Statistical Analysis Doc
Ghassan Slim
2025-12-18
Section 1
(1.1.2) Demographics Table for Baseline Population Visit 1
Here we define the “baseline-population” demographics with the first instance for patients who completed at least one survey by phone-call interview.
Demographics have been re-coded into the following
Education:
- College Graduates vs others
Employment
- Full/part-time employed vs others
Income
- Greater to or equal to median (75k) vs less than median
Marital Status
- Married vs others
Insurance
- Insured (Commercial + Medicaid) vs Uninsured (Medicaid Part D + No Coverage)
Line of Therapy
- First line vs others
Months on Therapy
- Less than 13 months vs greater than or equal to 13 months
Daily Dosing Frequency
- once daily vs twice daily vs other (more complex)
##
## level Overall
## n 99
## Q6 (mean (SD)) 63.12 (14.43)
## Q1 (%) Male 59 (60.2)
## Female 39 (39.8)
## daily_freq_recoded (%) Once daily 76 (76.8)
## Other 13 (13.1)
## Twice daily 10 (10.1)
## edu_recoded (%) HighSchool_or_Less 28 (28.6)
## Other 70 (71.4)
## employment_recoded (%) Employed 31 (31.3)
## Other 68 (68.7)
## income_recoded (%) <75k 53 (53.5)
## >=75k 32 (32.3)
## Missing 14 (14.1)
## marital_recoded (%) Married 72 (72.7)
## Other 27 (27.3)
## insurance_recoded (%) Covered 46 (49.5)
## Other 47 (50.5)
## line_therapy_recoded (%) FirstLine 64 (66.7)
## Other 32 (33.3)
## months_therapy_recoded (%) <13 months 46 (46.5)
## >=13 months 46 (46.5)
## Missing 7 (7.1)
## wilson_adherence (mean (SD)) 92.15 (13.14)| Table 1B. Recoded Demographics and Baseline Characteristics | ||
| Baseline Population Characteristics for Visit 1 | ||
| Variable | level | Overall |
|---|---|---|
| n | ||
| 99 | ||
| Age (Years) | ||
| Median [Min, Max] | 66 [25, 87] | |
| Sex | ||
| Male | 59 (60.2) | |
| Female | 39 (39.8) | |
| Daily Dosing Frequency | ||
| Once daily | 76 (76.8) | |
| Other | 13 (13.1) | |
| Twice daily | 10 (10.1) | |
| Education | ||
| HighSchool_or_Less | 28 (28.6) | |
| Other | 70 (71.4) | |
| Employment | ||
| Employed | 31 (31.3) | |
| Other | 68 (68.7) | |
| Income | ||
| <75k | 53 (53.5) | |
| >=75k | 32 (32.3) | |
| Missing | 14 (14.1) | |
| Marital Status | ||
| Married | 72 (72.7) | |
| Other | 27 (27.3) | |
| Insurance Coverage | ||
| Covered | 46 (49.5) | |
| Other | 47 (50.5) | |
| Line of Therapy | ||
| FirstLine | 64 (66.7) | |
| Other | 32 (33.3) | |
| Months on Therapy | ||
| <13 months | 46 (46.5) | |
| >=13 months | 46 (46.5) | |
| Missing | 7 (7.1) | |
| wilson_adherencemeanSD | ||
| 92.15 (13.14) | ||
(1.1.3) Demographics Table for Final Population Visit 1
In this section we define the demographics section for the “final-population”, where this table includes demographics of patients who completed all three patients at their first instance.
##
## level Overall
## n 42
## Q6 (mean (SD)) 61.05 (12.16)
## Q1 (%) Male 25 (59.5)
## Female 17 (40.5)
## daily_freq_recoded (%) Once daily 36 (85.7)
## Other 3 (7.1)
## Twice daily 3 (7.1)
## edu_recoded (%) HighSchool_or_Less 14 (34.1)
## Other 27 (65.9)
## employment_recoded (%) Employed 15 (35.7)
## Other 27 (64.3)
## income_recoded (%) <75k 28 (66.7)
## >=75k 14 (33.3)
## marital_recoded (%) Married 30 (71.4)
## Other 12 (28.6)
## insurance_recoded (%) Covered 19 (45.2)
## Other 23 (54.8)
## line_therapy_recoded (%) FirstLine 24 (57.1)
## Other 18 (42.9)
## months_therapy_recoded (%) <13 months 22 (52.4)
## >=13 months 19 (45.2)
## Missing 1 (2.4)
## wilson_adherence (mean (SD)) 90.49 (13.03)| Table 1B. Recoded Demographics and Baseline Characteristics | ||
| Final Population Characteristics for Visit 1 | ||
| Variable | level | Overall |
|---|---|---|
| n | ||
| 42 | ||
| Age (Years) | ||
| Median [Min, Max] | 65 [27, 79] | |
| Sex | ||
| Male | 25 (59.5) | |
| Female | 17 (40.5) | |
| Daily Dosing Frequency | ||
| Once daily | 36 (85.7) | |
| Other | 3 (7.1) | |
| Twice daily | 3 (7.1) | |
| Education | ||
| HighSchool_or_Less | 14 (34.1) | |
| Other | 27 (65.9) | |
| Employment | ||
| Employed | 15 (35.7) | |
| Other | 27 (64.3) | |
| Income | ||
| <75k | 28 (66.7) | |
| >=75k | 14 (33.3) | |
| Marital Status | ||
| Married | 30 (71.4) | |
| Other | 12 (28.6) | |
| Insurance Coverage | ||
| Covered | 19 (45.2) | |
| Other | 23 (54.8) | |
| Line of Therapy | ||
| FirstLine | 24 (57.1) | |
| Other | 18 (42.9) | |
| Months on Therapy | ||
| <13 months | 22 (52.4) | |
| >=13 months | 19 (45.2) | |
| Missing | 1 (2.4) | |
| wilson_adherencemeanSD | ||
| 90.49 (13.03) | ||
Section 2
(2.1.1) Missing Wilson Adherence Data
This Table Describes how many patients in each instance had missing Wilson Adherence values. Wilson Adherence is calculated as a function of 3 questions self-rated adherence performance, self-rated frequency performance, and self-rated # of missed medication days. If any of these responses were missing, a Wilson adherence score was not calculated
| instance | n_total | n_missing | n_nonmissing | pct_missing |
|---|---|---|---|---|
| 1 | 99 | 4 | 95 | 4.0 |
| 2 | 51 | 2 | 49 | 3.9 |
| 3 | 42 | 0 | 42 | 0.0 |
(2.1.2) Adherence IQR Table
This table displays the Interquartile ranges, means, minimum, maximum and sample size of the baseline population (anyone with visit 1), and the final population (those who completed 3 surveys), at visits one, two and 3. The change in sample size amongst the final population is caused by N/A Wilson adherence values not being used in the calculations.
| Summary of Adherence Scores Across Visits | ||||
| Mean (SD), Median (IQR), Min–Max, and Sample Size | ||||
| Visit | Mean (SD) | Median (IQR) | Min, Max | Sample Size |
|---|---|---|---|---|
| 1 (baseline population) | 92.15 (13.14) | 100.00 [86.67, 100.00] | 30.00, 100.00 | 95 |
| 1 (final population) | 90.49 (13.03) | 100.00 [84.44, 100.00] | 53.33, 100.00 | 41 |
| 2 (final population) | 95.01 (7.64) | 100.00 [91.11, 100.00] | 73.33, 100.00 | 41 |
| 3 (final population) | 93.25 (11.81) | 100.00 [87.50, 100.00] | 56.67, 100.00 | 42 |
(2.1.3) Adherence change over time Sankey Plots
These plots show how adherence values, binned 100%, <100-90, and <90, change across the entire data set and across only the final population.
(2.1.4) Distributions of Adherence by visit
The following tables show the skewed nature of Wilson adherence values across the baseline population and all three instances of the final population.
(2.2.1) Linear Regression - Wilson Adherence
This section shows the results of a linear regression using the binary covariates as predictors and Wilson adherence as the outcome. This model uses the baseline population.
How to interpret:
Those with a non-traditional dosing regimen indicated as “other” have more than 10 points lower adherence values on average when compared to the reference group.
Those who are not married have more than 8 points lower adherence values on average when compared to the reference group.
##
## Call:
## lm(formula = wilson_adherence ~ age + sex + edu + emp + income +
## marital + insurance + line + months + dose, data = baseline)
##
## Residuals:
## Min 1Q Median 3Q Max
## -32.251 -4.176 3.100 7.145 15.931
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 89.7078 7.4448 12.050 <2e-16 ***
## age 0.1084 0.1104 0.981 0.3298
## sexFemale -1.0969 2.8505 -0.385 0.7015
## eduOther -0.3217 3.4312 -0.094 0.9256
## empOther 3.2456 3.2928 0.986 0.3277
## income>=75k -1.9532 3.3054 -0.591 0.5565
## incomeMissing -0.8120 4.7557 -0.171 0.8649
## maritalOther -8.0152 3.5269 -2.273 0.0261 *
## insuranceOther -3.4669 2.9159 -1.189 0.2385
## lineOther 1.9704 3.0175 0.653 0.5159
## months>=13 months -1.0550 2.7061 -0.390 0.6978
## monthsMissing 4.3492 5.8031 0.749 0.4561
## doseOther -10.2194 3.9688 -2.575 0.0121 *
## doseTwice daily -1.1378 5.7565 -0.198 0.8439
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 11.67 on 70 degrees of freedom
## (15 observations deleted due to missingness)
## Multiple R-squared: 0.1975, Adjusted R-squared: 0.04847
## F-statistic: 1.325 on 13 and 70 DF, p-value: 0.2197
(2.2.2) Logistic Regression; Add Reference Groups
The following sections use the aforementioned covariates as predictors, this time with adherence as a binary/categorical variable using several different cutoffs. This also uses the baseline population.
