CMS set an 80 percent participation goal for Medicaid child health screenings in 1990 with a 1995 deadline — the national participation rate was 51 percent in fiscal 2023

What is happening?

Federal law makes preventive health screening a mandatory Medicaid benefit for enrolled children from birth through age 20. The benefit is called EPSDT — Early and Periodic Screening, Diagnostic, and Treatment — and it is defined at 42 U.S.C. 1396d(r). The Omnibus Budget Reconciliation Act of 1989, Pub. L. 101-239 section 6403(c), directed the Secretary of Health and Human Services to set participation goals for each state. In 1990 CMS set the participant ratio goal at 80 percent, set an identical 80 percent goal for the screening ratio, and set the deadline at federal fiscal year 1995.

The participant ratio measures the share of eligible children who received at least one recommended well-child screening in the year. GAO reported that ratio nationally as 64 percent in fiscal 2010, then 61, 62, 63, 59, 58, 58 and 59 percent through fiscal 2017. In June 2025 GAO reported 51 percent for fiscal 2023, down from the 59 percent it had reported for fiscal 2019.

A second requirement is stricter and is not a percentage goal at all. Since 1998 the State Medicaid Manual has required blood lead testing for 100 percent of children at 12 months and again at 24 months. GAO reported in June 2025 that CMS still does not have accurate and complete national data with which to determine whether the states comply.

The shortfall has been documented across three decades by two federal oversight bodies. HHS Office of Inspector General reported in 1997 that 60 percent of children in managed care did not receive required screenings. In May 2010 the same office examined a nine-state sample and found that 76 percent of children, about 2.7 million, missed at least one required medical, vision or hearing screening, and that 41 percent received no medical screening at all.

Whose problem is this?

The body that set the goal in 1990 is the same body that receives the annual reports showing whether it was met.

Where does this problem end?

The boundary sits at delivery rather than eligibility, because these children are already enrolled and the benefit is already mandatory under federal law.

What is missing is not a rule and not a target but a routine comparison between the target and the result.

What is the state now, and what should it be?

Now

Should be

The target state here does not have to be inferred, because the federal agency published both a number and a deadline more than three decades ago.

No year for which the sources opened here report a national figure reaches that number, and the deadline that governed it passed in 1995.

How big is it?

About 14 million children in a single year. Fiscal 2017 is the one year for which the sources opened here give a rate and a count from the same report and the same fiscal year: about 20.2 million children, or 59 percent, received at least one recommended well-child screening. The remainder of that denominator is roughly 13.7 to 14.4 million children who did not.

Participation counts a child as reached after a single screening, even when the recommended schedule calls for several.

That makes the figure a floor rather than a ceiling. In the nine-state sample HHS OIG examined for 2007, 59 percent of the children who did receive a medical screening were missing at least one of the five statutory components, most often a laboratory test such as blood lead. It also excludes lead testing entirely, because the standard there is 100 percent and no national compliance rate exists to subtract from it.

The most recent national rate is lower than the one used above. GAO reported 51 percent for fiscal 2023, but that report does not publish the eligible denominator for that year, so no count is derived from it here. Multiplying a fiscal 2023 rate by a fiscal 2022 enrollment total would combine two different populations and produce a number that no source supports.

Under what conditions does it arise?

1. The goal exists and nothing compares performance to it on a schedule. In August 2019 GAO made six recommendations to HHS. The department concurred with three and did not concur with three, among them the recommendation that CMS regularly assess state performance against the goals CMS had itself set.

2. Delivery runs through contracts, and oversight of those contracts was not exercised in the states GAO sampled. About 85 percent of enrolled children received services through managed care in fiscal 2022. In April 2024 GAO found that none of the four states in its sample reviewed a representative sample of the EPSDT service denials made by managed care plans.

3. The measurement basis changed part way through the series. By fiscal 2023, 31 states, the District of Columbia, Guam, Puerto Rico and the US Virgin Islands were producing the Form CMS-416 report from federal T-MSIS data rather than from their own records. A series that runs from 1990 to 2023 therefore does not rest on one method throughout.

