How We Track U.S. Childhood Vaccination Declines: Our Methodology

What We Mean by a Decline

A decline in U.S. childhood vaccination can refer to several different outcomes: fewer children receiving at least one dose, lower completion of a multi-dose series, more vaccinations occurring after the recommended age, or higher school-entry exemption rates.

These measures are related but not interchangeable. A child vaccinated late differs from a child who never received the vaccine. Similarly, one dose of a three-dose series does not constitute series completion. A school exemption is not necessarily evidence of intentional refusal because exemption categories vary by state and may include medical, religious, administrative, or legal classifications.

We therefore define a decline as a measurable reduction in a clearly specified outcome, including:

• Vaccination coverage.
• Vaccine uptake.
• Series completion.
• Timely vaccination.
• Documented doses.
• School-entry compliance.
• Coverage within a defined population.

National estimates can conceal substantial differences by state, county, income, race and ethnicity, insurance status, geography, and school type. A small national change may reflect a broad shift or a severe decline concentrated in a few communities.

The central methodological question is how to measure vaccination changes without confusing coverage data with public opinion, policy developments, or media attention. The available source set provides context on vaccine trust, federal guidance, policy processes, and international comparisons. It does not establish the size of a national decline in childhood vaccination. Numerical claims require verified primary data.

Questions This Methodology Answers

Is childhood vaccination coverage declining?

First, we identify the vaccine, dose number, age group, geographic area, population, and comparison years. A valid comparison may examine the percentage of children who completed a specified series by a defined age in one year versus the same measure in a later year. A school survey should not be combined with a household survey unless methodological differences are understood and adjusted.

For every comparison, we record:

• Vaccine and dose number.
• Recommended or measured age.
• Geographic area.
• Included population.
• Data collection period.
• Source and calculation method.
• Uncertainty estimates.
• Changes in definitions or reporting.

We report absolute change first:

Coverage fell from X% to Y%, a decline of Z percentage points.

We may then report relative change:

The decline represented approximately R% of the earlier coverage level.

Percentage-point change is generally easier to interpret. A reduction from 90% to 85% is a five-percentage-point decline and an approximately 5.6% relative decrease.

We do not describe a national decline when the evidence covers only selected states, schools, clinics, or survey respondents. The geographic and population scope must match the claim.

Which children are affected?

A national average does not show which groups experienced the largest change. We examine differences by:

• Age.
• State and region.
• Insurance coverage.
• Urban, suburban, and rural location.
• Household income.
• Race and ethnicity, where definitions remain consistent.
• School enrollment.
• Exemption status.
• Access to routine medical care.

This analysis shows whether a trend is broad-based or concentrated among specific groups. Overall coverage may remain stable while coverage falls among children with interrupted insurance, missed preventive visits, or limited access to pediatric care. Conversely, a national decline may be driven by a small number of jurisdictions with unusually large changes.

Small samples can produce unstable estimates, and subgroup definitions may change between surveys. We do not compare subgroup percentages without confirming that the groups were defined and measured consistently.

What may explain the change?

Observed vaccination outcomes must be separated from possible explanations. Potential factors include:

• Missed routine care during and after the COVID-19 pandemic.
• Transportation, scheduling, cost, or insurance barriers.
• Changes in school-entry requirements.
• Changes in nonmedical exemption rules.
• Vaccine confidence and trust.
• Conflicting health information.
• Changes to federal recommendations.
• Administrative or reporting changes.
• Differences in registry participation.

Each explanation receives an evidence label:

• Evidence-supported: Direct data or strong research support the explanation.
• Plausible: The explanation is consistent with available evidence but has not been established.
• Unproven: The claim lacks sufficient evidence.

A change that follows a policy announcement does not prove that the policy caused it. Causal claims require stronger designs, such as longitudinal research, natural experiments, linked administrative data, or carefully controlled comparisons.

Data Sources

Vaccination coverage data

The strongest evidence comes from official surveillance, immunization registries, school-entry records, and nationally representative surveys.

For each dataset, we record the publisher, publication date, collection period, covered population, vaccine or series measured, geographic scope, sample size or reporting population, numerator, denominator, calculation method, uncertainty estimates, missing-data treatment, and reporting limitations.

Whenever possible, year-over-year comparisons use the same source series. Consistent sources reduce the risk that an apparent change reflects a different survey design rather than a change in vaccination behavior.

Administrative records can cover large populations but may exclude children outside participating systems or contain incomplete documentation. Surveys can include children whose records are unavailable but may be affected by nonresponse, recall error, and sampling limitations. School records show compliance among enrolled children but do not represent children who are not enrolled or whose records are unavailable.

