Methods and Limitations
Close Window

Methodology

We estimated the duration of diabetes using data from CDC’s National Health Interview Survey (NHIS) of the National Center for Health Statistics (NCHS). Conducted continuously since 1957, the survey provides information on the health of the civilian noninstitutionalized population of the United States. The multistage probability design of the survey has been described elsewhere.1

Adult respondents are asked whether a health professional had ever told them they had diabetes. To exclude gestational diabetes, women are asked whether they had been told they had diabetes other than during pregnancy. Adults who reported being diagnosed with diabetes are then asked at what age they were diagnosed. Duration was calculated by subtracting age at diagnosis from current age. The analyses were limited to adults aged 18–79 years because age of diagnosis of diabetes was not collected among persons aged 17 years and younger and there were limited data available on age of diagnosis of diabetes among adults aged 80 years and older.

We reported both medians and means, due to the skewness of the distribution of duration of diabetes. Medians of duration of diabetes are presented by sex, age, and race/ethnicity. Persons of Hispanic origin may be of any race. The race groups include persons of both Hispanic and non-Hispanic origin. Boxplots were constructed to display the distribution of duration for each year. Minimum values were always zero (indicating duration is less than one year) and maximum values were calculated as the 75th percentile plus 1.5 times the interquartile range (75th percentile-25th percentile). Three-year averages were used to improve the precision of the annual estimates.
 

Data Limitations

Persons with diabetes may have the disease many years before diagnosis2. Therefore, the NHIS may underestimate the true duration of diabetes. With these cross-sectional data it is impossible to determine the causes of an increase or decrease in duration of diabetes. Duration change can be influenced both by survival and incidence.

 

References

  1. Botman SL, Moore TF, Moriarity CL, Parsons VL. Design and estimation for the National Health Interview Survey, 1995–2004. National Center for Health Statistics, CDC. Vital and Health Statistics 2000;2(130).
  2. Harris MI, Klein R, Welborn TA, Knuiman MW. Onset of NIDDM occurs at least 4–7 years before clinical diagnosis. Diabetes Care 1992;15(7):815–9.