Choose the way to analyze

Add in characteristics from BAART

1. T-Tests

T-tests showed that there is a significant difference in the following.

## 
##  Welch Two Sample t-test
## 
## data:  wound by arm
## t = 3.772, df = 371.94, p-value = 0.0001883
## alternative hypothesis: true difference in means between group Placebo and group Active is not equal to 0
## 95 percent confidence interval:
##  1.860986 5.914234
## sample estimates:
## mean in group Placebo  mean in group Active 
##              25.59814              21.71053
## 
##  Welch Two Sample t-test
## 
## data:  contra by arm
## t = 2.2915, df = 367.89, p-value = 0.0225
## alternative hypothesis: true difference in means between group Placebo and group Active is not equal to 0
## 95 percent confidence interval:
##  0.2200463 2.8821301
## sample estimates:
## mean in group Placebo  mean in group Active 
##              19.83050              18.27941
## 
##  Welch Two Sample t-test
## 
## data:  VapoData$wound and VapoData$contra
## t = 7.6021, df = 657.91, p-value = 1.008e-13
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  3.533259 5.994130
## sample estimates:
## mean of x mean of y 
##  23.88153  19.11784
## 
##  Welch Two Sample t-test
## 
## data:  dif by arm
## t = 3.8885, df = 367.54, p-value = 0.0001197
## alternative hypothesis: true difference in means between group Placebo and group Active is not equal to 0
## 95 percent confidence interval:
##  1.683024 5.126741
## sample estimates:
## mean in group Placebo  mean in group Active 
##              6.836000              3.431118

Graphs

Healed Wound Site TEWL Box plots by Arms

Box Plot of Wound/Contralateral TEWL Differences by ARM

Effect Plots

What is impacting the difference between wound site TEWL and contralateral TEWL at the first visit.

  • The only thing significant was how long it took a patient to heal

    • Each additional week it took to heal was associated with an increase of about 0.319 [0.075-0.563] in TEWL

    • p-value: 0.031

  • That being said, it is not significant when the same marker is evaluated as in the mixed model below for DM (Dm duration vs monofilament) p value jumps to 0.086

Variable Estimate CI lower CI upper P-value
smoking2 Current Smoker 5.860 -0.569 12.289 0.108
smoking1 Former Smoker 2.690 -2.770 8.150 0.359
monofilament_score Monofilament Score 0.808 -0.231 1.847 0.162
weeks_since_healed Time Since Healed (weeks) 0.754 -0.286 1.795 0.189
age Age 0.332 -0.093 0.757 0.160
end2_weeks Weeks to Heal 0.319 0.075 0.563 0.031
bmi BMI 0.144 -0.221 0.509 0.459

Mixed Model: What has an impact on Wound Site TEWL?

Variable Estimate CI lower CI upper P-value
smoking2 Current Smoker 6.865 1.027 12.702 0.044
smoking1 Former Smoker 3.123 -2.111 8.357 0.269
contra Contralateral TEWL 0.783 0.618 0.949 0.000
age Age 0.358 0.009 0.706 0.071
end2_weeks Weeks to Heal 0.063 -0.177 0.302 0.620
bmi BMI -0.099 -0.338 0.140 0.436
weeks_since_healed Time Since Healed (weeks) -0.104 -0.215 0.006 0.065
dm_duration Duration of DM -0.232 -0.449 -0.015 0.063

Bootstrapping Data

I need to bootstrap the data, with an n of only 19, to stabilize the data to get better CI this will be the best way. We will re-sample the data over and over to stabilize. 1000 iterations used to stabilize the model.

## 
## ORDINARY NONPARAMETRIC BOOTSTRAP
## 
## 
## Call:
## boot(data = VapoData, statistic = boot_function, R = 1000)
## 
## 
## Bootstrap Statistics :
##        original        bias    std. error
## t1* -9.03217953  7.550905e-02  5.29169421
## t2* -0.10430219 -3.572098e-03  0.05150797
## t3*  0.35756795  9.969931e-04  0.06538626
## t4*  0.78317501 -4.712997e-03  0.09492828
## t5*  0.06263134 -1.269967e-03  0.05131629
## t6* -0.09885877 -4.558115e-05  0.03884631
## t7* -0.23174245 -9.348102e-04  0.04421200
## t8*  3.12296511  2.721089e-02  1.03211314
## t9*  6.86454927  3.824133e-02  1.12274961

