Letter to Editor: Positive predictive value of diabetes mellitus risk assessment

Diabetes mellitus (DM) is an important public health challenge [1 ].Different studies predicted that the frequency of diabetic patients will be increased to 642 million throughout the world by 2040 [2]. A notable percentage of diabetic patients are not aware of their disease (approximately 30% in Iran) [3]. Lag in the diagnosis of DM raises the expense of controlling disease and makes the prognosis poorer [4]. It is indicated the importance of diabetic risk assessment as a screening test for high risk populations.  However, most of the screening methods to detect high risk people are invasive [5].  So, detecting population at high risk of developing DM by an easy way that can be applied by health care providers in the health centers may lead to preventive measures of public health magnitude [4]. Griffin et al developed a questionnaire according to the risk factors commonly collected in clinical practice and evaluated the characteristics of the questionnaire. They reported the positive predictive value of 11% for diabetes screening questionnaire in England and Wales [6].In Iran, primary health care providers in rural regions were named “Behvarz”. They were performed diabetes mellitus risk assessment as a screening program in health houses. They worked in “Health House,” which is the small health center in the rural areas of Iran. In the present study, we evaluated DM risk assessment positive predictive value (PPV) on the 30 years and older rural population. The PPV is the probability of the diabetes in a person with a positive risk assessment result ( ) [7]. A cross-sectional study was done in three villages of Bostanabad, one of East Azerbaijan ‘s cities in Iran. Three villages out of more than fifteen hundred villages in the Bostanabad were selected as considered concerns about arranging the population in the villages. In these three villages, Behvarzes performed screening activities for diabetes between March2019 to January2020. Screening for diabetes was regarded for all individuals older than 30 years of age living in chosen villages. Participants were interviewed and asked the presence of risk factors of DM. Risk factors contained family history of DM, overweight or obesity (BMI>25), already detected pre-diabetes. Among persons took part in screening test, those even with one risk factor were regarded positive and they as soon as were visited by a primary care physician for detection of individuals with  unknown DM. The total population of the three villages, namely Saeid abad, Tikmedash and Kordkandy was 5137. More than half of them (57.62%) were more than 30years old in these villages. Out of 1305 people, who had at least one risk factor, 57 persons were diabetics. Accordingly, 57 was true positive .The value of1305 was the sum of true positive and false positive. Calculated positive predictive value was 4.36%. So, the positive predictive value of the risk factor assessment was low. In Conclusion, the ability of the risk factor assessment to predict individuals with DM was quite poor (96% of people with risk factors were not diabetic) and the risk assessment did not work well to identify at risk individuals.