Fay MP, Feuer EJ. Confidence intervals for directly standardized rates: a method based on the gamma distribution. Stat Med; 16: 791-801

National Cancer Institute, Division of Cancer Prevention and Control, Bethesda, Maryland 20892-7354, USA.
Statistics in Medicine (Impact Factor: 1.83). 05/1997; 16(7):791-801. DOI: 10.1002/(SICI)1097-0258(19970415)16:73.0.CO;2-#
Source: PubMed


We offer an approximation to central confidence intervals for directly standardized rates, where we assume that the rates are distributed as a weighted sum of independent Poisson random variables. Like a recent method proposed by Dobson, Kuulasmaa, Eberle and Scherer, our method gives exact intervals whenever the standard population is proportional to the study population. In cases where the two populations differ non-proportionally, we show through simulation that our method is conservative while other methods (the Dobson et al. method and the approximate bootstrap confidence method) can be liberal.

Download full-text


Available from: Michael P Fay, Aug 18, 2014
25 Reads
  • Source
    • "Mortality rates for males and females were calculated separately and were age-standardised to the 2006 NSW population, in 5 year age-groups, using the direct method [20]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Findings regarding the association between overweight and all-cause mortality range from significantly lower to higher risk, compared with body-mass-index (BMI) within the "normal" range. We examined empirically potential methodological explanations for these apparently conflicting results using questionnaire and linked mortality data from 246,314 individuals aged ≥45 years in the Australian 45 and Up Study (11,127 deaths; median follow-up 3.9 years). Hazard ratios (HR) for all-cause mortality associated with BMI were modelled according to different methods of accounting for illness at baseline, finer versus broader gradations of BMI and choice of reference group, adjusting for potential confounders. In analyses using the broad World Health Organization (WHO) categories, the all-cause mortality HR was significantly lower in the overweight category (25.0-29.99 kg/m(2)), than the normal weight (18.5-24.99 kg/m(2)) category. However, in analyses accounting for baseline illness, which excluded those with pre-existing illness at baseline, ever-smokers and the first 2 years of follow up, absolute age-standardised mortality rates varied up to two-fold between finer BMI categories within the WHO normal weight category; rates were lowest at 22.5-24.99 kg/m(2) and mortality HRs increased steadily for BMI above (ptrend<0.02) and below (ptrend<0.003) this reference category. Hence, the breadth of the BMI categories used and whether or not baseline illness is accounted for explain the apparent discrepancies between reported BMI-mortality associations. Using fine BMI categories and the category with the lowest absolute rates as the reference group and accounting for the potential confounding effects of baseline illness is likely to yield the most reliable risk estimates for establishing the independent relationship of BMI to all-cause mortality. These results and those of other studies indicate that a BMI of 22.5-24.99 kg/m(2), not the broad "overweight" category of 25-29.99 kg/m(2), was associated with the most favourable mortality risk.
    PLoS ONE 02/2014; 9(2):e88641. DOI:10.1371/journal.pone.0088641 · 3.23 Impact Factor
  • Source
    • "Again, we used the first-order Taylor series approximation for the variance of the ratio of two random variables to account for sampling variation in the travel estimates. Confidence intervals for the adjusted rates are based on the gamma distribution (Fay and Feuer, 1997). We calculated the variance of the standardized rate ratios following Flanders (1984). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Comparing the injury risk of different travel modes requires using a travel-based measure of exposure. In this study we quantify injury risk by travel mode, age, race/ethnicity, sex, and injury severity using three different travel-based exposure measures (person-trips, person-minutes of travel, and person-miles of travel) to learn how these metrics affect the characterization of risk across populations. We used a linked database of hospital and police records to identify non-fatal injuries (2001-2009), the Fatality Analysis Reporting System for fatalities (2001-2009), and the 2001 Wisconsin Add-On to the National Household Travel Survey for exposure measures. In Wisconsin, bicyclists and pedestrians have a moderately higher injury risk compared to motor vehicle occupants (adjusting for demographic factors), but the risk is much higher when exposure is measured in distance. Although the analysis did not control for socio-economic status (a likely confounder) it showed that American Indian and Black travelers in Wisconsin face higher transportation injury risk than White travelers (adjusting for sex and travel mode), across all three measures of exposure. Working with multiple metrics to form comprehensive injury risk profiles such as this one can inform decision making about how to prioritize investments in transportation injury prevention.
    Accident; analysis and prevention 08/2013; 60C:103-112. DOI:10.1016/j.aap.2013.08.005 · 1.65 Impact Factor
  • Source
    • "To account for difference of age at the diagnosis of sIA disease between the unruptured and ruptured groups, we standardized the age to the joint age distribution of both age-groups. Standardized rates and CIs were calculated according to the methodology of Fay and Feuer (35), as the occurrences of diabetes are low in different age-groups. The incidences of ADM in the ruptured and unruptured patients were calculated for 10-year age intervals, and the CIs in each age group were calculated as exact central Poisson CIs (36). "
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE Type 2 diabetes is a risk factor for other forms of stroke, but its association with subarachnoid hemorrhage (SAH) from ruptured saccular intracranial aneurysm (sIA) has remained unclear.RESEARCH DESIGN AND METHODS Kuopio Intracranial Aneurysm Database ( includes all ruptured and unruptured sIA cases from a defined catchment population in eastern Finland since 1980. We compared the age-adjusted incidences of type 2 diabetes in 1,058 ruptured and 484 unruptured sIA patients during 1994-2008, using the national registry of prescribed medicine purchases.RESULTSOf the 1,058 ruptured sIA patients, 43% were males and 57% females, with a median age at rupture of 51 and 56 years, respectively. From 1994 to 2008 or until death, 9% had been prescribed antidiabetes medication (ADM) with a median starting age of 58 years for males and 66 years for females. Of the 484 unruptured sIA patients, 44% were males and 56% females, with a median age at the diagnosis of 53 and 55 years, respectively, and 9% had used ADM, with a median starting age of 61 years for males and 66 years for females. The incidence of type 2 diabetes was highest in the age-group 60-70 years, with no significant differences between the ruptured and unruptured sIA patients.CONCLUSIONS Our study suggests that type 2 diabetes does not increase the risk of rupture of sIA, which is by far the most frequent cause of nontraumatic SAH.
    Diabetes care 03/2013; 36(7). DOI:10.2337/dc12-1048 · 8.42 Impact Factor
Show more