall bias in vaccination history. These unmeasured factors may have influenced the observed associations and warrant cautious interpretation of the findings.

The scope of the analysis was further shaped by data limitations related to vaccine delivery mechanisms. Although Pakistan continues to face challenges in polio eradication, polio vaccination was not included because the PSLM dataset does not capture information on supplementary immunization activities, door-to-door campaigns, or repeated oral polio vaccine doses. As polio delivery differs substantially from routine EPI vaccines, particularly pentavalent and measles, missed opportunities for polio vaccination cannot be reliably inferred from maternal healthcare visits. Consequently, the findings of this study should not be directly extrapolated to polio vaccination without dedicated data capturing campaign-based delivery.

Within these constraints, the policy recommendations emphasize interventions that are both evidence-informed and operationally feasible. Integrating vaccine screening prompts into digital health platforms used by private providers represents a high-feasibility, short-term intervention that leverages existing urban primary care infrastructure. Such systems can prompt providers to routinely assess a child’s immunization status during visits and refer eligible children to nearby government vaccinators, potentially yielding rapid gains in referral and reporting.⁵² Community engagement initiatives to increase vaccine demand require sustained outreach and coordination and are therefore more appropriately positioned as medium-term strategies. In contrast, formal integration of private providers into national immunization planning constitutes a longer-term structural reform requiring regulatory alignment and institutional capacity building.