A machine learning tool for identifying patients with newly diagnosed diabetes in primary care.

BACKGROUND AND AIM: It is crucial to identify a diabetes diagnosis early. Create a predictive model utilizing machine learning (ML) to identify new cases of diabetes in primary health care (PHC). METHODS: A case-control study utilizing data on PHC visits for sex-, age, and PHC-matched controls. Stochastic gradient boosting was used to construct a model for predicting cases of diabetes based on diagnostic codes from PHC consultations during the year before index (diagnosis) date and number of consultations. Variable importance was estimated using the normalized relative influence (NRI) score. Risks of having diabetes were calculated using odds ratios of marginal effects (ORME). Four groups by age and sex were studied, age-groups 35-64 years and ≥ 65 years in men and women, respectively. RESULTS: The most important predictive factors were hypertension with NRI 21.4-29.7 %, and obesity 4.8-15.2 %. The NRI for other top ten diagnoses and administrative codes generally ranged 1.0-4.2 %. CONCLUSIONS: Our data confirm the known risk patterns for predicting a new diagnosis of diabetes, and the need to test blood glucose frequently. To assess the full potential of ML for risk prediction purposes in clinical practice, future studies could include clinical data on life-style patterns, laboratory tests and prescribed medication.

Colonization of ß-hemolytic streptococci in patients with erysipelas-a prospective study.

Erysipelas is a common skin infection causing significant morbidity. At present there are no established procedures for bacteriological sampling. Here we investigate the possibility of using cultures for diagnostic purposes by determining the perianal colonization with beta-hemolytic streptococci (BHS) in patients with erysipelas. Patients with erysipelas and a control group of patients with fever without signs of skin infection were prospectively included and cultures for BHS were taken from the tonsils, the perianal area, and wounds. BHS were grouped according to Lancefield antigen, species-determined, and emm-typed. Renewed cultures were taken after four weeks from patients with erysipelas and a positive culture for BHS. 25 patients with erysipelas and 25 with fever were included. In the group with erysipelas, 11 patients (44%) were colonized with BHS, ten patients were colonized in the perianal area, and one patient in the throat. In contrast, only one patient in the control group was colonized (p = 0.005 for difference). All of the patients with erysipelas colonized with BHS had an erythema located to the lower limb. The BHS were then subjected to MALDI-TOF MS and most commonly found to be Streptococcus dysgalactiae. Renewed cultures were taken from nine of the 11 patients with BHS and three of these were still colonized. Streptococcus dysgalactiae colonizes the perianal area in a substantial proportion of patients with erysipelas. The possibility of using cultures from this area as a diagnostic method in patients with erysipelas seems promising.