2018
DOI: 10.1530/eje-17-0921
|View full text |Cite
|
Sign up to set email alerts
|

HAPT2D: high accuracy of prediction of T2D with a model combining basic and advanced data depending on availability

Abstract: Our models provide an estimation of patient's risk over time and outweigh FINDRISC and Framingham traditional scores for prediction of T2D risk. Of note, the models developed in Scenarios 1 and 2, only exploited variables easily available at general patient visits.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
10
0

Year Published

2018
2018
2024
2024

Publication Types

Select...
6
2
1

Relationship

6
3

Authors

Journals

citations
Cited by 13 publications
(10 citation statements)
references
References 38 publications
0
10
0
Order By: Relevance
“…However, it is legitimate to ask whether a different method would perform better or worse in terms of ability to predict, e.g., the survival time. At this purpose, we trained a LASSO ‘least absolute shrinkage and selection operator’ regression analysis [26] coupled with Cox survival model [27] and recursive feature elimination as done in [28], on the same training data used to learn the DBN model. We obtained an AU-ROC equal to 0.79 and 0.64 at months 12 and 24, respectively; at the same time points the DBN achieved an AU-ROC equal to 0.96 and 0.90.…”
Section: Discussionmentioning
confidence: 99%
“…However, it is legitimate to ask whether a different method would perform better or worse in terms of ability to predict, e.g., the survival time. At this purpose, we trained a LASSO ‘least absolute shrinkage and selection operator’ regression analysis [26] coupled with Cox survival model [27] and recursive feature elimination as done in [28], on the same training data used to learn the DBN model. We obtained an AU-ROC equal to 0.79 and 0.64 at months 12 and 24, respectively; at the same time points the DBN achieved an AU-ROC equal to 0.96 and 0.90.…”
Section: Discussionmentioning
confidence: 99%
“…It is possible that metabolites could further improve the predictive value of diabetic subgroups, because we have recently shown that type 2 diabetes can be subdivided into five subgroups with six clustering variables (age at diagnosis, BMI, HbA1c, GAD autoantibodies, homeostatic model assessment of β-cell function, and homeostatic model assessment of INS resistance) with C-peptide measurements (34). The predictive accuracy may also be further improved by including additional anthropometric and lifestyle data (35).…”
Section: Discussionmentioning
confidence: 99%
“…12 The baseline study was conducted in 2004-2008 and the prospective study in 2010-2015. 18 For the present randomized parallel group clinical trial (RCT, ClinicalTrials. gov identifier NCT02131701) we recruited consecutive male and female participants from the PPP-Botnia Study who fulfilled the following criteria: (1) no diabetes and aged 30-70 years ; (2) poor physical fitness based on a 2 km walking test (fitness index <90); and (3) no contraindications for physical training based on a physical examination including ECG.…”
Section: Research Design and Methods Study Participantsmentioning
confidence: 99%