Impact of COVID-19 pandemic on childhood immunization coverage in Indonesia: lesson learned from a nationwide analysis of the Expanded Programme on Immunization

This nationwide analysis of Indonesia's Expanded Programme on Immunization reveals that the COVID-19 pandemic caused a significant, uneven decline in childhood immunization coverage during its first year, particularly in districts with high infection rates and weak health system capacity, though coverage largely recovered by the second year.

The Gist

Imagine Indonesia's healthcare system as a massive, intricate web of 514 tiny fishing villages (districts) scattered across a vast ocean. Each village has its own boat (health resources) and crew (doctors and midwives) responsible for catching a very specific kind of fish: vaccinated children. Before the storm, most villages were catching enough fish to keep their nets full and the community healthy.

Then, a massive hurricane named COVID-19 hit.

This paper is a report card on how that hurricane shook the fishing fleet, why some villages lost their catch while others barely noticed, and what we can learn to build better boats for the next storm.

The Big Picture: The Storm Hits

Before the pandemic (2019), about 83% of children in Indonesia got their full set of vaccines. It was a well-oiled machine.

When the storm arrived in 2020, the winds blew hard. The government had to close the ports (lockdowns) to stop the virus from spreading. Unfortunately, the boats carrying vaccines couldn't sail, and the fishermen (health workers) were too scared or too busy fighting the storm to go out.

• Year 1 of the Storm (2020-2021): The national catch dropped to 75%. That sounds small, but in a country of 280 million people, it meant over a million children missed their vaccines.
• Year 2 of the Storm (2021-2022): The winds died down a bit. The catch recovered to 88%, which is actually higher than before the storm! However, this recovery was uneven. Some villages bounced back quickly, while others were still struggling.

The Uneven Damage: Why Some Villages Suffered More

The most important finding of this study is that the hurricane didn't hit everyone equally. It's like a storm that destroys the weakest houses first while leaving the sturdy ones standing.

The researchers looked at the 514 districts and found that the ones hit hardest by the drop in vaccinations shared three specific "weak spots":

1. The Virus Was Stronger There: Districts where the virus spread like wildfire (high infection rates) saw their vaccination programs collapse. The health system was so overwhelmed by sick people that there was no time or energy left to vaccinate healthy kids.
2. The Crew Was Too Small: In villages where there weren't enough midwives (the local health heroes who often do the vaccinating), the system broke down. When the pandemic hit, these few midwives were pulled away to fight the virus, leaving no one to vaccinate the children. It's like a small fishing crew trying to fix a broken net while also fighting off sharks; they just can't do both.
3. The "Hospital-Only" Trap: Surprisingly, districts where most mothers gave birth in hospitals saw a bigger drop in vaccinations. Why? Because during the early pandemic, hospitals were scary places. They required negative virus tests for delivery, and many health workers were reassigned to the front lines. Mothers stayed home, and the chain of getting babies vaccinated right after birth was broken.

The "Zero-Dose" Problem

The study highlights a scary reality: when you stop vaccinating, you don't just pause the progress; you invite monsters back. The paper mentions that because of this drop, Indonesia is now seeing outbreaks of diseases like measles and polio that we thought were under control. It's like stopping the fire department for a year; eventually, the small sparks turn into a blaze.

The Lesson Learned: Building Storm-Proof Boats

So, what's the takeaway for the future?

• Don't rely on just one type of boat: The study suggests that if midwives are busy, we need other trained people (like dentists or pharmacists) to step in and vaccinate. We need a backup crew.
• Strengthen the weak links: The districts that struggled the most were the ones that were already struggling before the storm. We need to give extra resources to the villages with the smallest crews and the weakest boats before the next hurricane hits.
• Keep the lights on: Even during a crisis, we must find a way to keep the "vaccine boats" sailing. Stopping routine health services to fight a new disease often leads to more deaths from old diseases.

In a Nutshell

The COVID-19 pandemic was a stress test for Indonesia's health system. It showed that while the system is strong in some places, it is fragile in others. The storm didn't just cause direct damage (people getting sick with COVID); it caused indirect damage by knocking out the routine care that keeps children safe.

The paper concludes that to survive the next crisis, we need to build a health system that is resilient—one that can keep vaccinating children even when the sky is falling. We need to make sure that no matter how hard the wind blows, the fishing nets stay full.

Technical Summary

1. Problem Statement

The COVID-19 pandemic caused significant disruptions to routine health services globally, with routine immunization being among the most affected. While global data indicated a decline in vaccination coverage, evidence regarding the specific magnitude, heterogeneity, and underlying drivers of this disruption in low- and middle-income countries (LMICs) remained limited.

• Context: Indonesia, the fourth most populous country with a decentralized health system (514 districts), faced the highest COVID-19 burden in Southeast Asia.
• Gap: Prior to this study, there was no comprehensive evaluation of how the pandemic impacted childhood immunization coverage across all 514 districts in Indonesia, nor was there an analysis of the district-level factors (health system capacity, human development, and pandemic burden) associated with these declines.
• Consequence: Disruptions led to a rise in "zero-dose" children and subsequent outbreaks of vaccine-preventable diseases (VPDs), including measles and vaccine-derived polio.