No single source answers every question.

Public opinion and trust research

Polling describes attitudes, trust, information sources, and policy preferences. It does not directly measure whether children received a vaccine.

KFF published the “KFF Tracking Poll on Health Information and Trust: Fall 2026 Update on Vaccine Views” on October 20, 2026. The available summary identifies the topic but does not provide specific findings. We therefore do not attribute numerical results to that source without reviewing the original publication. Source 1

Pew Research Center also examines views on childhood vaccines, vaccine research, and related policy. The available summary provides no specific numerical findings, so it establishes the research topic but not a quantified conclusion. Source 9

We treat reported confidence or concern as context, not proof of a coverage decline.

Policy and schedule sources

Policy analysis distinguishes four concepts:

1. A federal recommendation.
2. A state school-entry requirement.
3. A child’s actual vaccination status.
4. A proposed or pending policy change.

KFF reporting on changes to the federal vaccine schedule may help explain how guidance could affect clinicians and families. The source summary describes potential implications but does not establish an actual change in coverage. Source 3

A schedule revision may change recommended vaccines, timing, or eligibility. It does not demonstrate that children received fewer vaccines. That conclusion requires comparable outcome data collected before and after implementation.

Policy-process reporting must also remain separate from coverage measurement. Claims that an executive order bypassed evidence-based review may inform analysis of transparency, clinical input, and institutional safeguards, but they are not quantitative evidence of a vaccination decline. Source 7

International comparisons

International comparisons require matching vaccine definitions, dose counts, age groups, reporting periods, data systems, population coverage, and completion standards.

Denmark and the United States differ in health-system organization, administration, school requirements, access, and data infrastructure. NPR’s reporting cautions against transferring Denmark’s vaccine policy directly to the United States. Source 5

Another country’s vaccination rate is not a direct benchmark unless these differences are addressed.

How We Calculate Trends

Establish the denominator

The denominator is the population eligible for the vaccine or series. It may include all children in an age group, children enrolled in school, children with available medical records, survey respondents, or children living in reporting jurisdictions.

Changing the denominator can create an apparent trend even when vaccination behavior is unchanged. Every result must state who was counted.

Standardize the numerator

The numerator counts children meeting the source’s definition of vaccinated. This may mean at least one dose, a completed primary series, a dose received by the recommended age, documentation in a registry or medical record, or compliance with a school requirement.

Partial vaccination cannot be combined with series completion. A child who received one dose should not be counted as completing a three-dose series.

Calculate absolute and relative change

We present percentage-point change first, followed by relative change when useful:

Coverage fell from X% to Y%, a decline of Z percentage points. That represents an approximately R% relative change.

Confidence intervals or margins of error are included when available. A difference within measurement uncertainty should not be described as a confirmed decline.

Track timing, not only completion

Final coverage can conceal delays. When data allow, we examine whether children received doses by the recommended age. Delayed vaccination differs from permanent nonvaccination, but delays can leave children vulnerable when protection is needed. Later catch-up vaccination may increase final coverage without eliminating the risks associated with missed or late doses.

Comparing States and Years

Comparisons require consistent vaccine formulations, dose counts, age cutoffs, reporting methods, and definitions. We check for changes in survey wording, registry participation, school enrollment coverage, state reporting requirements, exemption categories, federal recommendations, and collection periods.

National estimates are reported separately from state findings. We do not average state percentages without considering population size. State outliers receive attention but are not treated as representative of the country. A state with high overall coverage may contain counties, schools, or neighborhoods with substantially lower coverage.

Multiple years help distinguish long-term trends, short-term disruptions, reversals, and differences between routine childhood vaccines and newer or updated vaccines. Language remains proportional to the evidence, using descriptions such as “declined during the measured period,” “remained below the earlier level,” “evidence is mixed,” or “the available data do not establish a national trend.”

How Policy Changes Enter the Analysis

A federal schedule revision may affect recommendations, timing, eligibility, or clinical communication. We compare old and new schedules directly before considering possible effects. Guidance is not the same as law, and neither establishes whether an individual child is vaccinated.

Evidence-based review, clinical input, transparency, and conflict-of-interest safeguards matter when evaluating vaccine policy. However, policy commentary alone cannot establish the size or direction of an effect on coverage. That effect must be tested with outcome data.

We also track changes in school-entry requirements and nonmedical exemptions. An exemption is not automatically equivalent to intentional refusal. Where possible, exemption rates are compared with documented coverage because the measures answer different questions.

Trust and Information

Polling can assess confidence in vaccine safety, trust in health institutions, perceived importance of childhood vaccination, exposure to health information, and support for mandates or policy changes. It cannot replace immunization records, registries, school records, or representative coverage surveys.

