Design,Development And Application Of An Information Registry System In Systemic Lupus Erythematosus

Objectives 1)To design and develop an information registry system for patients with systemic lupus erythematosus(SLE),and form a database platform for clinical researches in SLE.2)To collect clinical data of the hospitalized SLE patients via information registry system and set up a retrospective cohort in a single center,and then conduct a follow-up study.3)To carry out descriptive and statistical analyses of baseline and follow-up data for the SLE inpatients,and provide new scientific evidences for guiding clinical practice and improving prognosis of SLE.Methods 1)We utilized the methods of literature review,case analysis,experts' convention,structured method and rapid prototyping to design the clinical data questionnaire and system requirements specification.Then an information registry system was designed and developed based on Browser/Server network structure via C# as programme language,Visual Studio 2008 as application development platform and SQL Server 2008 as database platform.2)We took the disease code of SLE(M32.9)as the main keyword and retrieved the medical record of all SLE inpatients who admitted between January 1,2010 and December 31,2017 through the electronic medical record system in Anhui Provincial Hospital.According to the inclusion criteria of the study subjects,a list of SLE inpatients was finally generated after the duplications were removed.Subsequently,we collected the clinical data of all SLE inpatients in the list via information registry system and finally set up a retrospective cohort for the SLE inpatients.3)We used one-way ANOVA tests to compare different continuous variables with normal distributions,while nonparametric Mann-Whitney U tests in continuous variables with non-normal distributions for two groups and Kruskal-Wallis H tests for more than two groups.The Chi-square tests or Fisher exact tests were applied for comparisons of categorical variables.4)We selected an unsupervised systematic cluster analysis using Ward method and squared Euclidean distance algorithm to perform cluster analysis on nine autoantibodies in the SLE inpatient cohort.5)SLE inpatients were categorized into three groups(i.e.<P25,P25-P75 and >P75,respectively)based on the 25 th and 75 th percentiles of days stay in hospital at the first admissions.Ordinal regression models were applied to identify variables associated with the days of hospital stay at the first admissions.Besides,we analysed influence factors associated with pulmonary arterial hypertension(PAH)and Sj?gren's syndrome(SS)secondary to SLE by Logistic regression models.6)The end point of follow-up observation was set as December 31,2018.We collected the follow-up data by telephone interviews and SLE inpatients along with definited outcomes were included in the further descriptive and statistical analyses of follow-up data.7)In this study,Poisson regression models were applied to explore influence factors for increased hospitalizations,and Cox proportional hazard model for the prognosis of death in SLE inpatients.Results 1)We drew up the clinical data questionnaire and system requirements specification.The clinical data questionnaire had nine data collection modules,including basic information module,previous history module,diagnosis confirmation module,admission/discharge module,laboratory indices module,concomitant diseases module,disease activity module,damage index module and clinical treatments module.The system requirements specification exhibited the demands of six system function which contained creating users' characters function,assigning users' permissions function,registering patients function,collecting medical records function,checking medical records function and requesting recheck function.Finally,we designed and developed the information registry system for the SLE patients.2)A total of 620 inpatients were included in our SLE cohort.The age of the SLE inpatients cohort ranged from 10 to 84 years old.The distribution of age at SLE diagnosis was consistent with age at SLE onset with a slight time of delay.Forty-eight(7.7%)inpatients were males and five hundred and seventy-two(92.3%)were females.The ratio of males to females was 1:12.The number of inpatients with early-onset SLE,adult-onset SLE and late-onset SLE were 72(11.6%),464(74.8%)and 84(13.6%),respectively.Male SLE inpatients were associated with high incidences of having elevated creatinine and lupus nephritis,but with low chances of having decreased hemoglobin,increased ESR and PAH.Early-onset SLE inpatients were more likely to have low education level.They were associated with high incidences of having decreased C3 and high SLEDAI-2K score,but with low chances of having positive anti-SSA antibody.Late-onset SLE inpatients also had low education level.They were associated with high incidences of having serositis,decreased eGFR,diabetes,stroke and high SLICC/ACR damage index,but with low chances of having positive anti-RNP antibody,positive anti-SSA antibody and high SLEDAI-2K score.3)Four clusters were identified: Cluster 1(anti-Sm and anti-RNP antibodies),Cluster 2(anti-SSA and anti-SSB antibodies),Cluster 3(anti-dsDNA,anti-nucleosome and anti-C1 q antibodies)and Cluster 4(anti-rRNP and anti-phospholipidq antibodies).The associations between positive autoantibodies and system damages displayed that anti-Sm and anti-RNP antibodies were positive associated with mucocutaneous involvement.Anti-SSA,anti-dsDNA and anti-nucleosome antibodies were positive associated with musculoskeletal involvement,but anti-SSA antibody was negative associated with