Upon the suggestion of Dr. Li a quasibinomial distrubition was investigated. This model would be applicable under the condition that there was not overdispersion. The dispersion of the logistic regression model was equal to 1, therefore no additional paramter was added to the model.
logistic regression equation:
- \[ \log\left(\frac{P(Y_i = 1)}{1 - P(Y_i = 1)}\right) = \beta_0 + \beta_1 \text{Age}_i + \beta_2 \text{Sex}_{i,\text{Female}} + \beta_3 \text{Sex}_{i,\text{PreferNot}} + \beta_4 \text{Education}_{i,\text{Other}} + \beta_5 \text{Employment}_{i,\text{Employed}} + \beta_6 \text{Income}_{i,\ge 75k} + \beta_7 \text{Income}_{i,\text{Missing}} + \beta_8 \text{Marital}_{i,\text{Married}} + \beta_9 \text{Insurance}_{i,\text{Covered}} + \beta_{10} \text{LineTherapy}_{i,\text{Other}} + \beta_{11} \text{MonthsTherapy}_{i,\ge 13} + \beta_{12} \text{MonthsTherapy}_{i,\text{Missing}} + \beta_{13} \text{Dose}_{i,\text{TwiceDaily}} + \beta_{14} \text{Dose}_{i,\text{Other}} \]
The null hypothesis is that for all coefficients:
- \[ H_0: \beta_j = 0 \]
The alternative hypothesis is that at least one coefficient:
- \[ H_A: \beta_j \neq 0 \]
How to interpret:
- For any statistically significant covariates, the odds increases by a factor of \(OR\) when compared to the reference group
(2.2.2.1) 100 vs <100
## [1] 1
| term | OR | OR_low | OR_high | p_label |
|---|---|---|---|---|
| Age (Years) | 1.0454656 | 1.00081017 | 1.0999404 | 0.060 |
| Sex: Female | 0.6140329 | 0.20139303 | 1.8281298 | 0.381 |
| Education: Other | 1.1503679 | 0.29906641 | 4.2953449 | 0.835 |
| employment_recodedOther | 1.5648611 | 0.45915566 | 5.3783739 | 0.471 |
| Income: >=75k | 0.4482122 | 0.11050386 | 1.6365486 | 0.237 |
| Income: Missing | 0.3891570 | 0.05337266 | 2.8075829 | 0.339 |
| marital_recodedOther | 0.2591365 | 0.06148264 | 0.9677439 | 0.052 |
| insurance_recodedOther | 0.5453849 | 0.16257262 | 1.6758237 | 0.303 |
| Line of Therapy: Other | 2.0431898 | 0.61804811 | 7.5883279 | 0.258 |
| Months of Therapy: >=13 | 0.6930642 | 0.24303937 | 1.9467856 | 0.486 |
| Months of Therapy: Missing | 3.0392245 | 0.31490362 | 72.5333031 | 0.386 |
| Dose Regimen: Other | 0.2466343 | 0.04656360 | 1.1410160 | 0.081 |
| Dose Regimen: Twice daily | 0.1384896 | 0.01242514 | 1.1726514 | 0.078 |
## [1] 1(2.2.2.2) >=95 vs <95
| term | OR | OR_low | OR_high | p_label |
|---|---|---|---|---|
| Age (Years) | 1.0462381 | 1.00086116 | 1.1017480 | 0.060 |
| Sex: Female | 0.7981758 | 0.26064514 | 2.4307700 | 0.689 |
| Education: Other | 1.0751006 | 0.27547125 | 4.0594957 | 0.915 |
| employment_recodedOther | 1.9003785 | 0.55986814 | 6.5866345 | 0.302 |
| Income: >=75k | 0.5596621 | 0.14045884 | 2.0655612 | 0.390 |
| Income: Missing | 0.4297634 | 0.05888038 | 3.1410598 | 0.394 |
| marital_recodedOther | 0.2755105 | 0.06551329 | 1.0291944 | 0.063 |
| insurance_recodedOther | 0.5475143 | 0.16115798 | 1.7072684 | 0.312 |
| Line of Therapy: Other | 2.3021493 | 0.67245449 | 9.1418001 | 0.204 |
| Months of Therapy: ≥13 | 0.7174947 | 0.25025359 | 2.0336739 | 0.531 |
| Months of Therapy: Missing | 2.7489103 | 0.27877818 | 67.0329618 | 0.434 |
| Dose Regimen: Other | 0.2113518 | 0.03972981 | 0.9722677 | 0.052 |
| Dose Regimen: Twice daily | 0.1314847 | 0.01181185 | 1.1264572 | 0.072 |
(2.2.2.2) >=90 vs <90
| term | OR | OR_low | OR_high | p_label |
|---|---|---|---|---|
| Age (Years) | 1.0105498 | 0.96446609 | 1.056541 | 0.646 |
| Sex: Female | 0.6389413 | 0.20449257 | 1.977177 | 0.434 |
| Education: Other | 1.6720635 | 0.42599075 | 6.599741 | 0.456 |
| employment_recodedOther | 3.1702775 | 0.87114328 | 12.525764 | 0.086 |
| Income: ≥75k | 0.6791506 | 0.16630607 | 2.627930 | 0.576 |
| Income: Missing | 2.2285736 | 0.25619259 | 50.744264 | 0.519 |
| marital_recodedOther | 0.3872457 | 0.09508918 | 1.433184 | 0.164 |
| insurance_recodedOther | 0.7051001 | 0.21072541 | 2.246603 | 0.558 |
| Line of Therapy: Other | 1.7426779 | 0.49302330 | 6.967925 | 0.404 |
| Months of Therapy: ≥13 | 0.9763625 | 0.33065615 | 2.936620 | 0.965 |
| Months of Therapy: Missing | 1.2020973 | 0.12357053 | 27.091981 | 0.883 |
| Dose Regimen: Other | 0.2525363 | 0.04975346 | 1.223640 | 0.086 |
| Dose Regimen: Twice daily | 1.2365734 | 0.11364012 | 30.821438 | 0.873 |
(2.2.3) Uni-variate analysis of baseline population
Here we performed univariate analysis between a single covariate and Wilson Adherence. These tests examine whether there exists a difference in distribution of Wilson adherence within demographic groups. These are non-parametric tests used because of the non-normal distribution of adherence demonstrated in section 2.1.4. P-values are included in box-plots.
Covariates with 2 categories were compared using the Wilcoxon rank-sum test.
Covariates with 3 categories or more were compared using the Kruskal–Wallis test.
Marital Status remains the only statistically significant covariate in this portion of the analysis.
Section 3
(3.1) Missing PRO data
Below is a summary of how many times a given question was missed/missing in the dataset. The following approach was adopted for data imputation.
For 2nd and 3rd surveys of a patient if a question is missing, their score from their first survey will be used to replace the missing score
For 1st surveys, a missing question will be imputed with the population mode of the particular question at visit 1.
From this point onward, the aforementioned changes have been implemented in the PRO data.
| Table 3.1 — Missing PRO Scores by Visit | |||
| N (%) of patients missing each PRO item | |||
| Question | Visit 1 (Baseline) | Visit 2 (Final Pop.) | Visit 3 (Final Pop.) |
|---|---|---|---|
| Q21 – In general, would you say your health is... | 0 (0%) | 0 (0%) | 0 (0%) |
| Q22 – In general, would you say your quality of life is... | 0 (0%) | 0 (0%) | 0 (0%) |
| Q23 – In general, how would you rate your physical health? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q24 – In general, how would you rate your mental health? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q25 – In general, how would you rate your satisfaction with your social activities? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q26 – In general, how well do you carry out social activities and roles? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q27 – To what extent are you able to carry out everyday physical activities? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q28 – In the past 7 days: emotional problems such as anxiety or irritability | 0 (0%) | 0 (0%) | 0 (0%) |
| Q29 – In the past 7 days: fatigue on average | 0 (0%) | 0 (0%) | 0 (0%) |
| Q30 – How would you rate your pain on average? | 0 (0%) | 0 (0%) | 0 (0%) |
| Q31 – Mouth or throat sores – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q32 – Mouth or throat sores – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q33 – Nausea – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q34 – Nausea – severity at its worst | 0 (0%) | 0 (0%) | 0 (0%) |
| Q35 – Constipation – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q37 – Diarrhea – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q38 – Rash – presence | 0 (0%) | 0 (0%) | 0 (0%) |
| Q39 – Hand–foot syndrome – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q40 – Numbness/tingling – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q41 – Numbness/tingling – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q42 – Blurry vision – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q43 – Blurry vision – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q44 – Problems with concentration – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q45 – Problems with concentration – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q46 – Pain – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q47 – Pain – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q48 – Pain – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q49 – Headache – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q50 – Headache – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q51 – Headache – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q52 – Aching muscles – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q53 – Aching muscles – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q54 – Aching muscles – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q55 – Aching joints – frequency | 0 (0%) | 0 (0%) | 0 (0%) |
| Q56 – Aching joints – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q57 – Aching joints – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q58 – Insomnia – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q59 – Insomnia – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
| Q60 – Fatigue – severity | 0 (0%) | 0 (0%) | 0 (0%) |
| Q61 – Fatigue – interference with daily activities | 0 (0%) | 0 (0%) | 0 (0%) |
(3.2.1) Domain Separation Guide
Based on the guidance of clinician Dr. Muluneh, the following “domains” have been created, each containing relevant questions from the PRO questionnaire.
| Table: PRO-CTCAE Items by Symptom Domain | ||
| Domain mapping for PRO-CTCAE Questions Q31–Q61 | ||
| Domain | Q# | Symptom / Description |
|---|---|---|
| GI | Q31 | Mouth/throat sores – severity |
| GI | Q32 | Mouth/throat sores – interference |
| GI | Q33 | Nausea – frequency |
| GI | Q34 | Nausea – severity |
| GI | Q35 | Constipation – severity |
| GI | Q37 | Diarrhea – frequency |
| Dermatologic | Q38 | Rash – presence |
| Dermatologic | Q39 | Hand–foot syndrome – severity |
| Pain | Q40 | Numbness/tingling – severity |
| Pain | Q41 | Numbness/tingling – interference |
| Pain | Q46 | Pain – frequency |
| Pain | Q47 | Pain – severity |
| Pain | Q48 | Pain – interference |
| Pain | Q52 | Muscle aches – frequency |
| Pain | Q53 | Muscle aches – severity |
| Pain | Q54 | Muscle aches – interference |
| Pain | Q55 | Joint aches – frequency |
| Pain | Q56 | Joint aches – severity |
| Pain | Q57 | Joint aches – interference |
| Neurologic | Q42 | Blurry vision – severity |
| Neurologic | Q43 | Blurry vision – interference |
| Neurologic | Q44 | Concentration problems – severity |
| Neurologic | Q45 | Concentration problems – interference |
| Neurologic | Q49 | Headache – frequency |
| Neurologic | Q50 | Headache – severity |
| Neurologic | Q51 | Headache – interference |
| Constitutional | Q58 | Insomnia – severity |
| Constitutional | Q59 | Insomnia – interference |
| Constitutional | Q60 | Fatigue – severity |
| Constitutional | Q61 | Fatigue – interference |
(3.2.2) Data Imputation
(3.2.3) Composite Score visualization by Domain
For tracking domain level changes over time, a composite score has been created for each domain. A higher score corresponds to a higher symptom burden, which is also the case in the individual PRO questions. The plots below plot the composite score by domain vs the time from the first survey of each patient.
