Construction And Application Of A Clinical Decision Support System In Alzheimer’s Disease

[Objective]Alzheimer’s disease(AD)is a common neurodegenerative disease of the elderly.By 2015,the number of people with the disease had reached 46.8 million worldwide.The inability to identify AD patients early and the high rate of delayed diagnosis or missed diagnosis.This may be due to the lack of systematic training on AD among primary health care workers in China,the lack of diagnostic skills and relevant ancillary tests in community service centers.This study develops a clinical decision support system(CDSS)for AD screening and diagnosis.It is suitable for use in community-general hospital referrals.We also evaluate the screening diagnostic efficacy of the system.[Methods]Draw an AD screening and diagnosis flowchart based on the Alzheimer’s disease screening and diagnosis knowledge base.Software engineers develop a CDSS for joint community-general hospital referrals based on the clinical diagnostic flowchart.Initial screening of older adults over 60 years of age is first conducted at community hospitals,and then those with suspected cognitive dysfunction are referred to memory and cognitive impairment clinics at general hospitals.Comprehensive hospitals complete neuropsychological scales and MRI.we then import the collected clinical data into CDSS to create a database.SPSS and Graphpad Prism software were used to calculate the diagnostic rates of AD and MCI,and to compare and analyze the diagnostic compliance rates of manual diagnosis and CDSS-assisted diagnosis of AD.SPSS was used to compare the correlation between each neuropsychological scale and MMSE.Graphpad Prism was employed to plot the subjects’ working curve(ROC)and to calculate the area under the curve(AUC),sensitivity and specificity.[Results]1.A total of 3897 subjects were included in this study.With the assistance of CDSS,community hospitals screened 224 cases with possible cognitive impairment who required referral to general hospitals(screening rate of 5.748%).Of the 224 residents who will be referred to the general hospital,6 residents had abnormal blood tests and 28 refused to be referred.Finally,190 cases were submitted to general hospitals.A total of 50 patients(1.283%diagnosis rate)were diagnosed as AD and 51 patients(1.309%diagnosis rate)were diagnosed as MCI after evaluation in the memory and cognitive disorders clinic of the general hospital.The prevalence of AD diagnosis in the Jiabei community before the application of the CDSS was 0.479%.The diagnosis rate of AD was 1.283%and the diagnosis rate of MCI was 2.592%after the application of the CDSS.2.In both community screening and hospital screening,the CDSS-assisted diagnosis and the traditional manual screening diagnosis were consistent,and they had a 100%compliance rate.3.The mean age of the 3897 cases included in the study was 69.4 ± 6.5 years.Among them,there were 2214 cases(56.8%)aged 60-69,with a mean age of 64.8±2.7 years;1403 cases(36.0%)aged 70-79,with a mean age of 73.6±2.7 years;273 cases(7.0%)aged 80-89,with a mean age of 83.0±2.5 years;7 cases(0.2%)aged 90-100,with a mean age of 93.0±2.8 years.-According to the gender classification,1684 cases(43.2%)were male and 2213 cases(56.8%)were female.According to the education level,there were 1197 cases of illiteracy,accounting for 30.7%;1241 cases of elementary school,accounting for 31.8%;1459 cases of junior high school and above,accounting for 37.4%.4.Among the 190 subjects referred to the general hospital,the mean age of the AD group(50 cases)was 70.1±6.5 years,19(38.0%)males and 31(62.0%)females;the mean age of the MCI group(51 cases)was 67.6±6.4 years,19(37.3%)males and 32(62.7%)females;the mean age of the HC group(89 cases)70.3±6.4 years,40 cases(44.9%)in males and 49 cases(55.1%)in females.The three groups were statistically different in terms of education level and presence of hypertension(p<0.05).5.We analyzed the between-group differences in cognitive assessment scale scores in the AD,MCI,and HC groups.The results suggested that the differences in AD8,MMSE,AVLT,BNT,TMT-A,and TMT-B were statistically significant(p<0.001).The scores of each scale in the three groups were HC group>MCI group>AD group,which were consistent with clinical symptoms.6.AVLT and BNT scores were positively correlated with MMSE scores(r=0.7218,p<0.001;r=0.5298,p<0.001),and TMT-A and TMT-B hours were negatively correlated with MMSE scores(r=-0.4262,p<0.001;r=-0.5449,p<0.001).7.In the AD screening diagnostic CDSS,the AUCs for the AVLT,TMT-A,TMT-B,and BNT scales to differentiate between patients with AD and HC were 0.9987(p<0.001,95%CI:0.996-1.000),0.7699(p<0.001,95%CI:0.679-0.861),0.8989(p<0.001,95%CI:0.8210.977),0.8449(p<0.001,95%CI:0.781-0.909).The AUCs for the AVLT,TMT-A,TMT-B,and BNT scales to differentiate between MCI patients and HC were 0.6336(p=0.009,95%CI:0.539-0.728),0.7215(p<0.001,95%CI:0.622-0.821),0.512(p<0.001,95%CI.0.410-0.614),0.6273(p=0.012,95%CI:0.533-0.722).[Conclusion]The CDSS designed in this study has excellent diagnostic efficacy for classification and is suitable for application in primary community and general hospitals.This CDSS can help community hospitals in early mass screening of older adults who may have impaired cognitive function.Through this system,older adults who may have cognitive impairment are accurately referred to a comprehensive hospital for evaluation.Eventually,the diagnosis rate of AD was increased.