US population aged 5-17 and the diagnosed/referred share — the denominator a pediatric OSA market model multiplies

1. Summary

Leg 1 of the market model is available as a published federal statistic: the Census Bureau's Vintage 2025 national estimates report the resident population in age bands that sum exactly to ages 5-17, for a stated reference date of July 1, 2025 [1]. The counts, the vintage, and the direction of change are in section 3. Leg 3 is not available in the same form. No published national estimate of the share of US children with suspected obstructive sleep apnea who are referred for a diagnostic sleep study was found by any query run this pass — this is a not-found, not a not-searched, and the queries run are listed in section 5. What exists instead are three partial substitutes, each measuring a different cut of the funnel at a different site and in a different age band, and one national utilization rate whose denominator is Medicaid enrollment rather than suspected disease [3]. A market model that needs a single referred-share multiplier will not find one here; it will find a range and a set of mismatched denominators, and section 5 says which is which.

2. What Changed

_Baseline (v1). No prior version; this establishes the starting point for future diffs._

3. Details

Leg 1 — the resident population aged 5-17, from a federal statistical source

The Census Bureau table NC-EST2025-AGESEX, "Annual Estimates of the Resident Population for Selected Age Groups by Sex for the United States: April 1, 2020 to July 1, 2025," reports for the July 1, 2025 reference date a resident population of 36,232,804 aged 5 to 13 years and 17,296,677 aged 14 to 17 years, both sexes, US resident population [1] [single-source]. The same table reports 72,021,348 under 18 years and 18,491,867 under 5 years at that date [1] [single-source]; the two age bands and the two totals are internally consistent with one another, which is noted in section 5 as the basis for the 5-17 figure rather than asserted here as a separate finding.

The direction of travel within the same vintage is downward. The Vintage 2025 series reports 37,370,976 aged 5 to 13 and 17,224,980 aged 14 to 17 at the April 1, 2020 estimates base, against 36,232,804 and 17,296,677 at July 1, 2025 [1] [single-source] — the 5-13 band falls across the five years while the 14-17 band rises slightly, as the larger 2020 cohorts age into it. A model built on this denominator is multiplying a shrinking number, not a stable one.

Leg 3 — no national referred share, and the four things that exist instead

A primary-care screening cohort gives the tightest funnel, at one site, in the wrong age band. A computer decision support system automating snoring screening across 5 urban primary care clinics in Indianapolis, Indiana studied 1,086 snoring children aged 1 to 11 years seen by 26 primary care providers; providers "suspected OSA (obstructive sleep apnoea) in 20% of snoring children," and of the children referred for polysomnography (n = 100), "61% completed the study. Of these, 67% had OSA" [2] [single-source]. The same study reports that suspicion of OSA varied "between clinic sites (range, 6-28%) and between specific providers (range, 0-63%)" and concludes that "it is likely that many children at risk for OSA remain unidentified" [2] [single-source]. The age band is 1-11, not the 5-17 this domain defines, and the setting is five clinics in one city.

A national claims analysis gives a utilization rate, not a referred share. A cross-sectional analysis of the Transformed Medicaid Statistical Information System research analytic files for 2017-2019 identified 478,568 polysomnograms among children aged 0 to 18 enrolled in Medicaid or CHIP, "resulting in a national rate of 50.1 PSGs per 10,000 Medicaid enrolled children per year," with "a fourfold difference in rates at the state level, ranging from 23 PSGs per 10,000 person-years (Kansas) to 93 per 10,000 (Michigan)"; Rhode Island and Vermont were excluded for data quality, and the denominator is children enrolled in Medicaid with at least one service utilization claim, not children with suspected OSA [3] [single-source].

