NHANES (CDC), derived Conicity Index percentiles
About this reference
FitnessNorms calculated the Valdez (1991) Conicity Index from NHANES waist circumference, standing height, and weight, then calculated weighted empirical P5, P25, P50, P75, and P95 values within sex and five-year age brackets. Adults with complete inputs were included and participants recorded as pregnant were excluded, leaving n=18,716. The pool contains 2015-2016, 2017-March 2020, and 2021-2023 data. MEC weights were multiplied by 2/7.2, 3.2/7.2, and 2/7.2 respectively so each survey year contributes proportionally. CDC documents the 3.2-year pre-pandemic duration and duration-based adjustment; applying the principle across the post-pandemic release is a FitnessNorms choice, not CDC guidance for this exact pool. A comparison with 2021-2023 alone found 13 of 130 cells beyond the 0.02 internal threshold, with a maximum difference of 0.037. The derivation script lists those cells and refuses to write the pooled artifact if a new unlisted breach appears. Survey-design standard errors were not calculated.
Known limitations
- US adults only. These values should not be assumed to represent other countries or waist-measurement protocols.
- This is an internal derivation from public microdata, not a peer-reviewed percentile publication. Weighted empirical quantiles are reported without survey-design standard errors.
- The pool spans the COVID-19 interruption and should be interpreted as a 2015-2023 period reference, not a current-year distribution. CDC cautions analysts about combining pre- and post-pandemic releases.
- Duration-adjusted pooling across all three releases is a FitnessNorms design choice. The 2017-March 2020 file covers 3.2 years rather than two, so equal division by three would misstate its duration.
- Known pregnancies were excluded using RIDEXPRG. That variable identifies pregnancy status only where NHANES reports it and does not exclude postpartum participants.
- RIDAGEYR is top-coded at 80 in the public releases, so the 80+ bracket cannot show variation within older ages.
- Cross-sectional age differences do not show how an individual's Conicity Index changes over time.
- No universal clinical Conicity Index cut-off is established; these are distribution percentiles rather than diagnostic boundaries.