4. The strictest requirement has no national measurement at all. Blood lead testing is an absolute standard rather than a percentage goal, and GAO reported in both 2019 and June 2025 that CMS lacks accurate and complete national data on compliance.

A requirement that nobody measures nationally produces no number that anyone can be shown to have missed.

5. Where money is attached, the amount attached is small. As GAO reported in June 2025, North Carolina, Ohio and Washington ran withhold or incentive arrangements tied to child screening performance in 2024 and 2025, with 1 to 3 percent of capitation payments linked to them. GAO also reported that performance data for the North Carolina arrangement was not expected to be available until 2026.

What has been tried?

Two directions have run in parallel for three decades — measure the shortfall, and recommend that someone act on the measurement. The first has produced a continuous record. The second has produced recommendations that were partly accepted, partly declined and, where accepted, in one case still open eight years later.

What was found?

Why is it still unsolved?

Enforcement absent — the goal, the deadline and the reporting form all exist, and the step that would connect a missed goal to a consequence does not.

The federal government set a number and a date, collected the data showing the number was not reached, and did not build the routine that compares one to the other. In August 2019 GAO recommended that CMS regularly assess state performance against the goals it had set, and HHS did not concur with that recommendation.

The second part is that delivery and oversight sit at different levels. About 85 percent of the children are served under contracts between state agencies and managed care plans. In April 2024 GAO examined four states and found that none of them reviewed a representative sample of the EPSDT service denials made by the plans they contracted with, and as of March 2025 the two recommendations GAO issued on that point had not been implemented.

The third part is that a slow decline does not arrive as an event. The rate fell from 64 percent in fiscal 2010 to 59 percent in fiscal 2017, before the pandemic, and to 51 percent by fiscal 2023. Each step is a point or two, and a goal that has been missed every year for three decades produces no year in which it is newly missed.

There is movement on the record, and it should be read as the response of the agencies rather than as silence. Congress required a review of EPSDT in managed care in the Bipartisan Safer Communities Act of 2022. CMS issued best practice guidance to state health officials in September 2024 and engaged an outside research contractor whose study runs through fiscal 2027. Three states have tied a small share of managed care payment to screening performance. What none of the sources opened here attaches to any of this is a date by which the participation ratio has to be higher.

What observation would mean it is solved?

Candidates — a. the national participant ratio reported on Form CMS-416 reaches 80 percent, the level CMS set in 1990. b. CMS publishes an assessment comparing each state to that goal on a stated schedule. c. CMS publishes a national blood lead testing compliance rate measured against the 100 percent requirement.

a alone is weaker than it looks. The participant ratio counts a child as reached after one screening in a year, so it can rise while the recommended schedule is still not delivered. The measure also changed basis part way through the series, and a rise that coincides with the switch to T-MSIS reporting could not be read as a change in delivery without a separate comparison.

b alone counts process. An assessment that is published but carries no consequence changes the record without changing the screening.

Coverage on a report is not the same as a screening reaching the child who needs it.

c is narrow but unambiguous. The lead requirement is absolute rather than proportional, so there is no threshold that can be partly met. What publishing a rate would end is the state of not knowing, which is where GAO has left that item since 2019. The three have to be read together, and a is only meaningful alongside a measure of whether the full recommended schedule was delivered.

What is it connected to?

Fills with researchchildhood lead exposure policy, oversight of managed care contracts for other Medicaid populations, school-based health services, the federal share of special education costs, and comparable child preventive care obligations in other countries. Relation type and evidence grade were not established in this round.