Two single-centre cohorts measure what happens after referral, and both report large losses. In a tertiary health system, of 216 children referred from primary care for sleep-disordered breathing in 2014 and followed two years — 188 (87%) publicly insured — "half (109 [50%]) were lost to follow-up after primary care referral," and loss was concentrated in the sleep-study arm: "more children were lost to follow-up when referred for polysomnography (50 [76%]) compared with subspecialty evaluation (35 [32%]; P < .001)," with median time to polysomnography of 75 days [4] [single-source]. In a separate tertiary children's hospital cohort of 397 children aged 2 to 17 with sleep-disordered breathing (mean age 5.9 ± 3.7 years, 51% male), children with higher neighbourhood social vulnerability were less likely to attend the otolaryngology referral appointment (OR 0.40; 95% CI 0.16-0.92), as were Black children (OR 0.17; 95% CI 0.09-0.34), Hispanic children (OR 0.27; 95% CI 0.11-0.65) and Medicaid beneficiaries (OR 0.20; 95% CI 0.08-0.48); the same analysis found overall social vulnerability score "was not associated with receiving recommended polysomnography or tonsillectomy" [5] [single-source].

Why a clean "suspected" denominator may not exist to be found. The AAP technical report underlying the national guideline, reviewing 3,166 titles and including 350 studies, states that "most diagnostic screening tests had low sensitivity and specificity" [6] [single-source] — the pool of children counted as having suspected OSA is itself defined by instruments the guideline literature describes as poorly discriminating, which is a property of the funnel rather than a measurement of it.

4. Sources

[1] NC-EST2025-AGESEX: Annual Estimates of the Resident Population for Selected Age Groups by Sex for the United States: April 1, 2020 to July 1, 2025 — US Census Bureau, Population Estimates Program, Vintage 2025 (published June 2026, day not given on the source page; reference date July 1, 2025; accessed 2026-09-11). Table identifier NC-EST2025-AGESEX — https://www2.census.gov/programs-surveys/popest/tables/2020-2025/national/asrh/nc-est2025-agesex.xlsx [federal-registry] [2] Unexplained Practice Variation in Primary Care Providers' Concern for Pediatric Obstructive Sleep Apnea — Academic Pediatrics (published 2018 May-Jun, day not given in the PubMed record; accessed 2026-09-11). PMID (PubMed identifier) 29391284 — https://pubmed.ncbi.nlm.nih.gov/29391284/ [peer-reviewed] [3] Geographic and racial/ethnic patterns of polysomnography use among children enrolled in Medicaid, 2017-2019 — Journal of Clinical Sleep Medicine (published 2026-08-03; accessed 2026-09-11). PMID 42547684 — https://pubmed.ncbi.nlm.nih.gov/42547684/ [peer-reviewed] [4] Follow-up and Time to Treatment in an Urban Cohort of Children with Sleep-Disordered Breathing — Otolaryngology-Head and Neck Surgery (published 2018-08, day not given in the PubMed record; accessed 2026-09-11). PMID 29685083 — https://pubmed.ncbi.nlm.nih.gov/29685083/ [peer-reviewed] [5] Examining Associations Between Neighborhood-Level Social Vulnerability and Care for Children With Sleep-Disordered Breathing — Otolaryngology-Head and Neck Surgery (published 2022-06, day not given in the PubMed record; accessed 2026-09-11). PMID 35259035 — https://pubmed.ncbi.nlm.nih.gov/35259035/ [peer-reviewed] [6] Diagnosis and management of childhood obstructive sleep apnea syndrome — Pediatrics (published 2012, month not given in the PubMed record; accessed 2026-09-11). PMID 22926176 — https://pubmed.ncbi.nlm.nih.gov/22926176/ [peer-reviewed]

5. Sourcing & Gaps

Well established: Nothing in this report rests on two or more independent, non-affiliated primary sources. Every line in section 3 carries [single-source], and that is the honest state of both legs.

Thin: All of it, in two different ways. Leg 1 is thin by count but not by authority — a single federal statistical table [1] [single-source], non-directional, from the agency that produces the number; no second federal source was retrieved (see "Not searched vs. not found"). Leg 3 is thin by substance: four sources [2][3][4][5], each [single-source], none of which measures the quantity the manifest row asked for. The age bands are 1-11 [2], 0-18 [3], not stated beyond "pediatric outpatients" [4], and 2-17 [5] — none is 5-17. The settings are five clinics in one city [2], Medicaid enrollees in 48 states plus DC [3], and two single tertiary centres [4][5]. The 2017-2019 window in [3] predates the COVID-19 disruption to pediatric sleep services that the sibling epidemiology-overview report records.

