Content uploaded by Qiuling Shi
Author content
All content in this area was uploaded by Qiuling Shi on Apr 04, 2022
Content may be subject to copyright.
ORIGINAL ARTICLE – THORACIC ONCOLOGY
Establishment of Minimal Clinically Important Improvement
for Patient-Reported Symptoms to Define Recovery After Video-
Assisted Thoracoscopic Surgery
Wei Xu, MS
1
, Wei Dai, MD
2
, Zhen Gao, MD
3,4
, Xin Shelley Wang, MD, MPH
5
, Li Tang, MD, MPH
6
,
Yang Pu, MS
1
, Qingsong Yu, Bsc
1
, Hongfan Yu, Bsc
1
, Yuxian Nie, MEc
6
, Weitao Zhuang, MD
3
,
Guibin Qiao, MD
3,4
, Charles S. Cleeland, PhD
5
, and Qiuling Shi, MD, PhD
1,6,7
1
School of Public Health and Management, Chongqing Medical University, Chongqing, China;
2
Department of Thoracic
Surgery, Sichuan Cancer Center, School of Medicine, Sichuan Cancer Hospital and Institute, University of Electronic
Science and Technology of China, Chengdu, China;
3
Department of Thoracic Surgery, Guangdong Provincial People’s
Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China;
4
The Second School of Clinical Medicine,
Southern Medical University, Guangzhou, China;
5
Department of Symptom Research, The University of Texas MD
Anderson Cancer Center, Houston, TX;
6
State Key Laboratory of Ultrasound in Medicine and Engineering, Chongqing
Medical University, Chongqing, China;
7
Center for Cancer Prevention Research, Sichuan Cancer Center, School of
Medicine, Sichuan Cancer Hospital and Institute, University of Electronic Science and Technology of China, Chengdu,
Sichuan, China
ABSTRACT
Purpose. The aim of this study was to define a threshold
of minimal clinically important improvement (MCII) for
interpreting patient condition following video-assisted
thoracoscopic surgery (VATS).
Methods. Patients undergoing VATS were recruited for
this multicenter, prospective, observational cohort study.
Symptoms were measured using the MD Anderson
Symptom Inventory–Lung Cancer Module perioperatively.
To define MCIIs, we first identified index symptoms,
defined as the most severe symptoms showing the largest
reduction from day 1 post-surgery to discharge. MCIIs for
each index symptom were then obtained via an anchor-
based approach. Symptom recovery was defined as an
MCII after post-surgery day 1. Cox regression models were
used to identify risk factors for unrecovered index
symptoms.
Results. Using 366 patients, we identified pain and fatigue
as index symptoms after VATS. MCII was defined as a
30% reduction in pain or fatigue. At discharge, 22.6% of
patients had not recovered from pain and 22.4% had not
recovered from fatigue. Cox models found that risk factors
for unrecovered pain were Charlson Comorbidity Index
score C1 (hazard ratio [HR] 1.36, 95% confidence interval
[CI] 1.04–1.77; p=0.02) and preoperative neoadjuvant
therapy (HR 2.78, 95% CI 1.13–6.83; p=0.02). Malig-
nancy was a risk factor for unrecovered fatigue (HR 1.47,
95% CI 1.02–2.13; p=0.04).
Conclusion. Pain and fatigue can be used as index mea-
sures for symptom recovery in patients following VATS. A
30% MCII represented meaningful recovery after VATS
and could identify patients who may need extensive care
after discharge.
Video-assisted thoracoscopic surgery (VATS) is a
minimally invasive technique that has been increasingly
used for treating lung cancer
1
and has emerged as an
alternative for advanced resections.
2
Currently, VATS is
employed in 80% of lung cancer surgeries.
3
Studies
showed that compared with thoracotomy, VATS for lung
cancer had comparable or better oncologic outcomes.
4,5
Furthermore, VATS is associated with a lower
ÓSociety of Surgical Oncology 2022
First Received: 23 September 2021
Accepted: 25 January 2022
Q. Shi, MD, PhD
e-mail: qshi@cqmu.edu.cn
Ann Surg Oncol
https://doi.org/10.1245/s10434-022-11629-7
postoperative symptom burden and lower functional
interference both in early-stage and locally advanced lung
cancers, which is beneficial for postoperative recovery.