Presence of rash in Q38 has been recoded from no = 1 and yes = 2 -> no = 1 and yes = 5 to match the scale of the other questions in the dataset.
| Descriptive Table: Composite PRO Domain Scores | |||
| Mean (SD), Median (IQR), and Min–Max by Domain | |||
| Domain | Mean (SD) | Median (IQR) | Min–Max |
|---|---|---|---|
| GI | 8.8 (3.6) | 8.0 [6.0, 10.0] | 5.0–27.0 |
| Dermatologic | 5.4 (1.4) | 6.0 [6.0, 6.0] | 1.0–8.0 |
| Pain | 27.4 (12.3) | 25.0 [17.0, 34.5] | 14.0–65.0 |
| Neurologic | 5.9 (3.1) | 4.0 [4.0, 6.0] | 3.0–18.0 |
| Constitutional | 8.4 (3.8) | 8.0 [6.0, 10.0] | 4.0–20.0 |
(3.2.4) Composite Score Descriptive Table across visits
Below is a descriptive table of the domain composite scores by population and visit.
| Descriptive Table: Composite PRO Domain Scores by Visit and Population | |||
| Mean (SD), Median (IQR), and Min–Max | |||
| Population / Visit | Mean (SD) | Median (IQR) | Min–Max |
|---|---|---|---|
| GI | |||
| Baseline population – Visit 1 | 8.9 (3.7) | 8.0 [6.0, 10.0] | 5.0–27.0 |
| Final population – Visit 1 | 9.6 (4.3) | 8.5 [7.0, 10.8] | 6.0–27.0 |
| Final population – Visit 2 | 8.4 (3.1) | 7.5 [6.0, 10.0] | 6.0–18.0 |
| Final population – Visit 3 | 8.7 (3.7) | 8.0 [6.0, 9.8] | 6.0–22.0 |
| Dermatologic | |||
| Baseline population – Visit 1 | 5.5 (1.3) | 6.0 [6.0, 6.0] | 1.0–7.0 |
| Final population – Visit 1 | 5.7 (0.9) | 6.0 [6.0, 6.0] | 2.0–6.0 |
| Final population – Visit 2 | 5.1 (1.7) | 6.0 [6.0, 6.0] | 2.0–8.0 |
| Final population – Visit 3 | 5.4 (1.4) | 6.0 [6.0, 6.0] | 2.0–6.0 |
| Pain | |||
| Baseline population – Visit 1 | 26.1 (12.8) | 23.0 [15.0, 34.0] | 14.0–65.0 |
| Final population – Visit 1 | 28.2 (12.6) | 25.0 [19.0, 34.8] | 14.0–65.0 |
| Final population – Visit 2 | 29.4 (12.4) | 27.0 [20.2, 34.8] | 14.0–63.0 |
| Final population – Visit 3 | 28.7 (11.0) | 26.0 [20.2, 36.8] | 14.0–58.0 |
| Neurologic | |||
| Baseline population – Visit 1 | 5.7 (2.9) | 4.0 [4.0, 6.0] | 3.0–16.0 |
| Final population – Visit 1 | 6.4 (3.1) | 5.5 [4.0, 8.0] | 4.0–16.0 |
| Final population – Visit 2 | 6.3 (3.6) | 4.5 [4.0, 6.0] | 4.0–18.0 |
| Final population – Visit 3 | 5.6 (3.0) | 4.0 [4.0, 6.0] | 4.0–16.0 |
| Constitutional | |||
| Baseline population – Visit 1 | 8.1 (3.7) | 8.0 [5.0, 10.0] | 4.0–18.0 |
| Final population – Visit 1 | 8.3 (3.5) | 8.0 [6.0, 10.0] | 4.0–18.0 |
| Final population – Visit 2 | 8.4 (3.7) | 8.0 [6.0, 10.0] | 4.0–18.0 |
| Final population – Visit 3 | 8.9 (4.0) | 8.0 [6.0, 11.0] | 4.0–19.0 |
(3.3.1) Ordinal Regression of composite scores (final population visit 1)
This model treats the composite scores as an ordinal variable since there is a finite number of possible outcomes, and larger composite score are indicative of greater symptomatic burden. This model uses the final population visit 1.
Because of the wide range of responses, sometimes ranging between 10-60, the composite scores are collapsed into four quantiles of the composite score.
| Ordinal Regression Results — GI Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.99 (0.92, 1.05) | 0.648 |
| sex2 | 2.55 (0.66, 10.50) | 0.179 |
| eduOther | 1.12 (0.24, 5.22) | 0.885 |
| empOther | 3.47 (0.65, 19.27) | 0.143 |
| income>=75k | 1.27 (0.23, 7.41) | 0.785 |
| maritalOther | 0.33 (0.06, 1.63) | 0.176 |
| insuranceOther | 0.41 (0.08, 1.85) | 0.247 |
| lineOther | 1.46 (0.38, 5.83) | 0.583 |
| doseOther | 0.12 (0.01, 1.45) | 0.103 |
| doseTwice daily | 1.39 (0.03, 51.41) | 0.858 |
| Ordinal Regression Results — Pain Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.98 (0.91, 1.05) | 0.579 |
| sex2 | 1.09 (0.30, 3.98) | 0.891 |
| eduOther | 0.96 (0.22, 4.28) | 0.956 |
| empOther | 2.41 (0.45, 13.23) | 0.299 |
| income>=75k | 0.79 (0.14, 4.36) | 0.789 |
| maritalOther | 0.73 (0.15, 3.49) | 0.689 |
| insuranceOther | 1.62 (0.39, 6.88) | 0.507 |
| lineOther | 0.58 (0.15, 2.16) | 0.416 |
| doseOther | 1.08 (0.10, 12.88) | 0.949 |
| doseTwice daily | 0.49 (0.01, 15.60) | 0.686 |
| Ordinal Regression Results — Neurologic Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.98 (0.91, 1.05) | 0.509 |
| sex2 | 0.48 (0.13, 1.72) | 0.267 |
| eduOther | 0.55 (0.11, 2.64) | 0.451 |
| empOther | 0.55 (0.10, 2.79) | 0.476 |
| income>=75k | 0.49 (0.08, 2.74) | 0.420 |
| maritalOther | 0.32 (0.06, 1.64) | 0.176 |
| insuranceOther | 1.31 (0.30, 5.79) | 0.721 |
| lineOther | 0.54 (0.14, 2.04) | 0.366 |
| doseOther | 0.68 (0.07, 6.51) | 0.729 |
| doseTwice daily | 0.07 (0.00, 2.40) | 0.152 |
| Ordinal Regression Results — Dermatologic Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.03 (0.96, 1.12) | 0.373 |
| sex2 | 0.18 (0.04, 0.72) | 0.020 |
| eduOther | 0.41 (0.07, 2.00) | 0.277 |
| empOther | 0.49 (0.09, 2.52) | 0.401 |
| income>=75k | 0.54 (0.08, 3.55) | 0.522 |
| maritalOther | 2.51 (0.48, 13.37) | 0.272 |
| insuranceOther | 4.24 (1.01, 20.05) | 0.056 |
| lineOther | 0.83 (0.21, 3.19) | 0.791 |
| doseOther | 1.55 (0.12, 21.04) | 0.731 |
| doseTwice daily | 0.31 (0.01, 17.80) | 0.551 |
| Ordinal Regression Results — Constitutional Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.96 (0.88, 1.03) | 0.253 |
| sex2 | 1.95 (0.53, 7.39) | 0.317 |
| eduOther | 0.73 (0.13, 3.95) | 0.709 |
| empOther | 5.80 (1.01, 38.01) | 0.053 |
| income>=75k | 0.86 (0.16, 4.80) | 0.858 |
| maritalOther | 0.42 (0.07, 2.26) | 0.319 |
| insuranceOther | 0.90 (0.20, 3.91) | 0.891 |
| lineOther | 0.39 (0.10, 1.46) | 0.168 |
| doseOther | 0.58 (0.06, 6.67) | 0.643 |
| doseTwice daily | 9.70 (0.14, 748.32) | 0.277 |
## # A tibble: 4 × 2
## quartile n
## <int> <int>
## 1 1 11
## 2 2 11
## 3 3 10
## 4 4 10
(3.3.2) continued. Univarite analysis of stat sig covariates
Statistically significant covariates from section 3.3.1, underwent nonparametric univariate tests to check for differences in distributions of composite scores between different groups.
Insurance and sex, while statistically significant in the ordinal regression model of 3.3.1 for dermatological domain, are difficult to interpret because of the identical means and IQR of the two groups. This is because of only two questions, including the binary Q38, contributing to the lack of variability.
Employment, while not statistically significant in the ordinal regression (p = .053), was statistically significant in this nonparametric test.
##
## Wilcoxon rank sum test with continuity correction
##
## data: Dermatologic_score by sex
## W = 234, p-value = 0.3387
## alternative hypothesis: true location shift is not equal to 0
##
## Wilcoxon rank sum test with continuity correction
##
## data: Constitutional_score by emp
## W = 117.5, p-value = 0.02518
## alternative hypothesis: true location shift is not equal to 0
| Dermatologic Domain — Significant Covariate | |
| Median (IQR) PRO score by sex (Final population, Visit 1) | |
| Sex | Dermatologic PRO Score |
|---|---|
| Male | 6.0 (6.0–6.0) |
| Female | 6.0 (6.0–6.0) |
| Constitutional Domain — Significant Covariate | |
| Median (IQR) PRO score by employment status (Final population, Visit 1) | |
| Employment Status | Constitutional PRO Score |
|---|---|
| Employed | 6.0 (4.0–9.0) |
| Other | 9.0 (6.5–11.0) |
###(3.4)
(3.4.1 & 3.4.2)
In this section we converted the PRO scores for each individual question into a binary outcome to compare whether the proportion of responses changes across visits. This analysis uses the final population.
The original scale is between 1 and 5. In table 3.4.1, for each question we compare whether the proportion of responses equal to 1 vs greater than 1 changes across the three visits.
