Baan et al developed predictive models to identify adults to screen for diabetes mellitus. This can help focus testing on high risk patients prior to overt symptoms. The authors are from the University of Rotterdam, Vrije University and other institutions in the Netherlands.
NOTE: This study shares some of the authors from the preceding paper as well as some similar parameters.
Parameters:
(1) age
(2) gender
(3) use of antihypertensive medications
(4) obesity or body mass index (BMI)
(5) physical inactivity
(6) family history of diabetes mellitus
Predictive Model 1 (PM1)
|
Parameters |
Findings |
Points |
|
age in years |
< 55 |
0 |
|
|
55 - 59 |
2 |
|
|
60 - 64 |
4 |
|
|
65 - 69 |
6 |
|
|
70 - 74 |
8 |
|
|
75 - 79 |
10 |
|
|
80 - 84 |
12 |
|
gender and obesity |
female |
0 |
|
|
male |
5 |
|
obesity |
BMI < 30 kg per square meter |
0 |
|
|
BMI >= 30 kg per square meter |
5 |
|
use of antihypertensive drugs |
none |
0 |
|
|
present |
4 |
Predictive Model 2 (PM2)
|
Parameters |
Findings |
Points |
|
age in years |
< 55 |
0 |
|
|
55 - 59 |
1 |
|
|
60 - 64 |
2 |
|
|
65 - 69 |
3 |
|
|
70 - 74 |
4 |
|
|
75 - 79 |
5 |
|
|
80 - 84 |
6 |
|
gender |
female |
0 |
|
|
male |
7 |
|
body mass index (BMI) |
|
<BMI> |
|
family history of diabetes |
none |
0 |
|
|
present |
7 |
|
use of antihypertensive drugs |
none |
0 |
|
|
present |
3 |
|
physical inactivity |
cycles |
0 |
|
|
does not cycle |
10 |
where:
• Transporting PM2 would require finding an alternative measure for physical activity other than bicycling.
• The questionnaires ask if the person is taking antihypertensive drugs, but the question would exclude untreated hypertensives.
total PM1 score =
= SUM(points for the 4 parameters)
total PM2 score =
= SUM(points for the 6 parameters)
Interpretatation:
• minimum PM1: 0
• maximum PM1: 26
• minimum PM2: depends on BMI
• maximum PM2: 73 or more
• The optimum cutoff varied with the population studied.
• In a study using about 8,000 adults from Rotterdam the optimum cutoffs for PM1 was >= 9 and for PM2 was >= 42.
• In a study using about 2,500 adults from Hoorn the optimum cutoffs for PM1 was >= 11 and for PM2 was >= 48.
Performance:
• The sensitivity and specificity varied with different cutoff points (see text) but in general these were fairly mediocre. However, the performance may be adequate for identifying population subsets for further testing.
Specialty: Endocrinology, Clinical Laboratory
ICD-10: ,