2,6
However, postoperative symptom and functional recovery
is not directly observable or quantifiable.
7
Studies of
interventions aimed at improving recovery commonly rely
on ‘proxy’ measurements such as length of hospital stay
(LOS) and complication rates,
8
which are of interest to
clinicians but do not reflect the complexity of the recovery
process and fail to capture the patient’s perspective.
9
Patient-reported outcomes (PROs) rely on self-assess-
ments that directly measure symptoms and functional
recovery post-surgery
10
and can be useful for improving
patient care.
9
Recently, the use of patient-reported symp-
toms as a measure of postoperative recovery has
increased.
11,12
However, systematic reviews of research
using PROs revealed a lack of measurable properties
9
that
were required by the International Society for Quality of
Life Research recommendation for using PROs in com-
parative effectiveness studies.
13
For example, it was
unclear whether a change in score was meaningful from a
patient’s perspective among patients who underwent
VATS, resulting in delayed implementation of PROs in
both research and practice. Although a statistically signif-
icant change in PRO score was commonly reported when
two surgical approaches were compared,
11
such a change
did not identify critical PROs with clinical significance
from the patient’s or clinician’s viewpoint, leading to dis-
cussions in the US FDA on the interpretation of PRO
results.
14
The concept of minimal clinically important improve-
ment (MCII) was created to ascertain the smallest change
that patients perceive as beneficial.
15
This PRO change is
defined as the minimal change representing a clinically
important improvement from the patient’s perspective.
16
For clinicians, using MCII helps guide treatment, which is
critical in understanding efficacy, interpreting study results,
and managing the recovery of patients who have undergone
VATS.
17
For example, recent research determined that in
patients with chronic obstructive pulmonary disease,
changes of 4.7 on the Functional Assessment of Chronic
Illness Therapy-Fatigue Subscale (FACIT-FS) and 3.8 on
the modified FACIT-FS represented clinically relevant
improvements in fatigue following pulmonary rehabilita-
tion.
18
However, as direct clinical measures of
postoperative recovery,
19
patient-reported symptoms in the
surgical settings have not been well defined or interpreted.
Furthermore, key patient-reported symptoms and the extent
to which they should inform clinical action from both
patient and clinician perspectives have not been estab-
lished. As a result, recovery status and timing remain
unknown, even when a patient is eligible for discharge.
Using a longitudinal study to collect PRO data over 4
weeks of the recovery course after VATS, we aimed to (1)
identify the symptoms with the largest reduction during
hospital stay as an index for defining recovery after VATS;
(2) determine and validate the MCII thresholds of the index
symptoms using an anchor-based method; (3) define patient
recovery status using the index symptoms and their
thresholds; and (4) estimate the timing of and factors
impacting MCII-defined recovery after VATS.
METHODS
Patients
We conducted a multicenter, prospective, observational
cohort study (CN-PRO-Lung 1) to measure the periopera-
tive symptom burden in patients who underwent lung
surgery. The study protocol was reviewed and approved by
the Ethics Committees of the six participating hospitals and
has been previously published.
20
All participants provided
written informed consent. Included in the study were
patients who underwent VATS during November 2017
through January 2020 across six hospitals in China.
Enhanced Recovery After Surgery (ERAS) protocols were
not routinely implemented at any participating center and
none of the patients underwent the ERAS pathway.
Epidurals or other types of regional blocks were not used in
any patients.
The external validation dataset was extracted from the
electronic patient records of the Department of Thoracic
Surgery at the Guangdong Provincial People’s Hospital,
where the MD Anderson Symptom Inventory–Lung Cancer
Module (MDASI-LC) is routinely evaluated in patients
who undergo lung surgery. The retrospective data analysis
protocol was approved, and informed consent was waived
by the Ethics Committee of Guangdong Provincial Peo-
ple’s Hospital.
Symptom Measurement and Clinical Data
Symptoms were assessed using the MDASI-LC, a valid,
reliable, and concise tool for measuring symptom severity
and interference with functioning whose reliability and
validity have been verified in Chinese patients.