Similarly in table 3.4.2, for each question we compare whether the proportion of responses less than or equal to 2 vs greater than 3 changes across the three visits.
| Table 3.2.1 — PRO Change Over Time (Binary: PRO > 1) | |||||
| Final population only; Cochran’s Q test across 3 visits | |||||
| Question | Q# | Visit 1 | Visit 2 | Visit 3 | Cochran’s Q p-value |
|---|---|---|---|---|---|
| Q21 – In general, would you say your health is... | Q21 | 41 (97.6%) | 42 (100%) | 40 (95.2%) | 0.2230 |
| Q22 – In general, would you say your quality of life is... | Q22 | 34 (81%) | 37 (88.1%) | 35 (83.3%) | 0.5290 |
| Q23 – In general, how would you rate your physical health? | Q23 | 40 (97.6%) | 40 (97.6%) | 40 (97.6%) | 1.0000 |
| Q24 – In general, how would you rate your mental health? | Q24 | 30 (71.4%) | 37 (88.1%) | 36 (85.7%) | 0.0681 |
| Q25 – In general, how would you rate your satisfaction with your social activities? | Q25 | 34 (81%) | 37 (88.1%) | 35 (83.3%) | 0.5290 |
| Q26 – In general, how well do you carry out social activities and roles? | Q26 | 32 (80%) | 36 (90%) | 30 (75%) | 0.0970 |
| Q27 – To what extent are you able to carry out everyday physical activities? | Q27 | 30 (71.4%) | 25 (59.5%) | 23 (54.8%) | 0.0617 |
| Q28 – In the past 7 days: emotional problems such as anxiety or irritability | Q28 | 22 (52.4%) | 23 (54.8%) | 24 (57.1%) | 0.8290 |
| Q29 – In the past 7 days: fatigue on average | Q29 | 38 (92.7%) | 38 (92.7%) | 34 (82.9%) | 0.1350 |
| Q30 – How would you rate your pain on average? | Q30 | 28 (66.7%) | 27 (64.3%) | 29 (69%) | 0.7410 |
| Q31 – Mouth or throat sores – severity | Q31 | 10 (23.8%) | 3 (7.1%) | 3 (7.1%) | 0.0169 |
| Q32 – Mouth or throat sores – interference with daily activities | Q32 | 7 (16.7%) | 3 (7.1%) | 3 (7.1%) | 0.1690 |
| Q33 – Nausea – frequency | Q33 | 16 (38.1%) | 12 (28.6%) | 14 (33.3%) | 0.3010 |
| Q34 – Nausea – severity at its worst | Q34 | 17 (40.5%) | 12 (28.6%) | 14 (33.3%) | 0.1780 |
| Q35 – Constipation – severity | Q35 | 13 (31%) | 12 (28.6%) | 10 (23.8%) | 0.7050 |
| Q37 – Diarrhea – frequency | Q37 | 20 (48.8%) | 15 (36.6%) | 15 (36.6%) | 0.2870 |
| Q38 – Rash – presence | Q38 | 37 (88.1%) | 32 (76.2%) | 32 (76.2%) | 0.2100 |
| Q39 – Hand–foot syndrome – severity | Q39 | 4 (9.5%) | 3 (7.1%) | 5 (11.9%) | 0.6870 |
| Q40 – Numbness/tingling – severity | Q40 | 22 (52.4%) | 19 (45.2%) | 17 (40.5%) | 0.3270 |
| Q41 – Numbness/tingling – interference with daily activities | Q41 | 18 (42.9%) | 18 (42.9%) | 14 (33.3%) | 0.4490 |
| Q42 – Blurry vision – severity | Q42 | 11 (26.8%) | 13 (31.7%) | 7 (17.1%) | 0.1350 |
| Q43 – Blurry vision – interference with daily activities | Q43 | 12 (28.6%) | 13 (31%) | 8 (19%) | 0.2470 |
| Q44 – Problems with concentration – severity | Q44 | 18 (42.9%) | 13 (31%) | 12 (28.6%) | 0.1610 |
| Q45 – Problems with concentration – interference with daily activities | Q45 | 17 (40.5%) | 14 (33.3%) | 12 (28.6%) | 0.3270 |
| Q46 – Pain – frequency | Q46 | 29 (69%) | 27 (64.3%) | 33 (78.6%) | 0.0970 |
| Q47 – Pain – severity | Q47 | 30 (71.4%) | 28 (66.7%) | 31 (73.8%) | 0.5290 |
| Q48 – Pain – interference with daily activities | Q48 | 24 (57.1%) | 25 (59.5%) | 24 (57.1%) | 0.9490 |
| Q49 – Headache – frequency | Q49 | 16 (38.1%) | 17 (40.5%) | 19 (45.2%) | 0.6620 |
| Q50 – Headache – severity | Q50 | 16 (38.1%) | 17 (40.5%) | 17 (40.5%) | 0.9360 |
| Q51 – Headache – interference with daily activities | Q51 | 12 (28.6%) | 13 (31%) | 11 (26.2%) | 0.7940 |
| Q52 – Aching muscles – frequency | Q52 | 18 (43.9%) | 27 (65.9%) | 26 (63.4%) | 0.0362 |
| Q53 – Aching muscles – severity | Q53 | 18 (42.9%) | 28 (66.7%) | 27 (64.3%) | 0.0160 |
| Q54 – Aching muscles – interference with daily activities | Q54 | 14 (33.3%) | 22 (52.4%) | 23 (54.8%) | 0.0260 |
| Q55 – Aching joints – frequency | Q55 | 20 (47.6%) | 27 (64.3%) | 23 (54.8%) | 0.1130 |
| Q56 – Aching joints – severity | Q56 | 21 (50%) | 27 (64.3%) | 23 (54.8%) | 0.2110 |
| Q57 – Aching joints – interference with daily activities | Q57 | 18 (42.9%) | 23 (54.8%) | 22 (52.4%) | 0.3500 |
| Q58 – Insomnia – severity | Q58 | 15 (35.7%) | 17 (40.5%) | 18 (42.9%) | 0.6920 |
| Q59 – Insomnia – interference with daily activities | Q59 | 15 (35.7%) | 16 (38.1%) | 19 (45.2%) | 0.5040 |
| Q60 – Fatigue – severity | Q60 | 33 (78.6%) | 33 (78.6%) | 30 (71.4%) | 0.4410 |
| Q61 – Fatigue – interference with daily activities | Q61 | 29 (69%) | 32 (76.2%) | 28 (66.7%) | 0.4860 |
| Table 3.2.2 — PRO Change Over Time (Binary: PRO > 3) | |||||
| Final population only; Cochran’s Q test across 3 visits | |||||
| Question | Q# | Visit 1 | Visit 2 | Visit 3 | Cochran’s Q p-value |
|---|---|---|---|---|---|
| Q21 – In general, would you say your health is... | Q21 | 18 (42.9%) | 18 (42.9%) | 16 (38.1%) | 0.6950 |
| Q22 – In general, would you say your quality of life is... | Q22 | 10 (23.8%) | 10 (23.8%) | 7 (16.7%) | 0.4070 |
| Q23 – In general, how would you rate your physical health? | Q23 | 17 (41.5%) | 16 (39%) | 18 (43.9%) | 0.7170 |
| Q24 – In general, how would you rate your mental health? | Q24 | 8 (19%) | 7 (16.7%) | 7 (16.7%) | 0.9050 |
| Q25 – In general, how would you rate your satisfaction with your social activities? | Q25 | 11 (26.2%) | 10 (23.8%) | 9 (21.4%) | 0.7790 |
| Q26 – In general, how well do you carry out social activities and roles? | Q26 | 13 (32.5%) | 11 (27.5%) | 12 (30%) | 0.8070 |
| Q27 – To what extent are you able to carry out everyday physical activities? | Q27 | 8 (19%) | 5 (11.9%) | 6 (14.3%) | 0.6070 |
| Q28 – In the past 7 days: emotional problems such as anxiety or irritability | Q28 | 2 (4.8%) | 4 (9.5%) | 5 (11.9%) | 0.4170 |
| Q29 – In the past 7 days: fatigue on average | Q29 | 2 (4.9%) | 8 (19.5%) | 6 (14.6%) | 0.0446 |
| Q30 – How would you rate your pain on average? | Q30 | 20 (47.6%) | 20 (47.6%) | 23 (54.8%) | 0.5490 |
| Q31 – Mouth or throat sores – severity | Q31 | 4 (9.5%) | 1 (2.4%) | 0 (0%) | 0.0743 |
| Q32 – Mouth or throat sores – interference with daily activities | Q32 | 3 (7.1%) | 0 (0%) | 0 (0%) | 0.0498 |
| Q33 – Nausea – frequency | Q33 | 5 (11.9%) | 2 (4.8%) | 4 (9.5%) | 0.4170 |
| Q34 – Nausea – severity at its worst | Q34 | 4 (9.5%) | 2 (4.8%) | 1 (2.4%) | 0.1740 |
| Q35 – Constipation – severity | Q35 | 1 (2.4%) | 0 (0%) | 3 (7.1%) | 0.1740 |
| Q37 – Diarrhea – frequency | Q37 | 4 (9.8%) | 4 (9.8%) | 4 (9.8%) | 1.0000 |
| Q38 – Rash – presence | Q38 | 37 (88.1%) | 32 (76.2%) | 32 (76.2%) | 0.2100 |
| Q39 – Hand–foot syndrome – severity | Q39 | 1 (2.4%) | 0 (0%) | 3 (7.1%) | 0.0970 |
| Q40 – Numbness/tingling – severity | Q40 | 5 (11.9%) | 3 (7.1%) | 4 (9.5%) | 0.6870 |
| Q41 – Numbness/tingling – interference with daily activities | Q41 | 6 (14.3%) | 5 (11.9%) | 4 (9.5%) | 0.7410 |
| Q42 – Blurry vision – severity | Q42 | 1 (2.4%) | 4 (9.8%) | 3 (7.3%) | 0.0970 |
| Q43 – Blurry vision – interference with daily activities | Q43 | 3 (7.1%) | 5 (11.9%) | 2 (4.8%) | 0.2470 |
| Q44 – Problems with concentration – severity | Q44 | 4 (9.5%) | 2 (4.8%) | 1 (2.4%) | 0.0970 |
| Q45 – Problems with concentration – interference with daily activities | Q45 | 4 (9.5%) | 5 (11.9%) | 2 (4.8%) | 0.4590 |
| Q46 – Pain – frequency | Q46 | 16 (38.1%) | 14 (33.3%) | 14 (33.3%) | 0.7660 |
| Q47 – Pain – severity | Q47 | 7 (16.7%) | 6 (14.3%) | 7 (16.7%) | 0.9130 |
| Q48 – Pain – interference with daily activities | Q48 | 10 (23.8%) | 6 (14.3%) | 6 (14.3%) | 0.3680 |
| Q49 – Headache – frequency | Q49 | 5 (11.9%) | 7 (16.7%) | 4 (9.5%) | 0.3680 |
| Q50 – Headache – severity | Q50 | 4 (9.5%) | 2 (4.8%) | 2 (4.8%) | 0.3680 |
| Q51 – Headache – interference with daily activities | Q51 | 4 (9.5%) | 4 (9.5%) | 2 (4.8%) | 0.5130 |
| Q52 – Aching muscles – frequency | Q52 | 8 (19.5%) | 14 (34.1%) | 13 (31.7%) | 0.1440 |
| Q53 – Aching muscles – severity | Q53 | 4 (9.5%) | 3 (7.1%) | 4 (9.5%) | 0.8820 |
| Q54 – Aching muscles – interference with daily activities | Q54 | 5 (11.9%) | 7 (16.7%) | 7 (16.7%) | 0.7170 |
| Q55 – Aching joints – frequency | Q55 | 10 (23.8%) | 11 (26.2%) | 8 (19%) | 0.6270 |
| Q56 – Aching joints – severity | Q56 | 9 (21.4%) | 4 (9.5%) | 4 (9.5%) | 0.1030 |
| Q57 – Aching joints – interference with daily activities | Q57 | 8 (19%) | 7 (16.7%) | 4 (9.5%) | 0.3380 |
| Q58 – Insomnia – severity | Q58 | 2 (4.8%) | 3 (7.1%) | 5 (11.9%) | 0.3110 |
| Q59 – Insomnia – interference with daily activities | Q59 | 5 (11.9%) | 3 (7.1%) | 7 (16.7%) | 0.1350 |
| Q60 – Fatigue – severity | Q60 | 6 (14.3%) | 6 (14.3%) | 9 (21.4%) | 0.5260 |
| Q61 – Fatigue – interference with daily activities | Q61 | 10 (23.8%) | 12 (28.6%) | 14 (33.3%) | 0.4240 |
(3.4.3) Ordinal Regression of Individual questions (Collapse questions)
In this section we once again perform ordinal regression, this time at the question level, across all three visits of the final population where questions have 5 naturally occuring ordinal categories . The questions were selected by Dr. Muluneh, including:
Diarrhea (frequency)
Nausea (frequency and severity)
Fatigue (severity and interference)
And maybe also: Rash
Currently we are unable to perform ordinal regression on nausea questions Q33 and Q34, possibly because of the lack of responses in some categories. To circumvent this we will attempt to collapse the questions into fewer responses.