Cross-sectional polling may identify an association between attitudes and reported behavior, but it usually cannot show that a particular opinion caused a child to miss a dose. Causal claims require longitudinal studies, natural experiments, or linked administrative data.

Changes in information access and trust may provide context, but they should not be assigned causal importance without supporting evidence. We distinguish original data from news summaries and commentary, then compare claims with primary evidence and transparent methods.

Source Quality and Limitations

We prioritize:

1. Original government surveillance and datasets.
2. Peer-reviewed research.
3. Transparent polling organizations.
4. Established news reporting linked to primary evidence.
5. Commentary and advocacy sources, clearly labeled.

Google News RSS links must be traced to the original article before publication. We record the original headline, author, publication date, and cited data sources.

Several supplied entries contain no usable evidence:

• Source 2, titled “amen thompson,” contains only “100+.”
• Source 4, titled “gianni infantino,” contains only “100+.”
• Source 6, titled “harga dolar hari ini,” contains only “200+.”
• Source 8, titled “tv one,” contains only “100+.”
• Source 10, identified as “detikcom,” contains only “2000+.”

These entries provide no publication date, URL, methodology, or substantive vaccination information and are excluded from the analysis.

The supplied summaries do not provide a numerical estimate of a U.S. childhood vaccination decline. Several sources address attitudes, policy, or international comparisons rather than coverage. The available KFF summary does not include underlying poll findings. Primary data must be obtained and verified before numerical claims are published. A national decline should not be presented as established until coverage data confirm it.

Quality-Control Checklist

Before publication, we:

• Confirm every statistic against the original source.
• Define the vaccine, dose, population, geography, and time period.
• Report percentage-point changes.
• Include uncertainty estimates where available.
• Check for changes in survey design or administrative reporting.
• Separate vaccination coverage from vaccine confidence.
• Separate recommendations from mandates.
• Avoid unsupported causal language.
• Identify missing data and excluded populations.
• Explain whether estimates are nationally representative.
• Date-stamp policy and schedule information.
• Obtain public-health or epidemiology review.
• Link to primary sources rather than relying only on news summaries.

What the Data Can and Cannot Show

Reliable data can show whether measured coverage changed over a defined period, which vaccines or age groups changed most, whether gaps widened or narrowed, and whether vaccination was delayed or missed. They can also show whether a policy change coincided with a shift in reported coverage.

Coverage data alone cannot establish:

• The single cause of a national decline.
• Whether every unvaccinated child reflects intentional refusal.
• Whether a policy change caused a specific coverage change.
• Whether polling predicts individual behavior.
• Whether another country’s approach will work in the United States.
• Whether incomplete reporting represents true nonvaccination.

These questions require additional evidence and stronger research designs.

Conclusion

Tracking U.S. childhood vaccination rates requires consistent definitions, reliable denominators, comparable time periods, and transparent source verification.

Coverage data establish whether a measured outcome changed. Polling can describe vaccine confidence, trust, and policy preferences. Policy reporting can explain changes in recommendations or legal requirements. Neither polling nor policy commentary can substitute for vaccination records or representative coverage estimates.

Incomplete and context-free sources should not influence the analysis. The most credible estimate identifies what was measured, explains how it was calculated, shows where uncertainty remains, and states what cannot yet be concluded.

FAQ

How do you measure a drop in childhood vaccination rates?

Compare coverage across consistent populations and time periods. Define the vaccine, dose count, age group, denominator, geography, and data source. Report the change in percentage points and include uncertainty where available.

Does lower vaccine confidence prove that vaccination rates are falling?

No. Polling measures attitudes, not completed vaccinations. Confidence data provide context, but coverage estimates, registries, surveys, or school records are required to measure actual uptake.

Can a federal vaccine schedule change reduce childhood vaccination rates?

A schedule change may affect recommendations, timing, eligibility, or clinical communication. It does not prove that coverage changed. The effect must be evaluated with comparable data collected before and after implementation.

Why are state vaccination rates difficult to compare?

States may use different reporting systems, school requirements, exemption definitions, age groups, and collection methods. Comparisons require standardized definitions and clear disclosure of methodological differences.

Can Denmark’s vaccine policy be applied directly in the United States?

Not necessarily. The countries differ in health systems, governance, access, school policies, and data infrastructure. International comparisons must account for these differences before supporting policy conclusions.

Why exclude some sources from the analysis?

Entries containing only labels or figures such as “100+” or “2000+” lack the information needed to verify a vaccination claim. A source must provide substantive content, provenance, and methodology before it can support a factual conclusion.