21
After
enrollment, patients were provided with the MDASI-LC
22
to rate each of the 16 symptoms, along with six
W. Xu et al.
interference items measuring symptom interference with
daily life (physical: general activity, work, and walking;
affective: relations with others, mood, and enjoyment of
life). Each symptom or interference item was rated ‘at its
worst’ in the previous 24 h on a 0–10 numerical scale, with
0 representing ‘not present’/‘did not interfere’ and 10
representing ‘as bad as you can imagine’/‘interfered
completely’.
These MDASI-LC data (PROs) were collected preop-
eratively, daily postoperatively during the in-hospital stay,
and weekly after discharge from the hospital for 4 weeks
after surgery or, if adjuvant therapy was administered, up
to the start of adjuvant therapy. Research staff extracted
information about demographics, surgery, anesthesia,
postoperative care, perioperative complications, and pain
management from patient medical records.
Statistical Analysis
Descriptive statistics (means and standard deviations)
were used to summarize the participants’ demographic
factors and PROs. We then conducted two independent
sample t-tests (e.g., age, body mass index) or the Wil-
coxon–Mann–Whitney test (e.g., length of postoperative
hospital stay) for continuous variables and the Chi-square
test or two-tailed Fisher exact test for categorical variables
to compare the patients with follow-up data available and
the patients with missing data at discharge. Next, we used
the mean MDASI-LC symptom scores during postopera-
tive hospitalization to identify the five most severe
symptoms, among which we defined index symptoms as
those sensitive to change from postoperative day (POD) 1
to discharge. This sensitivity to change was measured via
standardized response means of relative and absolute dif-
ferences between POD 1 and discharge in terms of
symptom scores. A standardized response means of C0.50
was considered acceptable for each.
23
Furthermore, a
paired t-test was performed to compare symptom scores
between POD 1 and discharge.
Patients who scored
24
their index symptoms as moderate
to severe (4–10) on POD 1 were included for MCII
determination. We calculated the changes in symptom and
interference scores between POD 1 (when symptoms
peaked) and discharge. If a patient’s stay in hospital
exceeded 2 weeks, we used the POD 14 symptom score as
the discharge score. Patients who were missing POD 1 or
discharge data for any of the index symptoms or six
interference items were excluded from the analysis.
Relative and absolute MCIIs were calculated for each
index symptom using the following formulae.
Relative change %ðÞ
¼symptom scores at discharge symptom scores on POD 1
symptom scores on POD 1
100%
Absolute change ¼symptom scores at discharge
symptom scores on POD 1
We used multivariate analysis of variance (MANOVA)
to determine MCII, with MDASI-LC interference as the
anchor, adjusted by American Society of Anesthesiologists
(ASA) physical status classification. A 0.3 or higher
correlation coefficient between the interference items and
symptoms was considered appropriate for the anchor.
25
The decreases in scores of index symptoms were
categorized using cut-off points (from 10 to 90% for
relative change and 1 to 9 for absolute change) to generate
variables of improvement (improved, greater than or equal
to the cut-off point; non-improved, less than the cut-off
point). For each MANOVA, the POD 1 to discharge
changes in scores of the six interference variables were
used as dependent variables, and each of the dichotomized
improvement variables was used as an independent
variable. The cut-off point generating the improvement
variable with the largest Fvalue in MANOVA was defined
as the MCII. Bootstrap resampling with 2000 samples was
used to test the robustness of the choice of optimal MCII
threshold.
We defined postoperative recovery for an index symp-
tom as a decrease in two consecutive postoperative scores
at or over the MCII compared with the score at POD 1
during post-surgery hospitalization, as reported by patients.
To identify potential risk factors for lack of pain and fati-
gue recovery, we first selected candidate variables (age,
body mass index, sex, pathological stage, smoking history,
surgical approach, pathological type, ASA physical status
classification, Charlson Comorbidity Index [CCI] score,
preoperative neoadjuvant therapy, percentage of predicted
forced expiratory volume in 1 s, postoperative pathological
type, complication grade, readmission) based on their
potential clinical relevance; thereafter, we used Cox
regression models stepwise variable selection with a
pvalue \0.05 as a criterion for inclusion (to keep the
model parsimonious). Kaplan–Meier analysis was then
used to evaluate the effect of the risk factors that were
significantly associated with recovery status on the median
recovery days for each index symptom.
