Table 1 Summary of clinical prediction models of RA, RA-CVD, and RA-ILD
Study diseases | Study population/sample size | Predictors | Outcome measures | Model | Model representation | External validation | Reference |
|---|---|---|---|---|---|---|---|
RA | American and Swedish/289–629 | 14 SNPs, 8 HLA alleles, age, gender, smoking | Occurrence of RA | Logistic regression | NR | Yes | Karlson/2010 [11] |
RA | American/542 | 31 SNPs, 8 HLA alleles, smoking | Seronegative, seropositive, erosive, seropositive and erosive RA phenotypes | Logistic regression | NR | Yes | Chibnik/2010 [12] |
RA | 5 countries/11,366 | 45 RA non-HLA susceptibility loci, imputed amino acids at HLA-DRB1 (11, 71, and 74), HLA-DPB1 (position 9), HLA-B (position 9), gender | Occurrence of RA | Logistic regression | NR | Yes | Yarwood/2015 [13] |
RA | American and Swedish/317–987 | 31 non-HLA alleles, 8 HLA-DRB1 alleles, HLA-SE*smoking, age, smoking, alcohol, education, parity, region, menopause, exposure to silica | Occurrence of RA | Logistic regression | NR | Yes | Karlson/2013 [14] |
RA | American and Swedish/381–1244 | 31 non-HLA alleles, 8 HLA-DRB1 alleles, age, smoking, alcohol, education, parity, BMI, family history | Occurrence of RA | Logistic regression | NR | Yes | Sparks/2015 [10] |
RA | England/80–2623 | 15 four-digit/10 two-digit HLA-DRB1 alleles, 31 SNPs, male ever-smoking status | Risk and age of onset of RA | COX regression | Diagram | No | Scott/2013 [15] |
RA | Multiple ethnicities/8683 | diabetes, depression, BMI, hypertension, gout, gender, ethnicity, smoking, sleep hours, income to poverty ratio | Occurrence of RA | Bayesian logistic regression | NR | Yes | Lufkin/2021 [16] |
RA | Singaporean Chinese/599 | 9 SNPs | Occurrence of RA | Machine learning | Webpage | Yes | Lim/2023 [17] |
RA | Netherlander/205 | RF, gender, smoking status, DAS, AMPD1, ATIC, ITPA, MTHFD1 | Insufficient response to MTX | Logistic regression | Risk score | Yes | Wessels/2007 [18] |
RA | Netherlander/285 | ABCB1 rs1045642 genotype, ABCC3 rs4793665 genotype, erythrocyte-folate, DAS28, HAQ, current smoking, BMI | Insufficient response to MTX | Logistic regression | Risk score | Yes | De Rotte/2018 [19] |
RA | Japanese/134 | SLCO3A1, CYP7A1, CHST10, GGH, SLC22A1, EPHX1, ATP7B, DAS28, folic acid | Efficacy and hepatotoxicity of MTX | Logistic regression | NR | No | Onishi/2020 [20] |
RA | Netherlander/91 | Erythrocyte-folate, DAS28, HAQ, current smoking, BMI | Insufficient response to MTX | Logistic regression | Webpage | No | Gosselt/2020 [21] |
RA | Norway/218 | Gender, SJC, RAI, PGA, ACPA, CRP, radiographic erosions, US, and MRI variables | Insufficient response to MTX and future structural damage progression | Logistic regression | NR | No | Sundin/2021 [22] |
RA | 40 countries/3280 | Gender, HAQ, presence of comorbidities, age, TJC, ESR | Treatment response to golimumab | Logistic regression | Diagram | Yes | Vastesaeger/2016 [23] |
RA | Netherlander/80 | CD14+, T cells, CD4+ T cells, PBMC RNA, PBMC DNA | Treatment response to adalimumab or etanercept | Random forest | Diagram | Yes | Tao/2021 [24] |
RA | Netherlander/93 | DAS28, interferon score, DMARDs use | Treatment response to rituximab | Logistic regression | NR | Yes | De Jong/2018 [25] |
RA | Multiple countries/164 | Synovial biopsy-based gene expression and histological data | Response to rituximab, tocilizumab and multidrug resistance | Elastic net regression and GBM | NR | No | Rivellese/2022 [26] |
RA-CVD | American/15,744 | Age, sex, diabetes, hyperlipidemia, hypertension, tobacco use, CDAI, modified HAQ, prednisone use, RA duration | Occurrence of CVD events, including MI, stroke, or fatal CVD in the next 10Â years | COX regression | Risk score | Yes | Solomon/2015 [27] |
RA-CVD | American/20,467 | MBDA score, age, diabetes, hypertension, tobacco use, CVD history, leptin, MMP-3, and TNF-R1 | Occurrence of CVD events, including MI, stroke, or fatal CVD in the next 3Â years | COX regression | Mathematical formula | No | Curtis/2020 [28] |
RA-CHD | Chinese/1012 | Age, hypertension, ACPA, LDL, HDL, TG, ESR | Occurrence of CHD | Logistic regression | Nomogram | Yes | Wei/2022 [29] |
RA-ILD | Italians/90 | VC | Occurrence of ILD | NR | NR | Yes | Pancaldi/2018 [30] |
RA-ILD | Chinese/183 | Male, smoke, cough, VC, have taken MTX, RF, ACPA, cold wet paralysis obstruction | Occurrence of ILD | Logistic regression | Nomogram | No | Ge/2021 [31] |
RA-ILD | Japanese/58 | FVC, PaO2/FiO2 ratio | mortality of after 90Â days of AE-RA-ILD | Recursive partitioning | Decision tree | No | Hozumi/2022 [32] |