When a study counts country–year–vaccine observations, it is describing a statistical panel — not counting individual children. That distinction matters when interpreting the findings.
- The assembled panel contained 44,472 country–year–vaccine observations; 38,405 had non-missing vaccination coverage in a reported analysis.
- In one shock-threshold specification, the authors estimated a 4.6-percentage-point lower coverage measure in the shock year, with uncertainty.
- The abstract reports 193 countries while a figure caption describes a 195-country estimation panel. The distinction was not resolved in the sections we checked.
- Paper date
- 17 November 2026
- Period
- 2000–2023
- Unit
- Country–year–vaccine observation
Original publication: 17 Nov 2026 · The date above refers to this brief.
Descriptive coverage by displacement quintile
percent
From national data to a cautious finding
Combine
Link national coverage estimates with displacement and other country-level indicators.
Compare
Estimate differences using specified thresholds and comparison groups.
Qualify
Examine uncertainty, reporting differences and limits of national averages.
What was compared?
The researchers combined national immunization coverage estimates with displacement, conflict, disaster and economic data. The outcome was a population-level proxy for service delivery. A country–year–vaccine observation is one entry for a country, year and vaccine; it is not a child or a directly observed vaccination event.
Source 1 ↗A descriptive difference is not the model estimate
In pooled descriptive comparisons, average coverage was 90.6% in the lowest displacement quintile and 80.4% in the highest: a 10.2-percentage-point difference calculated from the reported values. Countries and years can differ for many reasons. This descriptive gap is not the same quantity as the paper’s adjusted shock estimate.
Source 1 ↗A specific model, with uncertainty
For a threshold of at least 500,000 displaced persons and at least 100% year-on-year growth, the reported shock-year estimate was −4.6 percentage points, with a 95% confidence interval from −8.4 to −0.8 in the threshold analysis. It depends on the definition, comparison group and model assumptions. Larger thresholds involved fewer countries and wider intervals.
Source 1 ↗A reporting detail worth checking
The abstract and methods describe 193 countries. The Figure 3 caption describes an estimation panel of 195 countries, while Figure 1 describes a 193-country panel. The reviewed sections do not make clear whether this reflects different panel definitions or another issue. We do not choose one number for every analysis without clarification.
Source 1 ↗What country averages cannot resolve
National estimates can conceal unequal access within countries, including among displaced and non-displaced communities. The data do not identify individual children or prove the exact pathway behind a coverage change. The authors report sensitivity analyses, but unmeasured differences and the assumptions behind the comparison remain important.
Source 1 ↗A panel entry is not a child
Publisher HTML: abstract, Results, Figure 1–4 captions, Methods and ethics, accessed through a research-search interface. Publisher metadata and abstract also checked. Supplements, source-data files, code and clinical records were not independently audited. No publisher figure reproduced.
01Are there 44,472 individual child participants?
- What was observed
- The paper reports 44,472 country–year–vaccine observations.
- Where the conclusion stops
- Entries represent aggregate records, not individually enrolled children.
Source 1 · Publisher Results and Figure 1 caption
02Does the paper use one unambiguous country count everywhere?
- What was observed
- The abstract and methods describe 193 countries; the Figure 3 caption describes a 195-country estimation panel.
- Where the conclusion stops
- The reason for the difference was not resolved in the sections reviewed.