In addition, using generalized mixed-effects models, we
compared post-discharge trajectories of composite physical
score (average mean of MDASI-LC general activity, work,
MCII to Define Symptom Recovery After VATS
TABLE 1 Patient demographics and clinical characteristics in the development and validation cohorts
Development cohort Validation cohort
POD 1
[n=402]
Follow-up available
a
[n= 366]
Missing data at discharge
b
[n=36]
pValue
(follow-up vs. missing)
POD 1
[n=189]
Follow-up available
a
[n=155]
Missing data at discharge
[n=34]
pValue
(follow-up vs. missing)
Age, years [mean (SD)] 54.3 (10.4) 54.11 (10.4) 56.17 (9.7) 0.26 57.1 (11.5) 57.0 (11.4) 57.6 (11.9) 0.76
BMI, mean (SD) 23.13 (2.9) 23.2 (2.9) 22.46 (3.2) 0.15 – – –
Female, no. (%) 211 (52.5) 193 (52.7) 18 (50.0) 0.75 89 (47.1) 70 (45.2) 19 (55.8) 0.25
Length of postoperative hospital stay, days
[median (IQR)]
6 (5–7) 6 (5–7) 6 (5–7) 0.97 4 (3–5) 4 (3–5) 4 (3–4) 0.89
Pathological stage [n (%)]
0–I 236 (72.2) 215 (72.9) 21 (65.6) 0.7
c
141 (87.6) 117 (88.6) 24 (82.8) 0.37
c
II 35 (10.7) 31 (10.5) 4 (12.5) 17 (10.5) 12 (9.1) 5 (17.2)
III 32 (9.8) 28 (9.5) 4 (12.5) 3 (1.9) 3 (2.3) 0 (0)
IV 24 (7.3) 21 (7.1) 3 (9.4) – – –
Smoking history [n (%)]
No 277 (68.9) 255 (69.7) 22 (61.1) 0.29 160 (84.7) 130 (83.9) 30 (88.2) 0.52
Smoker 125 (31.1) 111 (30.3) 14 (38.9) 29 (15.3) 25 (16.1) 4 (11.8)
Surgical approach [n (%)]
Single-port 249 (61.9) 225 (61.5) 24 (66.7) 0.54 104 (55.3) 87 (56.5) 17 (50.0) 0.49
Multi-port 153 (38.1) 141 (38.5) 12 (33.3) 84 (44.7) 67 (43.4) 17 (50.0)
Pathological type [n (%)]
Adenocarcinoma 292 (89.3) 265 (89.8) 27 (84.4) 0.36
c
–– –
Other 35 (10.7) 30 (10.2) 5 (15.6) – – –
ASA classification [n (%)]
1 186 (46.3) 177 (48.4) 9 (25.0) \0.001 –– –
[1 216 (53.7) 189 (51.6) 27 (75.0) – – –
CCI score [n (%)]
0 126 (31.3) 114 (31.2) 12 (33.3) 0.79 – – –
C1 276 (68.7) 252 (68.8) 24 (66.7) – – –
Preoperative neoadjuvant therapy [n (%)]
No 395 (98.3) 360 (98.4) 35 (97.2) 0.48
c
180 (95.2) 147 (94.8) 33 (97.1) 0.99
c
Yes 7 (1.7) 6 (1.6) 1 (2.8) 9 (4.8) 8 (5.2) 1 (2.9)
Pulmonary function [n (%)]
FEV1% \80% 78 (19.4) 71 (19.4) 7 (19.4) 0.99 – – –
FEV1% C80% 324 (80.6) 295 (80.6) 29 (80.6) – – –
Postoperative pathological type [n (%)]
Benign 67 (16.7) 63 (17.21) 4 (11.1) 0.35 22 (11.6) 19 (12.3) 3 (8.8) 0.77
c
Malignant 335 (83.3) 303 (82.79) 32 (88.9) 167 (88.4) 136 (87.7) 31 (91.2)
Complication, grade [2 [n (%)]
