FINDRISC was developed in Finland from population based epidemiological studies and provides a total score that predicts future diabetes risk . The tool is a one page form that contains eight questions with weighted categorical answers about age, BMI, waist circumference, physical activity, daily consumption of fruits, berries or vegetables, history of antihypertensive drug treatment, history of high blood glucose, and family history of diabetes. The total risk score is a sum of the individual scores from each of the eight questions, with totals ranging from zero to 26 .
During recruitment for the GGT DPP, potential participants were asked to fill in the anthropometric information on the FINDRISC from self-report. This included data for questions about height, weight and waist circumference category. Study nurses were available to provide assistance if necessary, but were not required to perform anthropometric measures. However, after contacting the study nurses from all sites, one site was excluded from the analysis between self-report and measured data as the study nurse measured height, weight, and waist circumference during recruitment.
The study nurse was required to calculate BMI from self-reported height (m) and weight (kg), and then the total risk score. They were provided with a handheld BMI calculator to determine BMI and then record an individual’s BMI category. The BMI calculator has measurement markings for height between 1.4 and 2.1 m with 1 cm increments, weight between 40 and 120 kg with 1 kg increments, and BMI between 10 and 60 kg/m2 with 1 kg/m2 increments. Conversion charts for height from feet and inches (5’ 0” to 6’ 6”) to cm and weight from stone (8 to 26) to kg were also provided (all to 0 decimal places).
BMI categories were based on the World Health Organisation’s International Classifications, with normal BMI (<25.0 kg/m2) scored ‘0’, overweight (BMI 25.0-29.9 kg/m2) scored ‘1’, and obese (BMI ≥30.0 kg/m2) scored ‘3’ .
Waist circumference categories were different for men and women based on Lean’s (1995) ‘waist action levels’ that correspond to BMI categories of normal, overweight and obese . For both men and women normal categories (men <94.0 cm, women <80.0 cm) were scored ‘0’, overweight categories (men 94.0 cm-101.9 cm, women 80.0 cm-87.9 cm) were scored ‘3’, and obese categories (men ≥102.0 cm, women ≥88.0 cm) were scored ‘4’.
The recruitment criteria for the GGT DPP included a total score of 12 or more on the FINDRISC (moderate to high risk). This score was selected to be comparable with the GOAL Lifestyle Implementation Trial recruitment criteria . Individuals with cancer, recent myocardial infarction or stroke, cognitive impairment, substance abuse, pregnancy or previous type 2 diabetes diagnosis were excluded from the study .
Individuals recruited to the study attended baseline clinical tests within three weeks of being screened with the FINDRISC. Objectively measured height, weight, and waist circumference were among the anthropometric measures taken by specially trained study nurses using the European Health Risk Monitoring protocol . Specifically, waist circumference was measured with a non-stretched plastic tailor's measuring tape. Participants were measured at a level midway between the lower rib margin and the iliac crest with a tape around the body in a horizontal position with their clothes removed. Participants stood normally with their feet approximately 12–15 cm apart, so their weight should be equally distributed between each leg. Additionally, participants were asked to breathe normally with the reading of the measurement being taken at the end of gently exhaling.
Height and waist circumference were measured in cm to 1 decimal place, and weight was measured in kg to 1 decimal place. For waist circumference, the mean of two measurements was used.
Data were analysed in PASW Statistics (SPSS) version 18. Absolute numbers with percentages are presented in the results section.
After considering the time between recruitment and baseline clinical testing, daily fluctuation of weight , and potential waist circumference measurement error, borderline categories were created for BMI and waist circumference. The two extra categories of BMI were ‘borderline normal/overweight’ (borderline N/OW, 25 kg/m2 ± 2 kg) and ‘borderline overweight/obese’ (borderline OW/OB, 30 kg/m2 ± 2 kg). A 2 kg allowance has been previously used by Rossouw et al. (2000) in determining the accuracy of self-reported weight in overweight or obese subjects . Similarly, borderline categories were created for waist circumference that allow for a ± 2 cm margin of error.
Paired T-test was used to determine the difference between self-reported and measured BMI. Spearman-rho was used to calculate the correlation between original risk score total and risk score error in those who attended baseline clinical testing.