## # A tibble: 30 × 9
## term estimate std.error statistic p.value conf.low conf.high coef.type Item
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <chr> <chr>
## 1 age 1.01 0.0363 0.267 0.789 0.938 1.09 location Q37
## 2 sex2 0.390 0.717 -1.31 0.189 0.0891 1.54 location Q37
## 3 eduO… 0.766 0.835 -0.319 0.750 0.141 3.95 location Q37
## 4 empO… 2.32 0.887 0.950 0.342 0.415 14.5 location Q37
## 5 inco… 1.02 0.930 0.0243 0.981 0.167 6.78 location Q37
## 6 mari… 1.05 0.848 0.0531 0.958 0.193 5.64 location Q37
## 7 insu… 0.807 0.765 -0.280 0.780 0.171 3.61 location Q37
## 8 line… 1.27 0.723 0.330 0.742 0.305 5.36 location Q37
## 9 dose… 0.829 1.38 -0.136 0.892 0.0317 10.9 location Q37
## 10 dose… 1.67 1.92 0.266 0.790 0.0297 72.6 location Q37
## # ℹ 20 more rows| Ordinal Regression Results — Q37 | ||
| Final population, Visit 1 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.01 (0.94, 1.09) | 0.789 |
| sex2 | 0.39 (0.09, 1.54) | 0.189 |
| eduOther | 0.77 (0.14, 3.95) | 0.750 |
| empOther | 2.32 (0.42, 14.55) | 0.342 |
| income>=75k | 1.02 (0.17, 6.78) | 0.981 |
| maritalOther | 1.05 (0.19, 5.64) | 0.958 |
| insuranceOther | 0.81 (0.17, 3.61) | 0.780 |
| lineOther | 1.27 (0.30, 5.36) | 0.742 |
| doseOther | 0.83 (0.03, 10.92) | 0.892 |
| doseTwice daily | 1.67 (0.03, 72.64) | 0.790 |
| Ordinal Regression Results — Q60 | ||
| Final population, Visit 1 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.99 (0.92, 1.06) | 0.736 |
| sex2 | 1.19 (0.33, 4.34) | 0.795 |
| eduOther | 0.85 (0.18, 4.04) | 0.834 |
| empOther | 4.33 (0.82, 24.76) | 0.088 |
| income>=75k | 1.18 (0.20, 7.03) | 0.852 |
| maritalOther | 0.75 (0.14, 3.75) | 0.727 |
| insuranceOther | 1.46 (0.34, 6.21) | 0.607 |
| lineOther | 0.86 (0.22, 3.32) | 0.827 |
| doseOther | 0.79 (0.07, 8.50) | 0.842 |
| doseTwice daily | 0.43 (0.01, 14.96) | 0.640 |
| Ordinal Regression Results — Q61 | ||
| Final population, Visit 1 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.99 (0.92, 1.06) | 0.862 |
| sex2 | 2.81 (0.74, 11.28) | 0.134 |
| eduOther | 0.75 (0.15, 3.55) | 0.711 |
| empOther | 3.17 (0.61, 18.28) | 0.176 |
| income>=75k | 1.87 (0.32, 12.01) | 0.492 |
| maritalOther | 0.27 (0.05, 1.41) | 0.131 |
| insuranceOther | 1.08 (0.25, 4.58) | 0.913 |
| lineOther | 1.01 (0.28, 3.66) | 0.986 |
| doseOther | 1.10 (0.11, 10.93) | 0.931 |
| doseTwice daily | 0.70 (0.01, 25.66) | 0.846 |
## # A tibble: 20 × 9
## term estimate std.error statistic p.value conf.low conf.high coef.type Item
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <chr> <chr>
## 1 age 1.03 0.0345 0.727 0.467 0.957 1.10 location Q60
## 2 sex2 0.911 0.696 -0.134 0.893 0.229 3.61 location Q60
## 3 eduO… 0.620 0.823 -0.580 0.562 0.122 3.20 location Q60
## 4 empO… 3.88 0.910 1.49 0.136 0.660 24.7 location Q60
## 5 inco… 0.363 0.913 -1.11 0.266 0.0573 2.13 location Q60
## 6 mari… 0.242 0.875 -1.62 0.105 0.0409 1.30 location Q60
## 7 insu… 1.02 0.727 0.0298 0.976 0.242 4.33 location Q60
## 8 line… 1.33 0.682 0.419 0.675 0.349 5.18 location Q60
## 9 dose… 0.362 1.22 -0.835 0.404 0.0301 3.94 location Q60
## 10 dose… 0.945 1.82 -0.0309 0.975 0.0251 35.8 location Q60
## 11 age 1.02 0.0344 0.534 0.594 0.949 1.09 location Q61
## 12 sex2 2.61 0.732 1.31 0.190 0.631 11.5 location Q61
## 13 eduO… 0.764 0.812 -0.332 0.740 0.154 3.83 location Q61
## 14 empO… 4.65 0.920 1.67 0.0950 0.786 30.3 location Q61
## 15 inco… 0.270 0.936 -1.40 0.162 0.0401 1.64 location Q61
## 16 mari… 0.115 0.920 -2.35 0.0189 0.0175 0.663 location Q61
## 17 insu… 0.786 0.739 -0.327 0.744 0.179 3.34 location Q61
## 18 line… 0.757 0.706 -0.394 0.694 0.186 3.03 location Q61
## 19 dose… 0.323 1.19 -0.948 0.343 0.0294 3.49 location Q61
## 20 dose… 0.819 1.78 -0.112 0.911 0.0236 27.5 location Q61## NULL| Ordinal Regression Results — Q60 | ||
| Final population, Visit 2 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.03 (0.96, 1.10) | 0.467 |
| sex2 | 0.91 (0.23, 3.61) | 0.893 |
| eduOther | 0.62 (0.12, 3.20) | 0.562 |
| empOther | 3.88 (0.66, 24.70) | 0.136 |
| income>=75k | 0.36 (0.06, 2.13) | 0.266 |
| maritalOther | 0.24 (0.04, 1.30) | 0.105 |
| insuranceOther | 1.02 (0.24, 4.33) | 0.976 |
| lineOther | 1.33 (0.35, 5.18) | 0.675 |
| doseOther | 0.36 (0.03, 3.94) | 0.404 |
| doseTwice daily | 0.95 (0.03, 35.77) | 0.975 |
| Ordinal Regression Results — Q61 | ||
| Final population, Visit 2 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.02 (0.95, 1.09) | 0.594 |
| sex2 | 2.61 (0.63, 11.46) | 0.190 |
| eduOther | 0.76 (0.15, 3.83) | 0.740 |
| empOther | 4.65 (0.79, 30.33) | 0.095 |
| income>=75k | 0.27 (0.04, 1.64) | 0.162 |
| maritalOther | 0.12 (0.02, 0.66) | 0.019 |
| insuranceOther | 0.79 (0.18, 3.34) | 0.744 |
| lineOther | 0.76 (0.19, 3.03) | 0.694 |
| doseOther | 0.32 (0.03, 3.49) | 0.343 |
| doseTwice daily | 0.82 (0.02, 27.51) | 0.911 |
## # A tibble: 30 × 9
## term estimate std.error statistic p.value conf.low conf.high coef.type Item
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <chr> <chr>
## 1 age 0.983 0.0456 -0.371 0.711 0.898 1.08 location Q37
## 2 sex2 1.22 0.809 0.250 0.802 0.233 5.95 location Q37
## 3 eduO… 0.886 1.05 -0.115 0.909 0.110 7.55 location Q37
## 4 empO… 2.23 0.971 0.828 0.408 0.353 17.7 location Q37
## 5 inco… 3.02 1.08 1.02 0.306 0.394 30.5 location Q37
## 6 mari… 1.54 1.04 0.414 0.679 0.199 12.9 location Q37
## 7 insu… 4.67 0.956 1.61 0.107 0.763 35.8 location Q37
## 8 line… 0.517 0.834 -0.791 0.429 0.0948 2.63 location Q37
## 9 dose… 0.498 1.47 -0.475 0.635 0.0169 8.03 location Q37
## 10 dose… 3.06 2.03 0.552 0.581 0.0480 197. location Q37
## # ℹ 20 more rows| Ordinal Regression Results — Q37 | ||
| Final population, Visit 3 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.98 (0.90, 1.08) | 0.711 |
| sex2 | 1.22 (0.23, 5.95) | 0.802 |
| eduOther | 0.89 (0.11, 7.55) | 0.909 |
| empOther | 2.23 (0.35, 17.74) | 0.408 |
| income>=75k | 3.02 (0.39, 30.48) | 0.306 |
| maritalOther | 1.54 (0.20, 12.87) | 0.679 |
| insuranceOther | 4.67 (0.76, 35.77) | 0.107 |
| lineOther | 0.52 (0.09, 2.63) | 0.429 |
| doseOther | 0.50 (0.02, 8.03) | 0.635 |
| doseTwice daily | 3.06 (0.05, 196.66) | 0.581 |
| Ordinal Regression Results — Q60 | ||
| Final population, Visit 3 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.94 (0.87, 1.01) | 0.092 |
| sex2 | 0.59 (0.14, 2.29) | 0.452 |
| eduOther | 2.65 (0.55, 13.77) | 0.230 |
| empOther | 6.13 (1.16, 35.92) | 0.036 |
| income>=75k | 0.72 (0.12, 4.35) | 0.713 |
| maritalOther | 0.79 (0.16, 3.89) | 0.772 |
| insuranceOther | 1.12 (0.26, 4.70) | 0.879 |
| lineOther | 1.76 (0.43, 7.51) | 0.433 |
| doseOther | 1.04 (0.11, 9.55) | 0.971 |
| doseTwice daily | 0.86 (0.02, 29.87) | 0.933 |
| Ordinal Regression Results — Q61 | ||
| Final population, Visit 3 | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.94 (0.87, 1.01) | 0.107 |
| sex2 | 0.85 (0.20, 3.46) | 0.821 |
| eduOther | 1.81 (0.39, 9.02) | 0.456 |
| empOther | 7.28 (1.38, 46.97) | 0.024 |
| income>=75k | 0.64 (0.10, 4.10) | 0.629 |
| maritalOther | 0.53 (0.10, 2.91) | 0.466 |
| insuranceOther | 0.74 (0.15, 3.40) | 0.694 |
| lineOther | 2.31 (0.60, 9.36) | 0.227 |
| doseOther | 0.68 (0.07, 6.03) | 0.724 |
| doseTwice daily | 1.96 (0.04, 76.34) | 0.714 |
(3.4.4) Ordinal Regression of composite scores final populations visits 1 2 and 3
###Add visit 2 and 3
library(dplyr)
library(ordinal)
library(broom)
domain_vars <- c(
"GI_score",
"Dermatologic_score",
"Pain_score",
"Neurologic_score",
"Constitutional_score"
)
domain_ord_results <- lapply(domain_vars, function(v) {
df <- values_clean %>%
filter(
instance == 1,
patient_id %in% final_ids,
!is.na(.data[[v]])
) %>%
mutate(
age = suppressWarnings(as.numeric(Q6)),
