OBJECTIVES:

In modelling exercises, all-cause mortality is generally built up from disease and non-disease specific mortality data, where the latter is country specific general mortality that excludes disease specific mortality to avoid double counting (=non-specific mortality, NSM). The patients with type 1 diabetes (T1D) modelled tend to be young, so a reliable long-term prediction is crucial. In the IQVIA Core Diabetes Model (CDM), 4 mortality prediction approaches can be used: 1) the classical combination of diabetes specific mortality related to cardiovascular and microvascular complications combined with NSM (“non-combined mortality approach”, NCMA), 2) the UKPDS 68 approach applying NSM and its specific equations, 3) UKPDS 82 approach fully based on its mortality risk equations, and the 4) West Australian approach using T1/2D specific equations. UKPDS approaches are determined from patients with type 2 diabetes. We aimed to assess the validity of the 4 approaches in patients with T1D compared to published data.

METHODS:

Using the different mortality approaches, CDM9.5Plus analyses were conducted for a T1D cohort of average age of 27 years and 6 years of diabetes. Several HbA1c levels without progression over time were tested. Pittsburg cardiovascular risk equation was chosen. UK NSM was applied. The predicted 70-year survivals were compared against each other, but also against the findings of 2 publications (Pittsburg-Miller et al 2012; Australia-Huo et al 2016; same age and diabetes duration).

RESULTS:

Both publications reported a median survival between 68 and 73 years. For modelled patients at the age of 70 years, survival was 1)52% 2)60% 3)44% 4)67%, as per mortality approach chosen. The NCMA approach has similar survival compared with the long-term studies.

CONCLUSIONS:

The use of the NCMA approach results in reliable predictions of long-term survival in patients with T1D. It accounts for changes in the management of diabetes and general healthcare.