Home-based anthropometric, blood pressure and pulse measurements in young children by trained data collectors in the National Children's Study.

OBJECTIVE: The current study assessed whether home-based data collection by trained data collectors can produce high-quality physical measurement data in young children. DESIGN: The study assessed the quality of intra-examiner measurements of blood pressure, pulse rate and anthropometric dimensions using intra-examiner reliability and intra-examiner technical error of measurement (TEM). SETTING: Non-clinical, primarily private homes of National Children's Study participants in twenty-two study locations across the USA. SUBJECTS: Children in four age groups: 5-7 months (n 91), 11-16 months (n 393), 23-28 months (n 1410) and 35-40 months (n 800). RESULTS: Absolute TEM ranged in value from 0·09 to 16·21, varying widely by age group and measure, as expected. Relative TEM spanned from 0·27 to 13·71 across age groups and physical measures. Reliabilities for anthropometric measurements by age group and measure ranged from 0·46 to >0·99 with most exceeding 0·90, suggesting that the large majority of anthropometric measures can be collected in a home-based setting on young children by trained data collectors. Reliabilities for blood pressure and pulse rate measurements by age group ranged from 0·21 to 0·74, implying these are less reliably measured with young children when taken in the data collection context described here. CONCLUSIONS: Reliability estimates >0·95 for weight, length, height, and thigh, waist and head circumference, and >0·90 for triceps and subscapular skinfolds, indicate that these measures can be collected in the field by trained data collectors without compromising data quality. These estimates can be used for interim evaluations of data collector training and measurement protocols.

Premature death as the ultimate failure: predictors of death in the US drug user treatment population.

Premature death is the ultimate failure in public health. Failure to complete substance user treatment increases the likelihood of death. Using the five-year follow-up (1990/91-1995/96) of a representative sample of 3,047 clients discharged from drug user treatment, this article documents that deaths after treatment were 4.7 times higher for substance user treatment clients than for the U.S. population matched by age, sex, and race; death rates ranged from 3.5 times as likely for Black males to nine times as likely for White females. Logistic regression models show that completion of treatment is associated with a three-fifths decreased likelihood of death.