sex = factor(Q1),
edu = factor(edu_recoded),
emp = factor(employment_recoded),
income = factor(income_recoded),
marital = factor(marital_recoded),
insurance = factor(insurance_recoded),
line = factor(line_therapy_recoded),
dose = factor(daily_freq_recoded),
# Outcome collapsed into quartiles
outcome = factor(ntile(.data[[v]], 4), ordered = TRUE) #
)
fit <- clm(
outcome ~ age + sex + edu + emp + income +
marital + insurance + line + dose,
data = df,
link = "logit",
Hess = TRUE
)
tidy(fit, exponentiate = TRUE, conf.int = TRUE) %>%
filter(!grepl("\\|", term)) %>% # drop thresholds
mutate(Domain = gsub("_score", "", v))
}) %>%
bind_rows()
library(gt)
domain_tables <- lapply(unique(domain_ord_results$Domain), function(d) {
domain_ord_results %>%
filter(Domain == d) %>%
select(Domain, term, estimate, conf.low, conf.high, p.value) %>%
mutate(
OR_CI = sprintf("%.2f (%.2f, %.2f)", estimate, conf.low, conf.high),
p_value = ifelse(p.value < 0.001, "<0.001", sprintf("%.3f", p.value))
) %>%
select(term, OR_CI, p_value) %>%
gt() %>%
tab_header(
title = md(paste0("**Ordinal Regression Results — ", d, " Domain**")),
subtitle = md("Final population, Visit 1 (Odds Ratios with 95% CI)")
) %>%
cols_label(
term = "Covariate",
OR_CI = "Odds Ratio (95% CI)",
p_value = "p-value"
) %>%
opt_row_striping()
})
names(domain_tables) <- unique(domain_ord_results$Domain)
domain_tables$GI| Ordinal Regression Results — GI Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.99 (0.92, 1.05) | 0.648 |
| sex2 | 2.55 (0.66, 10.50) | 0.179 |
| eduOther | 1.12 (0.24, 5.22) | 0.885 |
| empOther | 3.47 (0.65, 19.27) | 0.143 |
| income>=75k | 1.27 (0.23, 7.41) | 0.785 |
| maritalOther | 0.33 (0.06, 1.63) | 0.176 |
| insuranceOther | 0.41 (0.08, 1.85) | 0.247 |
| lineOther | 1.46 (0.38, 5.83) | 0.583 |
| doseOther | 0.12 (0.01, 1.45) | 0.103 |
| doseTwice daily | 1.39 (0.03, 51.41) | 0.858 |
domain_tables$Pain| Ordinal Regression Results — Pain Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.98 (0.91, 1.05) | 0.579 |
| sex2 | 1.09 (0.30, 3.98) | 0.891 |
| eduOther | 0.96 (0.22, 4.28) | 0.956 |
| empOther | 2.41 (0.45, 13.23) | 0.299 |
| income>=75k | 0.79 (0.14, 4.36) | 0.789 |
| maritalOther | 0.73 (0.15, 3.49) | 0.689 |
| insuranceOther | 1.62 (0.39, 6.88) | 0.507 |
| lineOther | 0.58 (0.15, 2.16) | 0.416 |
| doseOther | 1.08 (0.10, 12.88) | 0.949 |
| doseTwice daily | 0.49 (0.01, 15.60) | 0.686 |
domain_tables$Neurologic| Ordinal Regression Results — Neurologic Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.98 (0.91, 1.05) | 0.509 |
| sex2 | 0.48 (0.13, 1.72) | 0.267 |
| eduOther | 0.55 (0.11, 2.64) | 0.451 |
| empOther | 0.55 (0.10, 2.79) | 0.476 |
| income>=75k | 0.49 (0.08, 2.74) | 0.420 |
| maritalOther | 0.32 (0.06, 1.64) | 0.176 |
| insuranceOther | 1.31 (0.30, 5.79) | 0.721 |
| lineOther | 0.54 (0.14, 2.04) | 0.366 |
| doseOther | 0.68 (0.07, 6.51) | 0.729 |
| doseTwice daily | 0.07 (0.00, 2.40) | 0.152 |
domain_tables$Dermatologic| Ordinal Regression Results — Dermatologic Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.03 (0.96, 1.12) | 0.373 |
| sex2 | 0.18 (0.04, 0.72) | 0.020 |
| eduOther | 0.41 (0.07, 2.00) | 0.277 |
| empOther | 0.49 (0.09, 2.52) | 0.401 |
| income>=75k | 0.54 (0.08, 3.55) | 0.522 |
| maritalOther | 2.51 (0.48, 13.37) | 0.272 |
| insuranceOther | 4.24 (1.01, 20.05) | 0.056 |
| lineOther | 0.83 (0.21, 3.19) | 0.791 |
| doseOther | 1.55 (0.12, 21.04) | 0.731 |
| doseTwice daily | 0.31 (0.01, 17.80) | 0.551 |
domain_tables$Constitutional| Ordinal Regression Results — Constitutional Domain | ||
| Final population, Visit 1 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.96 (0.88, 1.03) | 0.253 |
| sex2 | 1.95 (0.53, 7.39) | 0.317 |
| eduOther | 0.73 (0.13, 3.95) | 0.709 |
| empOther | 5.80 (1.01, 38.01) | 0.053 |
| income>=75k | 0.86 (0.16, 4.80) | 0.858 |
| maritalOther | 0.42 (0.07, 2.26) | 0.319 |
| insuranceOther | 0.90 (0.20, 3.91) | 0.891 |
| lineOther | 0.39 (0.10, 1.46) | 0.168 |
| doseOther | 0.58 (0.06, 6.67) | 0.643 |
| doseTwice daily | 9.70 (0.14, 748.32) | 0.277 |
values_clean %>%
filter(
instance == 1,
patient_id %in% final_ids,
!is.na(Pain_score)
) %>%
mutate(
quartile = ntile(Pain_score, 4)
) %>%
count(quartile)## # A tibble: 4 × 2
## quartile n
## <int> <int>
## 1 1 11
## 2 2 11
## 3 3 10
## 4 4 10domains <- c(
"GI_score",
"Dermatologic_score",
"Pain_score",
"Neurologic_score",
"Constitutional_score"
)
for (d in domains) {
p <- ggplot(
values_clean %>%
filter(
instance == 1,
patient_id %in% final_ids,
!is.na(.data[[d]])
),
aes(x = .data[[d]])
) +
geom_histogram(
binwidth = 1,
fill = "steelblue",
color = "black"
) +
theme_minimal(base_size = 14) +
labs(
title = paste("Distribution of", gsub("_score", "", d), "Domain Scores"),
subtitle = "Baseline (Visit 1), Final Population",
x = paste(gsub("_score", "", d), "Composite Score"),
y = "Count"
)
print(p)
}
library(dplyr)
library(ordinal)
library(broom)
library(gt)
domain_vars <- c(
"GI_score",
"Dermatologic_score",
"Pain_score",
"Neurologic_score",
"Constitutional_score"
)
domain_ord_results_v2 <- lapply(domain_vars, function(v) {
df <- values_clean %>%
filter(
instance == 2,
patient_id %in% final_ids,
!is.na(.data[[v]])
) %>%
mutate(
age = suppressWarnings(as.numeric(Q6)),
sex = factor(Q1),
edu = factor(edu_recoded),
emp = factor(employment_recoded),
income = factor(income_recoded),
marital = factor(marital_recoded),
insurance = factor(insurance_recoded),
line = factor(line_therapy_recoded),
dose = factor(daily_freq_recoded),
# Outcome collapsed into quartiles
outcome = factor(ntile(.data[[v]], 4), ordered = TRUE)
)
fit <- clm(
outcome ~ age + sex + edu + emp + income +
marital + insurance + line + dose,
data = df,
link = "logit",
Hess = TRUE
)
tidy(fit, exponentiate = TRUE, conf.int = TRUE) %>%
filter(!grepl("\\|", term)) %>%
mutate(Domain = gsub("_score", "", v))
}) %>%
bind_rows()
domain_tables_v2 <- lapply(unique(domain_ord_results_v2$Domain), function(d) {
domain_ord_results_v2 %>%
filter(Domain == d) %>%
select(term, estimate, conf.low, conf.high, p.value) %>%
mutate(
OR_CI = sprintf("%.2f (%.2f, %.2f)", estimate, conf.low, conf.high),
p_value = ifelse(p.value < 0.001, "<0.001",
sprintf("%.3f", p.value))
) %>%
select(term, OR_CI, p_value) %>%
gt() %>%
tab_header(
title = md(paste0("**Ordinal Regression Results — ", d, " Domain**")),
subtitle = md("Final population, Visit 2 (Odds Ratios with 95% CI)")
) %>%
cols_label(
term = "Covariate",
OR_CI = "Odds Ratio (95% CI)",
p_value = "p-value"
) %>%
opt_row_striping()
})
names(domain_tables_v2) <- unique(domain_ord_results_v2$Domain)
domain_tables_v2$GI| Ordinal Regression Results — GI Domain | ||
| Final population, Visit 2 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.00 (0.93, 1.07) | 0.986 |
| sex2 | 1.67 (0.43, 6.65) | 0.457 |
| eduOther | 0.68 (0.13, 3.50) | 0.645 |
| empOther | 1.17 (0.23, 5.91) | 0.850 |
| income>=75k | 0.73 (0.14, 3.90) | 0.715 |
| maritalOther | 0.70 (0.14, 3.32) | 0.655 |
| insuranceOther | 1.03 (0.26, 4.13) | 0.968 |
| lineOther | 0.25 (0.06, 1.00) | 0.054 |
| doseOther | 0.79 (0.08, 8.81) | 0.836 |
| doseTwice daily | 0.45 (0.01, 16.75) | 0.668 |
domain_tables_v2$Pain| Ordinal Regression Results — Pain Domain | ||
| Final population, Visit 2 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.97 (0.90, 1.04) | 0.376 |
| sex2 | 0.64 (0.17, 2.30) | 0.495 |
| eduOther | 1.23 (0.25, 6.30) | 0.800 |
| empOther | 2.30 (0.45, 12.13) | 0.314 |
| income>=75k | 0.53 (0.09, 2.93) | 0.471 |
| maritalOther | 0.79 (0.15, 3.98) | 0.775 |
| insuranceOther | 0.77 (0.16, 3.52) | 0.730 |
| lineOther | 0.95 (0.24, 3.71) | 0.939 |
| doseOther | 0.90 (0.09, 9.14) | 0.928 |
| doseTwice daily | 0.46 (0.01, 15.80) | 0.669 |