Source 1 · Publisher abstract, Figure 1 and Figure 3 captions
Numbers you can inspect
| Measure | Value & unit | Origin & method |
|---|---|---|
| Assembled panel observations | 44472 country-year-vaccine entries | Reported Panel observations, not children. Source 1 · Publisher Results and Figure 1 caption |
| Entries with coverage | 38405 country-year-vaccine entries | Reported Observations with non-missing vaccination coverage in the reported analysis. Source 1 · Publisher Figure 2 caption |
| Lowest displacement quintile | 90.6 percent | Reported Reported descriptive average vaccination coverage. Source 1 · Publisher Results, global associations |
| Highest displacement quintile | 80.4 percent | Reported Reported descriptive average vaccination coverage. Source 1 · Publisher Results, global associations |
| Descriptive coverage difference | 10.2 percentage points | Calculated 90.6 minus 80.4; unadjusted descriptive difference, not a model estimate. Source 1 · Publisher Results, global associations |
| Threshold-specific shock-year estimate | -4.6 percentage points | Reported Point estimate for at least 500,000 displaced persons and 100% year-on-year growth; 95% CI −8.4 to −0.8. Source 1 · Publisher Results, threshold specification |
| Country count in abstract | 193 countries | Reported Country count described in the abstract and methods. Source 1 · Publisher abstract and Methods |
| Country count in Figure 3 caption | 195 countries | Reported Estimation panel described in the figure caption; difference from 193 is unresolved here. Source 1 · Publisher Figure 3 caption |
Compare the actual experiments
These studies answer different questions. Read the unit and endpoint before comparing results.
| Study | Unit & setting | Readout | Interpretation boundary |
|---|---|---|---|
| Descriptive quintiles Source 1 · Publisher Results, global associations | Country–year–vaccine observations | Mean coverage in lowest and highest displacement groups | Not an adjusted estimate of a displacement shock. |
| Shock-threshold analysis Source 1 · Publisher Results and Figure 3 caption | Countries and their aggregate vaccine coverage | Difference-in-differences estimate under a specified threshold | Assumptions and confidence interval matter; it does not reveal individual outcomes. |
The export includes claims, available numbers, methods and source locations. It contains our reading notes and published summaries; it is not raw participant data or an independent reanalysis.
Evidence update · 17 Nov 2026
First publication. Reported observations, arithmetic and proposed follow-up questions are separated. No raw-data reanalysis. AI source check; no human editorial or clinical review.
Do not mix these quantities
| Reading question | What the paper reports | What remains open |
|---|---|---|
| 44,472 | Country–year–vaccine entries in the assembled panel | Not individual children |
| 38,405 | Entries with non-missing coverage in the reported analysis | A subset of observations, not a separate population |
| 10.2 percentage points | 90.6% minus 80.4%, descriptive quintile difference | Not the adjusted shock estimate |
| −4.6 percentage points | One threshold-specific shock-year model estimate | Depends on model assumptions and has uncertainty |
Original Press-News reading framework. Follow-up questions are our interpretation, not experiments performed by the authors.
Your questions, answered
Were 44,472 children enrolled?
No. Those are country–year–vaccine observations in an aggregate dataset.
Is the descriptive 10.2-point gap the adjusted effect?
No. The model estimate uses a different comparison and assumptions.
How many countries were analyzed?
The abstract describes 193; the Figure 3 caption says 195 in an estimation panel. The reason was not resolved in the sections checked.
Does a national average show who missed vaccination?
No. National data cannot identify the vaccination experience of individuals or specific local groups.
Limits of this interpretation
- The analysis uses aggregate national estimates and cannot identify individual vaccination experiences.
- Observed associations rely on modeling assumptions and can be affected by unmeasured differences and reporting quality.
- The abstract and Figure 3 caption report different country totals; the reason was not clear in the sections reviewed.
- A national coverage average cannot show which local populations experienced disruption.
Sources & transparency
- Headley, Shay and Tozan (2026): Forced displacement is associated with health service delivery disruption beyond the effects of armed conflict
Publisher HTML: abstract, Results, Figure 1–4 captions, Methods and ethics, accessed through a research-search interface. Publisher metadata and abstract also checked. Supplements, source-data files, code and clinical records were not independently audited. No publisher figure reproduced. · Accessed 17 Nov 2026
DOI: 10.1038/s41467-026-79274-2
Prepared and source-checked with AI. Press-news Team is the collective publication byline, not a medical reviewer. No human editorial or clinical review has taken place. This educational article examines research methods and laboratory or observational findings, not individual diagnosis or treatment. We did not conduct these experiments or reanalyse raw data. Reported results, our interpretation and proposed future tests are separated. Source-access limits are recorded below. Photographs are illustrative.
Source check: AI source check — selected primary results, captions and study boundaries
Clinical review: Not applicable to this educational guide
Suggest a correction