Skip to main content
. 2019 Nov 21;138(1):50–57. doi: 10.1001/jamaophthalmol.2019.4636

Table. Meta-regression From Frequentist and Bayesian Models.

No. of Studies in Bayesian/Frequentist Analyses (n = 34/n = 18) Covariate Frequentist Model Bayesian Model
OR (95% CI) P Valuea OR (95% CrI) Probability Relative OR ≤ 1.00b
Diagnosis
6/3 AMD 1.01 (0.46-2.20) 1 [Reference] 0.98 (0.33-3.41) 1 [Reference]
15/11 DME 1.68 (1.01-2.77) .17 1.75 (0.94-3.85) 0.15
13/4 RVO 0.45 (0.09-2.17) .72 0.30 (0.03-2.25) 0.82
No. of injections
33/17 Per 1 more 1.12 (1.04-1.22) .005 1.06 (0.98-1.15) 0.06
Drug
18/6 Ranibizumab 1.44 (0.72-2.87) 1 [Reference] 1.75 (0.74-4.82) 1 [Reference]
6/1 Bevacizumab 0.66 (0.03-14.6) .65 24.4 (0.10-∞) 0.24
3/1 Pegaptanib sodium 1.00 (0.39-2.21) .64 1.03 (0.21-4.97) 0.74
6/10 Aflibercept 1.47 (0.79-2.72) .56 1.46 (0.61-3.23) 0.62
Follow-up, mo
14/5 6 (6-8) 0.65 (0.16-2.66) 1 [Reference] 0.82 (0.15-6.53) 1 [Reference]
15/8 12 (11-12) 1.02 (0.54-1.92) .83 1.07 (0.51-2.46) 0.41
5/5 24 (23-24) 1.89 (1.06-3.36) .17 2.07 (1.06-4.49) 0.21
Previous cardiovascular disease
15/7 No 1.12 (0.70-2.06) 1 [Reference] 1.23 (0.52-3.22) 1 [Reference]
18/11 Yes 2.12 (0.56-8.08) .14 1.50 (0.41-4.72) 0.24

Abbreviation: OR, odds ratio.

a

Calculated for the comparison of the ORs of death in the subgroup compared with that in the reference category.

b

Indicates probability that the relative OR for the category is more than 1.00 compared with the reference category.