domain_tables_v2$Neurologic| Ordinal Regression Results — Neurologic Domain | ||
| Final population, Visit 2 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.01 (0.94, 1.08) | 0.823 |
| sex2 | 0.45 (0.11, 1.75) | 0.253 |
| eduOther | 1.64 (0.29, 10.16) | 0.579 |
| empOther | 1.02 (0.20, 4.88) | 0.979 |
| income>=75k | 0.34 (0.05, 1.80) | 0.218 |
| maritalOther | 0.76 (0.13, 4.42) | 0.759 |
| insuranceOther | 2.27 (0.53, 10.41) | 0.273 |
| lineOther | 1.23 (0.32, 4.81) | 0.764 |
| doseOther | 0.13 (0.00, 1.70) | 0.146 |
| doseTwice daily | 1.27 (0.04, 61.49) | 0.897 |
domain_tables_v2$Dermatologic| Ordinal Regression Results — Dermatologic Domain | ||
| Final population, Visit 2 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.08 (1.01, 1.17) | 0.034 |
| sex2 | 1.40 (0.35, 6.00) | 0.640 |
| eduOther | 0.09 (0.01, 0.52) | 0.010 |
| empOther | 1.54 (0.28, 8.81) | 0.616 |
| income>=75k | 27.90 (3.92, 261.40) | 0.002 |
| maritalOther | 2.14 (0.43, 11.38) | 0.360 |
| insuranceOther | 0.44 (0.09, 1.97) | 0.289 |
| lineOther | 0.62 (0.15, 2.47) | 0.498 |
| doseOther | 0.54 (0.03, 7.27) | 0.644 |
| doseTwice daily | 2.92 (0.05, 189.89) | 0.594 |
domain_tables_v2$Constitutional| Ordinal Regression Results — Constitutional Domain | ||
| Final population, Visit 2 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.04 (0.98, 1.12) | 0.206 |
| sex2 | 1.28 (0.32, 5.14) | 0.723 |
| eduOther | 0.23 (0.04, 1.20) | 0.091 |
| empOther | 1.44 (0.30, 6.99) | 0.643 |
| income>=75k | 0.18 (0.03, 1.01) | 0.057 |
| maritalOther | 0.11 (0.02, 0.58) | 0.013 |
| insuranceOther | 1.11 (0.25, 4.93) | 0.891 |
| lineOther | 0.45 (0.11, 1.83) | 0.272 |
| doseOther | 0.35 (0.03, 4.54) | 0.402 |
| doseTwice daily | 0.82 (0.03, 36.37) | 0.911 |
library(dplyr)
library(ordinal)
library(broom)
library(gt)
domain_ord_results_v3 <- lapply(domain_vars, function(v) {
df <- values_clean %>%
filter(
instance == 3,
patient_id %in% final_ids,
!is.na(.data[[v]])
) %>%
mutate(
age = suppressWarnings(as.numeric(Q6)),
sex = factor(Q1),
edu = factor(edu_recoded),
emp = factor(employment_recoded),
income = factor(income_recoded),
marital = factor(marital_recoded),
insurance = factor(insurance_recoded),
line = factor(line_therapy_recoded),
dose = factor(daily_freq_recoded),
outcome = factor(ntile(.data[[v]], 4), ordered = TRUE)
)
fit <- clm(
outcome ~ age + sex + edu + emp + income +
marital + insurance + line + dose,
data = df,
link = "logit",
Hess = TRUE
)
tidy(fit, exponentiate = TRUE, conf.int = TRUE) %>%
filter(!grepl("\\|", term)) %>%
mutate(Domain = gsub("_score", "", v))
}) %>%
bind_rows()
domain_tables_v3 <- lapply(unique(domain_ord_results_v3$Domain), function(d) {
domain_ord_results_v3 %>%
filter(Domain == d) %>%
select(term, estimate, conf.low, conf.high, p.value) %>%
mutate(
OR_CI = sprintf("%.2f (%.2f, %.2f)", estimate, conf.low, conf.high),
p_value = ifelse(p.value < 0.001, "<0.001",
sprintf("%.3f", p.value))
) %>%
select(term, OR_CI, p_value) %>%
gt() %>%
tab_header(
title = md(paste0("**Ordinal Regression Results — ", d, " Domain**")),
subtitle = md("Final population, Visit 3 (Odds Ratios with 95% CI)")
) %>%
cols_label(
term = "Covariate",
OR_CI = "Odds Ratio (95% CI)",
p_value = "p-value"
) %>%
opt_row_striping()
})
names(domain_tables_v3) <- unique(domain_ord_results_v3$Domain)
domain_tables_v3$GI| Ordinal Regression Results — GI Domain | ||
| Final population, Visit 3 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.02 (0.96, 1.10) | 0.509 |
| sex2 | 2.47 (0.67, 9.55) | 0.178 |
| eduOther | 1.33 (0.28, 6.67) | 0.721 |
| empOther | 1.65 (0.37, 7.48) | 0.510 |
| income>=75k | 0.58 (0.11, 3.00) | 0.518 |
| maritalOther | 0.96 (0.20, 4.55) | 0.961 |
| insuranceOther | 1.22 (0.26, 5.66) | 0.800 |
| lineOther | 1.29 (0.34, 4.96) | 0.711 |
| doseOther | 0.06 (0.00, 0.70) | 0.037 |
| doseTwice daily | 0.17 (0.00, 5.18) | 0.311 |
domain_tables_v3$Pain| Ordinal Regression Results — Pain Domain | ||
| Final population, Visit 3 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.93 (0.86, 1.00) | 0.072 |
| sex2 | 0.47 (0.11, 1.84) | 0.290 |
| eduOther | 1.93 (0.37, 10.78) | 0.437 |
| empOther | 3.13 (0.54, 18.80) | 0.202 |
| income>=75k | 0.73 (0.13, 3.92) | 0.713 |
| maritalOther | 1.66 (0.31, 8.95) | 0.552 |
| insuranceOther | 2.53 (0.49, 14.59) | 0.279 |
| lineOther | 0.65 (0.16, 2.61) | 0.547 |
| doseOther | 0.27 (0.02, 2.81) | 0.275 |
| doseTwice daily | 2.13 (0.07, 67.70) | 0.660 |
domain_tables_v3$Neurologic| Ordinal Regression Results — Neurologic Domain | ||
| Final population, Visit 3 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.97 (0.90, 1.03) | 0.283 |
| sex2 | 0.46 (0.12, 1.66) | 0.238 |
| eduOther | 2.26 (0.44, 13.07) | 0.342 |
| empOther | 1.36 (0.30, 6.33) | 0.691 |
| income>=75k | 1.13 (0.22, 5.78) | 0.883 |
| maritalOther | 0.87 (0.17, 4.62) | 0.872 |
| insuranceOther | 2.90 (0.71, 12.86) | 0.144 |
| lineOther | 1.05 (0.26, 4.20) | 0.946 |
| doseOther | 0.69 (0.06, 6.54) | 0.743 |
| doseTwice daily | 0.48 (0.01, 20.21) | 0.698 |
domain_tables_v3$Dermatologic| Ordinal Regression Results — Dermatologic Domain | ||
| Final population, Visit 3 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 1.08 (1.00, 1.18) | 0.053 |
| sex2 | 2.58 (0.61, 12.40) | 0.214 |
| eduOther | 0.55 (0.11, 2.70) | 0.458 |
| empOther | 0.74 (0.13, 4.17) | 0.731 |
| income>=75k | 11.40 (1.75, 85.86) | 0.013 |
| maritalOther | 7.81 (1.58, 45.02) | 0.015 |
| insuranceOther | 1.24 (0.28, 5.46) | 0.771 |
| lineOther | 1.08 (0.27, 4.27) | 0.914 |
| doseOther | 4.71 (0.40, 64.02) | 0.219 |
| doseTwice daily | 0.27 (0.01, 10.05) | 0.477 |
domain_tables_v3$Constitutional| Ordinal Regression Results — Constitutional Domain | ||
| Final population, Visit 3 (Odds Ratios with 95% CI) | ||
| Covariate | Odds Ratio (95% CI) | p-value |
|---|---|---|
| age | 0.88 (0.80, 0.96) | 0.005 |
| sex2 | 0.67 (0.16, 2.68) | 0.578 |
| eduOther | 11.58 (1.76, 103.14) | 0.017 |
| empOther | 14.52 (2.45, 108.77) | 0.005 |
| income>=75k | 0.45 (0.08, 2.45) | 0.355 |
| maritalOther | 0.54 (0.09, 3.20) | 0.495 |
| insuranceOther | 3.04 (0.58, 17.65) | 0.195 |
| lineOther | 0.63 (0.15, 2.54) | 0.515 |
| doseOther | 0.04 (0.00, 0.48) | 0.016 |
| doseTwice daily | 21.72 (0.33, 1573.50) | 0.136 |
(3.4.5) Sankey plot of Individual questions
(3.4.6) Linear Regression of Pooled Composite Scores
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_PRO)
##
## Residuals:
## Min 1Q Median 3Q Max
## -25.795 -8.988 -0.753 5.585 58.487
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 79.3147 23.4012 3.389 0.0021 **
## age -0.4247 0.3731 -1.138 0.2647
## sexFemale 0.4162 7.3631 0.057 0.9553
## eduOther 2.9437 8.6834 0.339 0.7371
## empOther 9.4084 9.2337 1.019 0.3170
## income>=75k -3.8378 9.8181 -0.391 0.6988
## maritalOther -5.7604 9.1010 -0.633 0.5319
## insuranceOther 2.9680 8.0090 0.371 0.7137
## lineOther -4.5230 7.3226 -0.618 0.5418
## months>=13 months -1.1630 7.4979 -0.155 0.8779
## monthsMissing 45.5393 22.6714 2.009 0.0543 .
## doseOther -5.9408 13.4858 -0.441 0.6629
## doseTwice daily -3.0115 19.7453 -0.153 0.8799
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 20 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.2656, Adjusted R-squared: -0.04917
## F-statistic: 0.8438 on 12 and 28 DF, p-value: 0.6079
(3.4.7) Linear Regression of Domain Composite Scores
##
## ==============================
## DOMAIN: GI_score
## ==============================
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_domain)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5.158 -2.027 0.000 1.619 6.894
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 12.09137 3.68061 3.285 0.00274 **