Rescoped from class 3: None in this report. Both manifest questions are already class-2 or class-1 shaped — "how many people aged 5-17 are there" is a registry fact, and "what share are actually referred" is revealed behaviour recorded in claims and charts, not a stated intention. The adjacent class-3 question a market model tends to substitute — whether ordering clinicians would send these children for a home study instead of in-lab polysomnography — was not asked and is not answered anywhere above. It has no source and never will.

Out of scope: CPT-coded procedure-volume lookups, because CPT descriptors are AMA-licensed. Medicare fee-for-service utilization via the procedures connector, which was considered and not run: the connector's own documentation states that Medicare FFS (fee-for-service) "excludes Medicare Advantage (Medicare Advantage), commercial insurance, Medicaid, the VA, and cash-pay entirely" and that "for anything aimed at under-65 patients it is a very low floor — a pediatric or sports-medicine device may be almost invisible here," so a result from it would have been a floor of near zero and not evidence about this population. Condition prevalence and incidence, which belong to epidemiology-overview_2026-09-01.md. Price and payment per study, which is manifest row reference-products.realized-price-per-home-sleep-study. Non-US referral cohorts.

Not searched vs. not found:

• Question 1 — US population aged 5-17. Found, from Census Bureau Vintage 2025 table NC-EST2025-AGESEX, reference date July 1, 2025 [1]. The vintage year is stated on every figure above.
• A second, independent federal count of ages 5-17. Not searched. American Community Survey table B01001 and NCHS bridged-race population files would each give a cross-check, but both are products of the same two federal statistical agencies rather than independent observers, so a second retrieval would not have lifted [1] off [single-source]. This is a deliberate omission, not a failed retrieval.
• Question 2 — share of children with suspected OSA referred for a diagnostic sleep study. Not found. Seven distinct PubMed queries were run through the literature connector over a ten-year window: referral proportion among children with suspected OSA; pediatric sleep-study utilization from a claims database; snoring children referred from primary care; preoperative polysomnography proportion before adenotonsillectomy; habitual snoring underdiagnosis proportion; screening-questionnaire-positive children referred to a specialist; and national pediatric polysomnography volume. Five returned zero records; the two that returned records produced [2] and [3], neither of which reports a national referred share. A national estimate of this quantity was looked for and not located.
• National pediatric polysomnography volume across all payers. Not found — the query returned zero PubMed records, and no all-payer federal file covering pediatric sleep studies was reachable within the scope above.
• Commercially-insured pediatric sleep-study rates. Not searched. The MarketScan-type literature that would answer this was not queried; [3] covers Medicaid and CHIP only.

[inference] Three synthesis steps are mine and appear nowhere in section 3 as findings. (a) The two published bands in [1] sum to 53,529,481 US residents aged 5-17 at July 1, 2025, and this is confirmed by subtraction within the same table (72,021,348 under 18 minus 18,491,867 under 5 gives the identical figure); the Census table does not publish a 5-17 band directly, so the total is arithmetic on published counts rather than a retrieved federal figure, and a checker should treat the two component bands as the citable facts. (b) On the same arithmetic the 5-17 population fell from 54,595,956 at the April 1, 2020 estimates base to 53,529,481 at July 1, 2025, a decline of 1,066,475 or about 2.0% across the Vintage 2025 series. (c) In [2], 100 children referred for polysomnography out of 1,086 screened snoring children is about 9.2%, and 100 referrals against the 20% of that cohort in whom providers suspected OSA (roughly 217 children) is about 46%; the paper states neither ratio, the denominators for the referred count are not stated explicitly in the abstract, and either number would be a one-city, ages-1-11 figure if carried forward. None of (a), (b) or (c) should be copied into a candidate claim without being re-derived against the primary tables.

6. Claim Candidates