Definitions
Trial
Trial 1 has follow-up at pre-op (FU.tr.1) and baseline 6 months
earlier (patients from LAPPRO, QoliCOL, QoliRECT enrolled).
Trial 2 has follow-up at 3 months (FU.tr.2) and baseline at pre-op
(patients from LAPPRO enrolled)
Trial 3 has follow-up at 12 months (FU.tr.3) and baseline at pre-op
(patients from LAPPRO, QoliCOL, QoliRECT enrolled)
Trial 4 has follow-up at 24 months (FU.tr.4) and baseline at 12
months (patients from LAPPRO, QoliRECT enrolled).
If follow-up date is missing, but there is no pick-up after time zero
patient will correctly end up in Control group
If follow-up date is missing but there is a pick-up after time zero
patient will have missing on Group assignment. Since follow-up is (in
theory) 3, 12 and 12 months after time zero in trial 2, 3 and 4,
respectively, we could set follow-up to time-zero + 3 or 12, but since
we don´t know exactly, we use a cut-off of half, that is 3/2 and
months, that is:
- If Follow-up missing and pick-up is within (time zero + 9 months) then GROUP=ACTIVE
- If Follow-up missing and pick-up is > (time zero + 9 months) then EXCLUDE
Patients with no date available for time zero could not be determined
whether to include or exclude (“NA”) and was therefore excluded.
Data fix
Patients with no expedition should get their expedition date
‘1970-01-01’. If there is an empty line for a patient, then it is
assumed he/she did not get BB (we code it as they picked up the
prescription 1970-01-01 ).
Study flow chart and demographics and patient characteristics
LPR: 3705 patients; QL: 1677 patients, QR: 852 patients. (total =
6234 patients)
Trial 3: 5598 patients has available date for both time zero and
follow-up
Trial 4: 3908 patients has available date for both time zero and
follow-up
Among those included, which will be included in the respective
group?
Trial 1: 1114+9 excluded, 5111 included
Trial 2: 564 excluded, 3141 included
Trial 3: 1263+9 excluded, 4962 included (1263 had BB before and 9 had no
time zero data)
Trial 4: 721+392 excluded, 3444 included (721 had BB before and 392 had
no time zero data)
Among those included:
Trial3: 69 Active, 4873 Control (20 could not be determined due to
missing FU date and BB not within 9 months after time zero).
Trial4: 95 Active, 3349 Control
LAPPRO: 3705 patients; QoliCOL: 1677 patients, QoliRECT: 852
patients. (total = 6234 patients)
For trial 3: (baseline at pre-op and follow-up at 12 months), Of 6234
enrolled patients (LAPPRO, QoliCOL, QoliRECT)
For trial 4: (baseline at pre-op and follow-up at 12 months), Of 4557
enrolled patients (LAPPRO, QoliRECT)
Q: What to present in demography table?
Statistical model
Endpoints modelled using ordinal regression assuming proportional
odds. Weakly regularizing priors (\(\beta \sim
N(0, 1)\)) on the logit scale were used for treatment factor
(Active vs Control). Baseline value of endpoints was included, with a
monotone effect using a Dirichlet distribution.
Variables adjusted för are: - Binge drinking (Drinking at least 6
glass on the same occassion, YES or NO) - Age - Mental health status
(History of psychiatric health issues of prescription of
anti-depressant)
Imputations using MICE (5 imputed data sets).
Study results
Analyse and present results for trial 3 and 4.
Primary:
How often in the last month have there been negative thoughts about
your cancer suddenly and by itself?
Co-primary:
How severe has the sudden negative intrusive thoughts been?
Secondary 1:
Depressed mood
Secondary 2:
How would you like to describe your quality of life in the last
month?
Secondary 3:
Have you experienced periods of intense anxiety, worry or
panic
Secondary 4:
Have you woken up at any time during the night with worry, anxiety
or an unpleasant feeling the last month?
Main analysis of endpoints
Odds ratio describes the relative (Active vs Control) of more worse
outcomes. That is large values (>1) of odds ratio means that patients
on Active are worse off compared to Control. Small values (<1) of
odds ratio means that patients on Active are better off compared to
Control.
Patients assigned to Active group had higher tendency to poor QoL (OR
2.0 and 1.6 for trial 3 and 4, respectively. Patients assigned to Active
had more periods of intense axiety, worry and panic in trial 4
(OR=2.4)
Primary:
How often in the last month have there been negative thoughts
about your cancer suddenly and by itself?
Large value of odds ratio favors Control. Posterior probability that
odds ratio > 1 is 19%.
Co-primary:
How severe has the sudden negative intrusive thoughts
been?
Secondary 1:
Depressed mood
There is a higher probabaility that patients on Active has depressed
mood.
Secondary 2:
How would you like to describe your quality of life in the last
month?
Patients on Active has poorer QoL
Secondary 3:
Have you experienced periods of intense anxiety, worry or
panic
Patients on Active has more anxiety and worry
Secondary 4:
Have you woken up at any time during the night with worry,
anxiety or an unpleasant feeling the last month?
Supplementary material
Supportive sensitivity analyses
Unadjusted and frequentist analysis
Unadjusted analysis is model with only treatment factor (Active or
Control) is in the model (complete cases, no imputations).
Frequentist analysis use same adjustment variables as main analysis
as well as pooled across the 5 imputed data sets
Prevalent users analysis
Patient assigned to Active if picked up at least one prescription any
time before follow-up. Patient assigned to Control if picked up no
prescription any time before follow-up.
If follow-up date is missing but there is a pick-up after time zero
patient will have missing on Group assignment. Since follow-up is (in
theory) 3, 12 and 12 months after time zero in trial 2, 3 and 4,
respectively, we could set follow-up to time-zero + 3 or 12, but since
we don´t know exactly, we use a cut-off of half, that is 3/2 and
months, that is:
- If Follow-up missing and pick-up is within (time zero + 9 months) then GROUP=ACTIVE
- If Follow-up missing and pick-up is > (time zero + 9 months) then EXCLUDE
Unadjusted analysis.
Trial 3
t1
##
## Control Active
## 4781 1439
Trial 4
t2
##
## Control Active
## 3594 887
MCMC diagnostics
Primary:
How often in the last month have there been negative thoughts
about your cancer suddenly and by itself?
Co-primary:
How severe has the sudden negative intrusive thoughts
been?
Secondary 1:
Depressed mood
Secondary 2:
How would you like to describe your quality of life in the last
month?
Secondary 3:
Have you experienced periods of intense anxiety, worry or
panic
Secondary 4:
Have you woken up at any time during the night with worry,
anxiety or an unpleasant feeling the last month?
