Skip to main content

Table 1 Patient characteristics in data for model development and external validation

From: Using real-world data to dynamically predict flares during tapering of biological DMARDs in rheumatoid arthritis: development, validation, and potential impact of prediction-aided decisions

Characteristics

Development data (n = 279)

DRESS data for external validation (n = 164)

P-value of differencea

General characteristics

 Age, mean in years (SD)

50.2 (17.1)

58.7 (9.9)

< 0.01

 Female, N (%)

203 (72.8%)

105 (64%)

0.05

 BMI, median in kg/m2 (IQR)

25.4 (22.6–30.0)

26.8 (23.3–29.5)

0.22

 Height, mean in cm (SD)

170.4 (12.5)

172 (9.2)

0.15

 Weight, mean in kg (SD)

77.0 (17.2)

78.8 (15.5)

0.27

 Follow-up time, median in months (SD)

21 (2.0)

18.7 (1.6)

< 0.01

RA characteristics

 Disease duration at start of bDMARD, median in years (IQR)

9.0 (5.0–16.5)

6.0 (2.7–12.7)

0.38

 Positivity for RF and/or ACPA, N (%)

236 (84.6%)

140 (85.4%)

0.83

Biological

 bDMARD type, N (%)

 Etanercept

77 (27.6%)

107 (65.2 %)

< 0.01

  Infliximab

5 (1.8%)

–

–

  Adalimumab

113 (40.5%)

57(34.8%)

0.23

  Certolizumab

10 (3.6%)

–

–

  Golimumab

18 (6.5%)

–

–

  Tocilizumab

8 (2.9%)

–

–

  Sarilumab

2 (0.7%)

–

–

 Abatacept

37 (13.3%)

–

–

 Time from start bDMARD baseline, mean in weeks (SD)b

10 (7.7)

42.1 (29.1)

< 0.01

 Prescribed dose during follow-up (mean, expressed as % of full dose)

76.7%

61.6%

< 0.01

Disease activity

 DAS28 at baseline, mean (SD)

2.79 (1.34)

2.15 (0.70)

< 0.01

 VAS GH at baseline, median (IQR)

30 (11–40)

20 (10–34)

0.76

 TJC at baseline, median (IQR)

0 (0–1)

0 (0–1)

–

 SJC at baseline, median (IQR)

0 (0–1)

0 (0–1)

–

 ESR at baseline, median in mm/hour (IQR)

7 (3–12)

13 (7–22)

< 0.01

 CRP at baseline, median mg/ml (IQR)

2.7 (1.3–5.0)

3 (3–3)

0.22

 Increase in TJC (yes/no)c, N (%)

49 (17.6%)

NA

–

 Increase in SJC (yes/no)c, N (%)

28 (10.0%)

NA

–

 No. of DAS28 measurements, mean (SD)

6.1 (3.8)

7.3 (1.2)

< 0.01

 Time between DAS28, mean in weeks (SD)

22.3 (12.3)

12.0 (5.4)

< 0.01

 Flare rate (# flares per patient year)

0.47

0.62

< 0.01

 DAS28 measurement rate (#DAS28 measurements per patient year)

2.18

4.72

0.04

  1. ACPA anti-citrullinated protein antibodies, bDMARD biological disease-modifying antirheumatic drug, CRP C-reactive protein, DAS28 disease activity score based on 28 joint count, EHR electronic health record, ESR erythrocyte sedimentation rate, IQR interquartile range, RF rheumatoid factor, SD standard deviation, TJ(C)/SJ(C) tender/swollen joint count, VAS GH an assessment of general health on a visual analog scale (0–100 mm)
  2. aP-values based on T-test for normally distributed continuous data, Mann-Whitney U test for continuous non-normally distributed data, and χ2 test for nominal data
  3. bIn the development data, baseline is defined as the first DAS28 ≤ 3.2
  4. cAn increase in TJC and/or SJC (yes/no) at baseline relative to the previous TJC/SJC measurement
Back to article page

Arthritis Research & Therapy

ISSN: 1478-6362

Contact us