In hospital 57 (14.2) 50 (13.7) 7 (19.4) 0.34 – – – –
After discharge 11 (2.7) 10 (2.7) 1 (2.8) 0.99
c
–– – –
Readmission [n(%)] 11 (2.7) 10 (2.7) 1 (2.8) 0.99
c
–– – –
W. Xu et al.
and walking scores) and composite affective interference
score (average mean of MDASI-LC mood, relations with
others, and enjoyment-of-life scores) between patients who
were recovered and those who were unrecovered on the
day of discharge. The mixed-effects models included the
recovery status, time (weeks after discharge), and ASA
physical status classification as fixed effects. The subject
and time random effects were included. Maximum likeli-
hood estimation was used. For repeated measures (time), a
variance components structure with heterogeneous vari-
ances was assumed.
Finally, we examined whether the obtained MCII-de-
fined recovery was related to functional status, using both
development and validation cohorts. We used a two-inde-
pendent sample t-test to compare functional scores between
recovered and unrecovered patients. All pvalues were
obtained from two-sided tests and the results were con-
sidered statistically significant at p\0.05. All statistical
analyses were conducted using JMP Clinical software
version 6.1 (SAS Institute Inc., Cary, NC, USA).
RESULTS
Sample Characteristics
A total of 366 patients who underwent VATS were
included in the development cohort and 155 were included
in the validation cohort. Selected characteristics of patients
in both cohorts are demonstrated in Table 1. Compared
with patients in the development cohort, those in the val-
idation cohort had shorter LOS.
Definition of Index Patient-Reported Outcomes
and Minimal Clinically Important Improvements
(MCIIs)
Over the period of hospitalization after surgery, the five
most severe postoperative symptoms were pain
(3.87 ±2.53), fatigue (3.33 ±2.53), coughing
(3.29 ±2.33), shortness of breath (2.95 ±2.68), and dis-
turbed sleep (2.94 ±2.68). All top five symptoms
improved substantially during the hospital stay (Table 2).
We identified pain and fatigue as the index symptoms
because their standardized response means were over the
0.50 threshold (1.08 absolute and 1.01 relative for pain,
0.73 absolute and 0.84 relative for fatigue) [Table 2]. The
correlation coefficients between the six interference items
(anchors) and pain and fatigue during hospitalization met
the 0.30 threshold (p\0.001 for both index symptoms).
We used nine sets of thresholds to fit the MANOVA
model to each pair of change in pain or fatigue scores as
absolute or relative. Table 3demonstrates the MANOVA
Table 1 (continued)
Development cohort Validation cohort
POD 1
[n=402]
Follow-up available
a
[n= 366]
Missing data at discharge
b
[n=36]
pValue
(follow-up vs. missing)
POD 1
[n=189]
Follow-up available
a
[n=155]
Missing data at discharge
[n=34]
pValue
(follow-up vs. missing)
Preoperative score [mean (SD)]
Pain 0.58 (1.4) 0.60 (1.4) 0.46 (1.2) 0.40 0.6 (1.4) 0.6 (1.5) 0.5 (0.8) 0.55
Fatigue 0.95 (1.6) 0.96 (1.7) 0.86 (1.4) 0.51 1.1 (1.6) 1.1 (1.7) 1.1 (1.4) 0.95
General activity 0.36 (1.1) 0.37 (1.2) 0.23 (0.7) 0.78 0.8 (1.6) 0.8 (1.6) 0.7 (1.2) 0.67
Mood 1.1 (2.0) 1.14 (2.0) 0.63 (1.3) 0.36 1.7 (2.2) 1.8 (2.3) 1.2 (1.5) 0.11
Work 0.7 (1.6) 0.73 (1.7) 0.40 (1.0) 0.34 1.1 (2.0) 1.2 (2.0) 0.8 (1.2) 0.13
Relations with others 0.52 (1.5) 0.54 (1.5) 0.26 (0.9) 0.18 0.9 (1.8) 0.9 (1.8) 0.8 (1.7) 0.80
Walking 0.37 (1.2) 0.37 (1.3) 0.31 (0.8) 0.89 0.8 (1.8) 0.7 (1.7) 1.2 (2.2) 0.26
Enjoyment of life 0.7 (1.6) 0.72 (1.7) 0.49 (1.3) 0.81 1.1 (2.0) 1.1 (2.0) 0.9 (1.6) 0.63
BMI body mass index (calculated as weight in kilograms divided by height in meters squared), ASA American Society of Anesthesiologists, CCI Charlson Comorbidity Index, SD standard deviation, FEV1% percentage of predicted
forced expiratory volume in 1 s, IQR interquartile range, POD postoperative day
a
Date available on POD 1 and discharge day
b
We excluded patients whose symptom severity or interference scores were missing on the day of discharge
c
Two-tailed Fisher exact test was used for categorical variables
MCII to Define Symptom Recovery After VATS
Fvalues for MCII determination. The 30% relative
decrease and 2-point absolute decrease for both pain and
fatigue generated the largest Fvalues, therefore these
thresholds were chosen as the MCIIs. These MCIIs then
underwent internal validation using the bootstrap resam-
pling results, where the least differences between the top
three highest Fvalues were observed for the MANOVA
model (electronic supplementary Table S1). The bootstrap
resampling confirmed that 30% and 2-point MCIIs were the
optimal parameters in the largest proportion of samples for
both pain and fatigue.