gastrointestinal involvement.Besides,anti-dsDNA and anti-C1 q antibodies were positive associated with renal and hematologic involvements,but were negative associated with ophthalmologic involvement.Anti-phospholipidq antibody(APL)was negative associated with renal involvement.4)Neuropsychiatric involvement(OR=3.177,95%CI: 1.540-6.554),gastrointestinal involvement(OR=2.168,95%CI: 1.217-3.865),increased CRP(OR=1.770,95%CI: 1.201-2.607),use of immunosuppressants(OR=2.234,95%CI: 1.523-3.277)and high SLEDAI-2K score(OR=1.068,95%CI: 1.030-1.109)were risk factors for prolonged hospitalizations,while new rural cooperative medical insurance(OR=0.503,95%CI: 0.283-0.894),positive anti-RNP antibody(OR=0.503,95%CI: 0.283-0.894)and use of hydroxychloroquine(OR=0.500,95%CI: 0.251-0.984)may be protective factors.5)SLE inpatients with PAH were more likely to be females and were associated with high incidences of having serositis,increased immunoglobulin IgG and high SLICC/ACR damage index,but with low chances of having oral ulcers.SLE inpatients with SS were more likely to be old and were associated with high incidences of having arthritis,increased immunoglobulin IgM,positive anti-SSA and anti-SSB antibodies,but with low chances of having positive anti-C1 q antibody and high SLEDAI-2K score.6)In the follow-up study,we obtained outcome data of 526 SLE inpatients and the effective follow-up rate was 84.9%.A total of 449 hospitalizations from 242 SLE inpatients were retrieved via EMR system in Anhui Provincial Hospital,and the annual hospitalization rate was about 18%.Infection,SLE damage and SLE activity are three leading causes of hospitalizations.In addition,male SLE inpatients were more likely to be admitted for infection.Early-onset SLE inpatients were more likely to be admitted for SLE damage and SLE activity,and late-onset SLE inpatients were more likely to be admitted for SLE activity and other comorbidities.Urban employee's basic medical insurance()RR=2.307,95%CI: 1.579-3.374,positive anti-nucleosome antibody(RR=1.370,95%CI: 1.087-1.726)and high SLICC/ACR damage index(RR=1.137,95%CI: 1.027-1.259)were risk factors for more frequencies of the cumulative hospitalizations,while positive anti-SSA antibody(RR=0.784,95%CI: 0.647-0.950)and use of hydroxychloroquine(RR=0.593,95%CI: 0.440-0.801)may be protective factors.7)Among 526 SLE inpatients,497(94.5%)patients were survived and 29(5.5%)patients were died.Infection(37.9%)and neuropsychiatric systemic lupus erythematosus(27.6%)were the main causes of death in SLE inpatients.The 1-,3-and 5-year cumulative survival rates in SLE inpatients were 96.4%,95.5% and 94.6 %,respectively.There were no significant differences in cumulative survival rates between males and females or between early-onset and adult-onset,while the cumulative survival rates in late-onset SLE inpatients was lower.Old at SLE diagnosis(HR=1.020,95%CI: 1.001-1.041),neuropsychiatric involvement(HR=4.291,95%CI: 1.920-9.593),cardiopulmonary involvement(HR=2.680,95%CI: 1.182-6.072),thrombocytopenia(HR=4.240,95%CI: 1.912-9.404),decreased eGFR(HR=4.240,95%CI: 1.644-10.932)and high SLICC/ACR damage index(HR=1.804,95%CI: 1.331-2.446)were risk factors for death in the SLE inpatients.Conclusions 1)The information registry system met the requirements of establishing a clinical database for the SLE patients,and achieved the initial goal in this study.Besides,it is also needed further module expansion and update.2)The baseline characteristics of the SLE inpatients were significant different between males and females,and among early-onset,adult-onset and late-onset SLE inpatients.Male inpatients were more likely to have renal involvement,while female inpatients were more likely to have anemia and PAH.Early-onset inpatients had high disease activity,but lower in late-onset inpatients.However,late-onset inpatients were more likely to have serositis and other chronic diseases,and also high system damage.3)Autoantibodies were significant associated with system damages in SLE and four clusters were identified.These findings may provide useful clues to assist disease diagnosis and predict organ damage in SLE.4)The influence factors associated with prolonged hospitalizations and increased hospitalizations could further confirm the importance of hydroxychloroquine in the treatments of SLE.Moreover,preventing organ damage and secondary infection should be potential targets in clinical practices.5)Comparing with SLE alone,inpatients with PAH or SS tend to have specific features in clinical and serological results,and these findings would be important to improve early diagnosis and prevent secondary injury in SLE.6)The annual hospitalization rate was 18%.Infection,SLE damage and SLE activity are three main causes of hospitalizations.There were significant differences in the causes of hospitalizations between males and females,and among early-onset,adult-onset and late-onset SLE inpatients.7)Infection and neuropsychiatric systemic lupus erythematosus were the main causes of death.The 1-,3-and 5-year cumulative survival rates were more than 90%.There were no significant differences in cumulative survival rates between males and females or between early-onset and adult-onset,while relative lower in late-onset SLE.Old at SLE diagnosis,neuropsychiatric and cardiopulmonary involvement,thrombocytopenia,decreased eGFR and high SLICC/ACR damage index were risk factors for prognosis of death in the SLE inpatients.8)The information registry system can provide standardized and diversified data for clinical researches in SLE and more valuable findings will be discovered in multicenter retrospective or prospective cohort studies.