Simple online tool predicts type 2 diabetes risk

Medical Tribune May 2009 P14
David Brill

A simple new tool could allow GPs to identify patients at high risk of developing type 2 diabetes without the need for laboratory tests.

The QDScore, developed with electronic medical record data from some 2.5 million patients, is quick and easy to use and is freely available online.

It is the first prediction algorithm to include ethnicity and social deprivation alongside conventional diabetes risk factors, and is intended for routine use in primary care.

The QDScore performed well in a recent study – accurately predicting 10-year diabetes risk in a diverse patient population drawn from 176 UK general practices. [BMJ 2009 Mar 17;338:b880]

“There is good evidence that lifestyle changes and medical intervention at an early stage can prevent type 2 diabetes in up to two-thirds of high-risk cases and that early diagnosis is likely to improve outcomes,” said lead researcher Professor Julia Hippisley-Cox, professor of clinical epidemiology and general practice at the University of Nottingham, UK.

“As the number of people diagnosed with diabetes in the UK continues to rise, this new algorithm will be an invaluable tool to help doctors identify those at greatest risk who are most likely to benefit from interventions.”

The QDScore calculates risk using age, sex, ethnicity, family diabetes history, personal cardiovascular disease history, hypertension medications, steroid usage, smoking status and body mass index. It was derived from a cohort of more than 2.5 million 25 to 79 year-olds – over 78,000 of whom developed type 2 diabetes over 10 years of follow up.

There were considerable differences in diabetes risk between ethnic groups. Bangladeshi men were the most likely to develop diabetes, with an adjusted hazard ratio (HR) of 4.53 as compared to Caucasian men. The subsequent male risk hierarchy was: Pakistani, Indian, Other Asian, Black African, Chinese and Black Caribbean (adjusted HRs 2.54, 1.93, 1.89, 1.67, 1.41 and 0.80 respectively).

The risk patterns were similar for women, although Chinese ethnicity conferred a greater risk than in men. Bangladeshi women were the most likely to develop diabetes, followed by those of Pakistani, Chinese, Indian, Other Asian, Black African and Black Caribbean origin (adjusted HRs 4.07, 2.15, 1.96, 1.71, 1.26, 0.81 and 0.80, respectively).

The QDScore was tested on records from more than 1.2 million patients – over 37,500 of whom developed diabetes. It scored highly on validation statistics, out-performing the Cambridge risk score in all domains for both men and women. It is available at: http://www.qdscore.org/.