Background and aims To set up the data base of Systemic Lupus Erymatosus using data capture technology (Hitales platform).

Methods Using Optical Character Recognition, Artificial Intelligence, Natural language processing technology to transfer the medical records into structured data which can be easily and freely explored. The medical records are acquired from admitted medical histories of department of Shanghai Renji Hospital.

Results Totally 4150 cases of admitted SLE patients in Dept. of Rheumatology Shanghai Renji Hospital from 2010–2015 were enrolled. The clinical patterns can be easily visualised. 3729 were females and 375 males; The average age was 36.2±14.1, with SLEDAI scores of 6.9±5.6. The most items frequently counted in SLEDAI were proteinuria (37.6%), low complement (33.7%) and rash (29.2%). Compared to female patients, male patients were tendency to have protenuira (48.4% vs.36.6%, p<0.01), hematuria (25.8% vs.19.7%, p<0.01). Disease activity evaluated by SLEDAI were highest in summer, however the highest cost in hospital were in winter. 47.0% (1948/4150) patients with lupus nephritis did renal biopsy. The majority pathology type was type IV (27.4%), while 23.6% for type V and 12.1% for V+IV. The most common features counted AI and CI were glomerular cell proliferation (89.6%) and interstitial fibrosis (62.4%) respectively.

Conclusions Using Hitales platform to set up our clinical database can extract medical information conveniently, quickly and efficiently with sufficient accuracy. So far, we only simply analysis the clinical features of SLE patients. With joint of biological specimens’ library and follow up data, the LUPUS puzzle could be learned more.