2. Methodology

• Study Design: A nationwide longitudinal analysis of the Expanded Programme on Immunization (EPI).
• Data Source:
  • Immunization Data: District-level data from the Ministry of Health (MoH) immunization registry covering 514 districts across 34 provinces (2019–2021).
  • Explanatory Variables: Data from the National COVID-19 Task Force, National COVID-19 vaccine database, 2018 Public Health Development Index (PHDI), and 2020 Human Development Index (HDI).
• Timeframe:
  • Pre-pandemic: January 2019 – February 2020.
  • Year 1 of Pandemic: March 2020 – February 2021.
  • Year 2 of Pandemic: March 2021 – December 2021.
• Outcome Variable: Annual childhood immunization coverage (proportion of children aged 0–12 months receiving all recommended doses: Hepatitis B, BCG, 3x DPT-HB-Hib, 4x Polio, 1x Measles).
• Statistical Analysis:
  • Descriptive: Comparison of national and island-level coverage trends.
  • Coverage Ratio: Calculated by comparing pandemic-year coverage to pre-pandemic baseline. Districts with a statistically significant decline (ratio < 1, p < 0.05) were classified as "affected."
  • Regression Modeling: Mixed-effect logistic regression was used to assess associations between explanatory variables and coverage decline in Year 1.
    • Random Effect: Province (to account for clustering; 23.4% of variance attributed to provincial clustering).
    • Selection: Forward selection approach; variables with p < 0.20 in bivariable analysis entered the multivariable model.
    • Significance: Two-sided p < 0.05.

3. Key Contributions

• Granular Geographic Analysis: This is one of the first studies to evaluate the pandemic's impact on immunization at the district level (n=514) in a large LMIC, revealing significant heterogeneity masked by national averages.
• Identification of Specific Drivers: The study isolates specific district-level factors (beyond just COVID-19 case counts) that exacerbated immunization failures, specifically focusing on health workforce adequacy and facility-based delivery patterns.
• Policy-Relevant Evidence: The findings provide a blueprint for strengthening health system resilience in decentralized systems, highlighting the need for targeted interventions in districts with weak health infrastructure.

4. Results

Immunization Coverage Trends:

• National Level: Coverage dropped from 83.2% (pre-pandemic) to 75.0% (Year 1), then recovered to 88.6% (Year 2).
• District Impact:
  • Year 1: 69.3% (356/514) of districts experienced a significant decline. The national coverage ratio was 0.92 (95% CI: 0.89–0.94).
  • Year 2: The number of affected districts decreased to 36.2% (186/514), with a national coverage ratio of 1.07 (95% CI: 1.05–1.10), indicating recovery.
• Geographic Variation: Java Island saw the largest absolute drop (-13.6%), while Sulawesi showed the strongest recovery (+32.3% in Year 2). Coverage ratios varied wildly, from 0.07 in South Monokwari to 4.52 in Paniai during Year 1.

Multivariable Analysis (Factors Associated with Decline in Year 1):
Three factors were significantly associated with a higher likelihood of immunization coverage decline:

1. High COVID-19 Incidence: Districts in the highest incidence quartile had a 2.19 times higher odds of decline compared to the lowest (aOR 2.19, 95% CI 1.01–4.73).
2. Low Midwife Adequacy: Districts with the lowest proportion of villages having adequate midwives had the highest odds of decline (aOR 5.84, 95% CI 2.40–14.16) compared to those with the highest adequacy.
3. High Proportion of Health Facility-Based Births: Contrary to expectations, districts with a higher proportion of facility births (middle-high vs. lowest) were more likely to see a decline (aOR 2.98, 95% CI 1.49–5.98).

Note: Factors such as COVID-19 mortality, stunting, malnutrition, doctor adequacy, and posyandu (community health post) adequacy were significant in bivariable analysis but lost significance in the multivariable model.

5. Significance and Implications

• Health System Resilience: The study demonstrates that the pandemic's impact was not uniform; it disproportionately affected districts with weaker health systems (specifically low midwife availability) and higher disease burdens.
• Workforce Criticality: The strong association between midwife adequacy and immunization coverage highlights the critical role of midwives in Indonesia's immunization delivery (often acting as vaccinators in community posts). Their redeployment to pandemic tasks severely disrupted routine services.
• Facility-Based Paradox: The finding that facility-based births correlated with lower immunization coverage suggests that pandemic policies (e.g., mandatory negative PCR tests for delivery) and staff shortages in hospitals created barriers that prevented mothers from accessing postnatal immunization services, a phenomenon not observed in other LMIC contexts.
• Future Preparedness: The authors recommend:
  • Protecting maternal and child health services during crises.
  • Redistributing healthcare workforces and recruiting volunteers.
  • Expanding the scope of non-physician vaccinators (e.g., dentists, pharmacists) to ensure continuity of care.
• Long-term Risk: Despite recovery in Year 2, the study notes a continued decline in districts meeting national targets in 2022–2024, correlating with recent VPD outbreaks (measles, polio), underscoring the long-tail effects of pandemic disruptions.