This diabetes diagnostic support tool has been created by MWDH to help clinicians in making the correct diabetes subtype diagnosis. In particular, it aims to highlight an individual's risk of maturity onset diabetes of the young (MODY) and autoimmune diabetes, and will give guidance around next steps in diagnostic testing.
Baseline
If an individual meets the minimum threshold for assessment of a diagnosis of either Autoimmune diabetes (in someone not currently diagnosed with autoimmune diabetes) or Monogenic diabetes (MODY) (diagnosed with diabetes before age 35) then either a risk score or an opportunity to calculate a risk score will appear on the home page (see figure 1). As you add more information into the risk calculator (see figure 2) this risk score may change.
Figure 1: Home screenshot
Machine Learning Models/ Algorithms:
Baseline prediction of risk of either Autoimmune Diabetes or Monogenic Diabetes is based on the models/ algorithms derived from the literature (see references below) that have then been further validated within a Scotland diabetes population.
References:
- The development and validation of a clinical prediction model to determine the probability of MODY in patients with young-onset diabetes.
author: B M Shields and T J McDonald and S Ellard and M J Campbell and C Hyde and A T Hattersley. Diabetologia (2012); 55 (5): 1265-72 - A clinical screening tool identifies autoimmune diabetes in adults.
author: Spiros Fourlanos and Christine Perry and Mark S Stein and Jim Stankovich and Leonard C Harrison and Peter G Colman. Diabetes care (2006); 29(5): 970-5
These models take into information from the existing clinical record such as BMI, age of participant, type of diabetes, age at diagnosis, medication (insulin use), current HbA1C. For additional data fields needed for the models, not readily available in the clinical record, information is collected through the data entry form (below)
Figure 2: Data Entry Form/ risk calculator
Risk Thresholds/ Advice around Additional Diagnostic Tests
This information in models will result in a baseline low, medium and high baseline prediction of risk of MODY and an increased/ not increased risk of auto-immune diabetes.
For MODY prediction, threshold for these categorisations is based on optimal model sensitivity thresholds (calculated from training / validation datasets) and/or thresholds stipulated in the original publications. Cut offs for further clinical tests are those deemed clinically reasonable as per consensus stated in original publications or in validated NHS protocols.
Clinical consensus protocols form the basis of a decision tree matrix that outputs suggestions around the need for additional tests (e.g. antibody, c-peptide testing or genetic testing), and linked advice.
Data/ Calculations
The calculator will retain the last information that has been inputted. Any further updates/ changes to the information supplied in the calculator (including the addition of lab results) will result in a refresh of the calculation and will change the guidance given.
Results/ Advice
The advice page (see screenshot below), will appear once the calculate button is pressed on the calculator, or may link directly from the home page result. The information/ advice will change as information is updated.
Disclaimer
This tool has been developed to support clinicians, but is purely advisory and should not replace clinician decision making. Any incorrect data either entered by the clinicians, or contained within the clinical system may affect the outputs and advice given. It is the responsibility of the clinician using the system to ensure accuracy of data, and to make the final decision on any diagnostic, testing or treatment decision.