What these sources do not say

• How many children missed a screening in fiscal 2023. GAO published the 51 percent rate for that year without the eligible denominator behind it. The 2019 report gave both figures for fiscal 2017; the 2025 report gives only the rate.
• The national blood lead testing compliance rate. No source opened here gives one. GAO reported in 2019 and again in June 2025 that the CMS data are not accurate and complete enough to determine whether the states comply.
• State-level participation for fiscal 2023. The most recent state distribution found in this round is for fiscal 2019, ranging from about 87 percent in Hawaii to 36 to 38 percent in Alaska and South Dakota. No source opened here gives a later one.
• Why the rate declined. The fall from 64 percent in fiscal 2010 to 59 percent in fiscal 2017 predates the pandemic. No source opened here carries a causal analysis of that decline from CMS or GAO. Pandemic effects are described only for the period after 2020.
• Whether the series is comparable across the T-MSIS transition. By fiscal 2023 a majority of reporting jurisdictions had changed the basis of the Form CMS-416 report. No source opened here quantifies what that change does to the national rate, so the 1990 goal and the 2023 figure are not stated by any source to rest on the same measurement.
• How complete screenings are today. The finding that 59 percent of screened children missed at least one statutory component comes from a nine-state sample of 2007 records. No source opened here gives a later national figure for component completeness.
• What CMS says about the 51 percent figure. No source opened here carries a published CMS response to the fiscal 2023 rate. The positions of the agencies that are on the record are these: HHS concurred with three of the six 2019 recommendations and did not concur with three, CMS issued best practice guidance to state health officials in September 2024, and CMS engaged an outside research contractor to conduct the review Congress required, with the study running through fiscal 2027.

See the evidence

No primary CMS document was read here. Eight URLs were opened directly and all eight are federal oversight reports, a congressional advisory page, a statutory text or a reanalysis of federal aggregates. The two CMS surfaces that would carry the underlying Form CMS-416 tables and the 2024 guidance letter returned HTTP 403 to automated fetch and are recorded as unconfirmed rather than filled in with a substitute address. Four further items are cited only through footnotes in reports that were opened, and they are marked as such rather than presented as directly consulted.

Where the opened sources overlap on a value they agree, and no conflicting figure was found for any number used above: the 80 percent goal, the 1995 deadline and the 100 percent lead testing requirement appear in the 2019 report, and the fiscal 2019 and fiscal 2023 rates appear in the 2025 report, which states the fiscal 2019 value as the one it had reported earlier. The one break in the record is not a disagreement between sources but a change of measurement basis — by fiscal 2023 a majority of reporting jurisdictions produced Form CMS-416 from federal T-MSIS data rather than from their own records, and no opened source states what that change does to comparability across the series. That break is why the 1990 goal and the 2023 figure are presented above as published values rather than as two points on one continuous measure. This is a Path A output, a research-based definition, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range 13,700,000–14,400,000 As of federal fiscal year 2017

Sensitivity The interval width comes only from rounding in the two published figures. Both are drawn from the same report and the same fiscal year, so the interval carries no allowance for spread in the underlying data and is not a confidence interval. The count is a floor rather than a ceiling in three ways. First, the participant ratio counts a child as reached after a single screening even when the recommended schedule calls for several, and in a nine-state sample of 2007 records HHS Office of Inspector General found that 59 percent of the children who did receive a medical screening were missing at least one of the five statutory components. Second, it excludes blood lead testing, for which the standard is 100 percent of children at 12 and 24 months and for which no national compliance rate is published. Third, it is a fiscal 2017 figure and the national rate has since fallen to 51 percent for fiscal 2023, but the report carrying that rate does not publish the eligible denominator for that year, so no current count is derived here and the fiscal 2023 rate is not multiplied by the fiscal 2022 enrollment total of 33 million, which is a different measure of a different population. In the opposite direction the count overstates the shortfall for any child who received a recommended screening that the Form CMS-416 report did not capture, and by fiscal 2023 a majority of reporting jurisdictions had changed the basis of that report from their own records to federal T-MSIS data, so capture is not stated by any opened source to be constant across the series.

Regional breakdown The rate and count pair used for this chain is national only. The most recent state-level distribution found in this round is for fiscal 2019, two years later than the pair, and it runs from about 87 percent in Hawaii down to 36 to 38 percent in Alaska and South Dakota, so state rates are far from uniform. Splitting the national total by state child population would be proportional allocation, which this repository does not accept, and it would be wrong in both directions at once because the states with the largest child populations are not the states with the lowest participation rates.