Table 3.
Marginal distributions of stratification variables under random sampling (RS) and maximum entropy sampling (MES) designs when using only EHR data and when using combined EHR and USC data. For a sample on 90 000 households, percentages of non-missing values are reported by population (pediatric, adult)
| Adult Sites |
Pediatric Sites |
|||||
|---|---|---|---|---|---|---|
| EHR Only |
EHR + USC |
EHR Only |
EHR + USC |
|||
| RS | RS | MES | RS | RS | MES | |
| Age | ||||||
| Low age group | 22.9 | 22.9 | 43.8 | 68.7 | 68.7 | 56.7 |
| High age group ears | 77.1 | 77.1 | 56.2 | 31.3 | 31.3 | 43.3 |
| Gender | ||||||
| Female | 58.3 | 58.3 | 52.9 | 47.6 | 47.6 | 49.7 |
| Male | 41.7 | 41.7 | 47.1 | 52.4 | 52.4 | 50.3 |
| Race | ||||||
| White | 87.1 | 87.6 | 34.3 | 66.0 | 69.3 | 33.0 |
| African American | 5.7 | 5.6 | 18.3 | 19.3 | 17.7 | 22.5 |
| Asian | 2.4 | 2.3 | 16.1 | 2.9 | 2.6 | 14.7 |
| AI/AN | 0.7 | 0.6 | 7.1 | 0.2 | 0.1 | 2.5 |
| NH/PI | 0.3 | 0.2 | 4.9 | 0.1 | 0.1 | 1.8 |
| Other | 3.8 | 3.6 | 19.2 | 11.6 | 10.2 | 25.5 |
| Missing | 14.1 | – | – | 13.1 | – | – |
| Ethnicity | ||||||
| Non-Hispanic/Latino | 95.9 | 95.6 | 69.3 | 93.3 | 93.9 | 69.5 |
| Hispanic /Latino | 4.1 | 4.4 | 30.7 | 6.7 | 6.1 | 30.5 |
| Missing | 19.3 | – | – | 11.8 | – | – |
| Education | ||||||
| < HS | – | 1.0 | 11.9 | – | 1.2 | 13.6 |
| HS + some college | – | 76.8 | 54.9 | – | 72.4 | 48.9 |
| ≥ Bachelor’s | – | 22.2 | 33.2 | – | 26.4 | 37.6 |
| Rurality | ||||||
| Suburban/Urban | 52.0 | 62.5 | 70.7 | 63.9 | ||
| Rural | – | 48.0 | 37.5 | – | 29.3 | 36.1 |
Low age group (< 12 in pediatric sites, < 35 in adult sites), high age group (≥ 12 in pediatric sites, ≥ 35 in adult sites), AI/AN = American Indian/Alaska Native, NH/PI = Native Hawaiian/Pacific Islander, HS = high school.