Overview

In this report we examine baseline characteristics (such as DFU wound area) and some other demographic information such as age and sex of patients with emphasis on healing status of patient at week 12.

Descriptive statistics - table

Table 1. Summary of patient characteristics
Characteristic Healed, N = 181 Unhealed, N = 321 Difference2 95% CI2,3 p-value2
Age (in years) 52 (10) 58 (13) -6.0 -13, 0.80 0.082
Sex
Female 3 (17%) 4 (12%)
Male 15 (83%) 28 (88%)
Baseline wound area (sq.cm) 2.93 (2.96) 3.65 (3.15) -0.72 -2.5, 1.1 0.4
cMyc (%) 69 (23) 72 (16) -2.7 -15, 9.8 0.7
1 Mean (SD); n (%)
2 Welch Two Sample t-test
3 CI = Confidence Interval

Descriptive figures - boxplots

Model and findings

We consider the following baseline logistic regression model \[\text{Healing status} \sim \log \left[\text{Baseline wound area (sq. cm.)} \right] + \text{Sex}.\]

Characteristic OR1 95% CI1 p-value
Baseline wound area (sq. cm.) 0.66 0.32, 1.28 0.2
Sex
Female — —
Male 0.57 0.10, 3.38 0.5
AUC for baseline model: 0.5998
1 OR = Odds Ratio, CI = Confidence Interval


Since we do not have complete information on \(\text{Wound duration (months)}\), so we shall focus on the univariate model with just baseline wound area. Note that for the baseline model, AUC = 0.6, which is close to the reported value of AUC = 0.66 in the protocol.

Next, we consider a larger logistic regression model \[\text{Healing status} \sim \log \left[\text{Baseline wound area (sq. cm.)} \right] + \text{Sex} + \text{cMyc (%)} + \text{cMyc (%)} \times \text{Sex}. \]

Characteristic OR1 95% CI1 p-value
Baseline wound area (sq. cm.) 0.62 0.28, 1.24 0.2
Baseline cMYC marker (%) 1.02 0.92, 1.13 0.7
Sex
Female — —
Male 4.24 0.00, 7,209 0.7
Baseline cMYC marker (%) * Sex
Baseline cMYC marker (%) * Male 0.97 0.87, 1.08 0.5
AUC for cMYC model: 0.6146
1 OR = Odds Ratio, CI = Confidence Interval


Note that the AUC for the larger model does not point to a significant improvement, AUC = 0.615. We compare the baseline and cMyc-added larger logistic model via their respective AUCs. The null hypothesis is that difference in AUC is equal to zero. The alternative hypothesis is that the true difference (larger - baseline) in AUC is greater than zero. The standardized z-score (95% CI) of the difference of AUCs is given by 0.015 (-0.104, 0.075). The p-value is 0.627.