Article Text

Download PDFPDF

407 Using decision tree to identify the itp with high probability of sle development from a nationwide cohort study
  1. TH Li1,2,
  2. YS Chang3 and
  3. CY Tsai4
  1. 1Taichung Venterans general hospital – Chiayi branch, Division of Allergy- Immunology- and Rheumatology-Department of Medicine, Chiayi City, Taiwan R.O.C
  2. 2National Yang-Ming University, Institute of Clinical Medicine, Taipei, Taiwan R.O.C
  3. 3Shuang-Ho Hospital- Taipei Medical University, Division of Allergy- Immunology- and Rheumatology- Department of Internal Medicine, New Taipei City, Taiwan R.O.C
  4. 4Taipei Veterans General Hospital, Division of Allergy- Immunology and Rheumatology- Department of Internal Medicine, Taipei, Taiwan R.O.C


Background and aims Idiopathic thrombocytopenic purpura (ITP) is an immune-related thrombocytopenia which may herald the development of systemic lupus erythematosus (SLE), and thus regular following up has been suggested. Whereas widespread surveillance on all ITP patients would be time and cost-consuming; therefore identifying those with high probability of development of SLE among ITP patients should be more practical.

Methods We enrolled ITP patients without previous SLE diagnosis from the Taiwan National Health Insurance research database between 1997 and 2012 and identified those with SLE diagnosis during follow up. We also analysed the symptoms and comorbidities as well as the dose of average oral steroid to derive the decision trees, which classified the ITP patients with different probability of development of SLE.

Results A total of 10 265 ITP patients were enrolled, among whom 80 patients developed SLE while following-up. The whole ITP patients were allocated to development group (7186 patients including 57 with SLE) and validation group (3079 patients including 23 with SLE); the former was used for derivation of the decision-tree based model (Figure 1) and the latter for validation of the previously mentioned model (Figure 2), and provided high sensitivity (78.2%), specificity (99.2%) and negative prediction value (99.8%). To reduce the complexity, we also proposed another models with different complexity parameters (Figure 3).

Conclusions We derived different decision tree models exempt from the necessity of laboratory data and adequate for various clinical scenarios of ITP patients, among whom those with high probability of development of SLE could be identified.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.