Effect Model with Bootstrapped CI

Table for Effect Size Plot (Bootstrapped)

Variable Estimate CI lower CI upper P-value
Current Smoker 6.865 4.673 9.019 < 0.001
Former Smoker 3.123 1.199 5.202 < 0.001
Contralateral TEWL 0.783 0.596 0.960 0.472
Age 0.358 -0.173 -0.021 < 0.001
Weeks to Heal 0.063 0.230 0.484 1.000
BMI -0.099 -0.035 0.157 < 0.001
Time Since Healed (weeks) -0.104 -0.212 -0.008 0.001
Duration of DM -0.232 -0.316 -0.147 < 0.001

Global TEWL Increase

## 
##  Pearson's product-moment correlation
## 
## data:  VapoData$wound and VapoData$contra
## t = 12.043, df = 368, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.4543719 0.6010660
## sample estimates:
##      cor 
## 0.531695
## 
##  Pearson's product-moment correlation
## 
## data:  VapoData$five_cm and VapoData$contra_five_cm
## t = 16.944, df = 368, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.6005896 0.7156582
## sample estimates:
##       cor 
## 0.6620072

For every increase of 1 in wound TEWL it is expected that the contralateral TEWL will increase by 0.3456 ( p = \(2.2 * 10 ^{-16}\)). The correlation between wound TEWL and contralateral TEWL is r = 0.532 with a p value of \(2.18 * 10^{-28}\) . This lets us know that there is a moderately strong positive correlation that is significant between the two.

Similarly, when comparing the TEWL from five cm from the wound to five cm from the contralateral site. It is expected that every increase of 1 in the five cm away TEWL will increase the TEWL of the site 5 cm away from the contralateral site by 0.746 ( p-value \(2 * 10^{-16}\) ). There was a moderatley strong positive correlation between them, r = 0.662 ( p-value \(5.15 * 10^ {-48}\) ).

Pearson Correlation Test Results
Correlation.Coefficient..r. Degrees.of.Freedom T.Statistic P.Value X95..Confidence.Interval
cor 0.531695 368 12.04303 0 0.454371881501052 - 0.601065983994448

Call: lm(formula = contra ~ wound, data = VapoData)

Residuals: Min 1Q Median 3Q Max -18.2979 -3.7370 -0.8646 3.8372 16.0812

Coefficients: Estimate Std. Error t value Pr(>|t|)
(Intercept) 10.6880 0.7581 14.10 <2e-16 wound 0.3456 0.0287 12.04 <2e-16 — Signif. codes: 0 ‘’ 0.001 ’’ 0.01 ’’ 0.05 ‘.’ 0.1 ’ ’ 1

Residual standard error: 5.599 on 368 degrees of freedom (15 observations deleted due to missingness) Multiple R-squared: 0.2827, Adjusted R-squared: 0.2808 F-statistic: 145 on 1 and 368 DF, p-value: < 2.2e-16

Pearson Correlation Test Results
Correlation.Coefficient..r. Degrees.of.Freedom T.Statistic P.Value X95..Confidence.Interval
cor 0.6620072 368 16.94402 0 0.600589642045215 - 0.715658229405885

Call: lm(formula = five_cm ~ contra_five_cm, data = VapoData)

Residuals: Min 1Q Median 3Q Max -20.6655 -3.7009 -0.5923 3.4042 18.6302

Coefficients: Estimate Std. Error t value Pr(>|t|)
(Intercept) 6.63556 0.82899 8.004 1.59e-14 contra_five_cm 0.74643 0.04405 16.944 < 2e-16 — Signif. codes: 0 ‘’ 0.001 ’’ 0.01 ’’ 0.05 ‘.’ 0.1 ’ ’ 1

Residual standard error: 4.976 on 368 degrees of freedom (15 observations deleted due to missingness) Multiple R-squared: 0.4383, Adjusted R-squared: 0.4367 F-statistic: 287.1 on 1 and 368 DF, p-value: < 2.2e-16

Scatter Plots Wound x Contra TEWL