## age -0.06541 0.05868 -1.115 0.27447
## sexFemale 1.97720 1.15809 1.707 0.09884 .
## eduOther 0.76850 1.36575 0.563 0.57812
## empOther 2.09506 1.45231 1.443 0.16023
## income>=75k -0.65181 1.54421 -0.422 0.67618
## maritalOther -2.71098 1.43144 -1.894 0.06861 .
## insuranceOther -0.97184 1.25968 -0.771 0.44688
## lineOther 0.01430 1.15172 0.012 0.99018
## months>=13 months 0.61280 1.17929 0.520 0.60740
## monthsMissing 17.86634 3.56582 5.010 2.7e-05 ***
## doseOther -4.14073 2.12108 -1.952 0.06098 .
## doseTwice daily 0.54877 3.10560 0.177 0.86101
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.146 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.6237, Adjusted R-squared: 0.4625
## F-statistic: 3.868 on 12 and 28 DF, p-value: 0.001546
##
## ==============================
## DOMAIN: Dermatologic_score
## ==============================
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_domain)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.7821 -0.1394 0.1829 0.3560 1.0081
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.587781 1.090860 5.122 1.98e-05 ***
## age 0.007704 0.017393 0.443 0.661
## sexFemale -0.386076 0.343234 -1.125 0.270
## eduOther -0.384628 0.404781 -0.950 0.350
## empOther -0.045117 0.430434 -0.105 0.917
## income>=75k 0.145251 0.457673 0.317 0.753
## maritalOther 0.545225 0.424249 1.285 0.209
## insuranceOther -0.204823 0.373345 -0.549 0.588
## lineOther 0.012721 0.341346 0.037 0.971
## months>=13 months -0.193389 0.349517 -0.553 0.584
## monthsMissing 0.289494 1.056838 0.274 0.786
## doseOther 0.689347 0.628645 1.097 0.282
## doseTwice daily 0.319359 0.920438 0.347 0.731
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9324 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.2015, Adjusted R-squared: -0.1406
## F-statistic: 0.589 on 12 and 28 DF, p-value: 0.8325
##
## ==============================
## DOMAIN: Pain_score
## ==============================
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_domain)
##
## Residuals:
## Min 1Q Median 3Q Max
## -17.462 -8.340 -0.125 3.868 39.918
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 38.6113 16.0918 2.399 0.0233 *
## age -0.2360 0.2566 -0.920 0.3655
## sexFemale -1.2182 5.0632 -0.241 0.8116
## eduOther 1.8431 5.9711 0.309 0.7599
## empOther 5.0400 6.3496 0.794 0.4340
## income>=75k -1.0246 6.7514 -0.152 0.8805
## maritalOther -0.6577 6.2583 -0.105 0.9170
## insuranceOther 3.7782 5.5074 0.686 0.4983
## lineOther -3.0181 5.0354 -0.599 0.5537
## months>=13 months -1.7985 5.1559 -0.349 0.7298
## monthsMissing 24.6636 15.5900 1.582 0.1249
## doseOther -0.5294 9.2735 -0.057 0.9549
## doseTwice daily -2.3094 13.5778 -0.170 0.8662
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 13.75 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.1876, Adjusted R-squared: -0.1605
## F-statistic: 0.539 on 12 and 28 DF, p-value: 0.8704
##
## ==============================
## DOMAIN: Neurologic_score
## ==============================
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_domain)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.0978 -1.9024 -0.6009 0.9250 8.4517
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 11.55553 4.07359 2.837 0.00838 **
## age -0.05078 0.06495 -0.782 0.44090
## sexFemale -0.82513 1.28174 -0.644 0.52497
## eduOther 0.24408 1.51157 0.161 0.87288
## empOther -0.48601 1.60737 -0.302 0.76461
## income>=75k -1.89187 1.70908 -1.107 0.27774
## maritalOther -1.74734 1.58427 -1.103 0.27945
## insuranceOther 0.11692 1.39418 0.084 0.93376
## lineOther -0.08192 1.27469 -0.064 0.94921
## months>=13 months -0.36592 1.30520 -0.280 0.78127
## monthsMissing 1.47877 3.94654 0.375 0.71071
## doseOther -1.12278 2.34755 -0.478 0.63617
## doseTwice daily -3.74674 3.43718 -1.090 0.28498
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.482 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.1473, Adjusted R-squared: -0.2182
## F-statistic: 0.403 on 12 and 28 DF, p-value: 0.9503
##
## ==============================
## DOMAIN: Constitutional_score
## ==============================
##
## Call:
## lm(formula = PRO_score ~ age + sex + edu + emp + income + marital +
## insurance + line + months + dose, data = baseline_domain)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.2408 -1.8826 -0.1184 1.6216 8.7712
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 11.46875 4.23898 2.706 0.0115 *
## age -0.08017 0.06759 -1.186 0.2455
## sexFemale 0.86837 1.33378 0.651 0.5203
## eduOther 0.47270 1.57294 0.301 0.7660
## empOther 2.80445 1.67263 1.677 0.1047
## income>=75k -0.41468 1.77848 -0.233 0.8173
## maritalOther -1.18961 1.64859 -0.722 0.4765
## insuranceOther 0.24963 1.45079 0.172 0.8646
## lineOther -1.44999 1.32644 -1.093 0.2836
## months>=13 months 0.58205 1.35819 0.429 0.6715
## monthsMissing 1.24103 4.10678 0.302 0.7647
## doseOther -0.83729 2.44286 -0.343 0.7343
## doseTwice daily 2.17659 3.57674 0.609 0.5477
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.623 on 28 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.2214, Adjusted R-squared: -0.1123
## F-statistic: 0.6634 on 12 and 28 DF, p-value: 0.7704
Section 4
(4.1) Binned plots by question (Item score vs adherence)
## [[1]]
##
## [[2]]
##
## [[3]]
##
## [[4]]
##
## [[5]]
##
## [[6]]
##
## [[7]]
##
## [[8]]
##
## [[9]]
##
## [[10]]
##
## [[11]]
##
## [[12]]
##
## [[13]]
##
## [[14]]
##
## [[15]]
##
## [[16]]
##
## [[17]]
##
## [[18]]
##
## [[19]]
##
## [[20]]
##
## [[21]]
##
## [[22]]
##
## [[23]]
##
## [[24]]
##
## [[25]]
##
## [[26]]
##
## [[27]]
##
## [[28]]
##
## [[29]]
##
## [[30]]
###Domain Emax Curves
(4.2) Spearman correlation
(4.3) Add Table for Correaltions
library(dplyr)
library(tidyr)
library(gt)
# -----------------------------
# Domains
# -----------------------------
domains <- c(
"GI_score",
"Dermatologic_score",
"Pain_score",
"Neurologic_score",
"Constitutional_score"
)
# -----------------------------
# Final population
# -----------------------------
final_pop <- values_clean %>%
filter(instance %in% c(1, 2, 3)) %>%
group_by(patient_id) %>%
summarise(n_visits = n_distinct(instance), .groups = "drop") %>%
filter(n_visits == 3) %>%
pull(patient_id)
# -----------------------------
# Collect correlations
# -----------------------------
corr_results <- list()
for (d in domains) {
for (v in c(1, 2, 3)) {
df <- values_clean %>%
filter(
patient_id %in% final_pop,
instance == v,
!is.na(wilson_adherence),
!is.na(.data[[d]])
)
if (nrow(df) < 5) next
ct <- cor.test(
df$wilson_adherence,
df[[d]],
method = "spearman",
exact = FALSE
)
corr_results[[length(corr_results) + 1]] <- tibble(
Domain = gsub("_score", "", d),
Visit = paste0("Visit ", v),
r = round(ct$estimate, 2),
p = ct$p.value
)
}
}
corr_results <- bind_rows(corr_results) %>%
mutate(
p_value = ifelse(p < 0.001, "<0.001", sprintf("%.3f", p))
)
# -----------------------------
# Wide-format table
# -----------------------------
corr_table <- corr_results %>%
select(Domain, Visit, r, p_value) %>%
pivot_wider(
names_from = Visit,
values_from = c(r, p_value),
names_glue = "{Visit}_{.value}"
) %>%
rename(
`Visit 1 r` = `Visit 1_r`,
`Visit 1 p-value` = `Visit 1_p_value`,
`Visit 2 r` = `Visit 2_r`,
`Visit 2 p-value` = `Visit 2_p_value`,
`Visit 3 r` = `Visit 3_r`,
`Visit 3 p-value` = `Visit 3_p_value`
)
# -----------------------------
# GT table
# -----------------------------
corr_gt <- corr_table %>%
gt() %>%
tab_header(
title = md("**Correlation Between Adherence and Domain Scores**"),
subtitle = md("Spearman correlation by visit (final population)")
) %>%
cols_align(align = "center", -Domain) %>%
opt_row_striping()
corr_gt| Correlation Between Adherence and Domain Scores | ||||||
| Spearman correlation by visit (final population) | ||||||
| Domain | Visit 1 r | Visit 2 r | Visit 3 r | Visit 1 p-value | Visit 2 p-value | Visit 3 p-value |
|---|---|---|---|---|---|---|
| GI | 0.13 | -0.12 | 0.20 | 0.412 | 0.454 | 0.216 |
| Dermatologic | -0.18 | 0.25 | 0.02 | 0.265 | 0.115 | 0.903 |
| Pain | 0.14 | 0.14 | 0.18 | 0.380 | 0.389 | 0.248 |
| Neurologic | 0.00 | 0.07 | 0.10 | 0.987 | 0.658 | 0.512 |
| Constitutional | 0.14 | -0.12 | -0.03 | 0.382 | 0.453 | 0.856 |
Export composite scores
# ============================================================
# EXPORT ANALYSIS DATASET WITH COMPOSITE SCORES
# ============================================================
library(dplyr)
library(writexl)
analysis_dataset <- values_clean %>%
dplyr::select(
# Identifiers
patient_id,
instance,
relative_date,
# Adherence
wilson_adherence,
# Composite scores
AE_score,
GI_score,
Dermatologic_score,
Pain_score,
Neurologic_score,
Constitutional_score,
# Demographics / covariates
age = Q6,
sex = Q1,
edu_recoded,
employment_recoded,
income_recoded,
marital_recoded,
insurance_recoded,
line_therapy_recoded,
daily_freq_recoded,
months_therapy_recoded
) %>%
arrange(patient_id, instance)
# ---- Export to CSV ----
write.csv(
analysis_dataset,
file = "analysis_dataset_with_composite_scores.csv",
row.names = FALSE
)
# ---- Optional Excel version (often preferred by advisors) ----
write_xlsx(
analysis_dataset,
path = "analysis_dataset_with_composite_scores.xlsx"
)Note that the echo = FALSE parameter was added to the
code chunk to prevent printing of the R code that generated the
plot.