Identification of Poor Recovery Status
We used the 30% and 2-point MCIIs to identify patients
who did not recover by the measured time points (elec-
tronic supplementary Table S2). None of the patients
reached an MCII-defined symptom recovery on POD 2.
For pain defined by the 30% MCII, 22.6% of patients at
discharge and 10.2% at 4 weeks after discharge did not
recover. For pain defined by the 2-point MCII, 19.3% of
patients on the day of discharge and 8.4% at 4 weeks after
discharge did not recover. For fatigue defined by the 30%
MCII, 22.4% of patients at discharge and 13.6% at 4 weeks
after discharge were unrecovered. For pain defined by the
2-point MCII, 18.7% of patients at discharge and 10.5% at
4 weeks after discharge were unrecovered (electronic
supplementary Table S2).
Patients who did not recover from pain and fatigue at
discharge, as defined by the 30% MCII, reported signifi-
cantly poorer physical and affective functional status
during the first 4 weeks after discharge via a generalized
mixed-effects models (all p\0.05), as shown in Fig. 1.
Cox regression models found that risk factors for unre-
covered pain included higher CCI score (C1 vs. 0; hazard
ratio [HR] 1.36, 95% confidence interval [CI] 1.04–1.77;
p=0.02) and lack of preoperative neoadjuvant therapy
(HR 2.78, 95% CI 1.13–6.83; p=0.02). Malignancy was
related to slower fatigue recovery as defined by the 30%
MCII (malignant vs. benign; HR 1.47, 95% CI 1.02–2.13;
p=0.04) (Table 4). Lack of neoadjuvant therapy was
again related to slower pain recovery as defined by the 30%
MCII (2 days vs. 4 days; p=0.02) (Fig. 2, electronic
supplementary Table S3). In Table 5, compared with
patients who had an MCII, the physical interference score
at discharge was significantly higher in patients with
unrecovered pain (mean difference in score 1.15, 95% CI
0.56–1.74; p\0.001) and in patients with unrecovered
fatigue (mean difference in score 1.27, 95% CI 0.71–1.84;
p\0.001).
External Validation of MCIIs
In the external validation cohort, 55 (28.6%) patients did
not recover from pain and 58 (28.7%) did not recover from
fatigue at discharge as defined by the 30% MCII. As shown
TABLE 2 Defining index symptoms using changes in symptoms in the development sample (n=366)
Symptom item on
MDASI-LC
(0–10 scale)
POD 1
a
(mean ±SD)
Discharge
b
(mean ±SD)
Mean score change
(absolute)
c
[range]
Mean score change
(relative)
d
[range]
Absolute
SRM
Relative
SRM
pValue
Pain 5.36 ±2.63 2.46 ±2.01 -2.90 ±2.69 [-10 to
10]
-0.50 ±0.45 [-1to
2]
1.08 1.11 \0.001
Fatigue 4.67 ±2.89 2.22 ±2.16 -2.45 ±2.91 [-10 to
8]
-0.45 ±0.62 [-1to
3]
0.84 0.73 \0.001
Coughing 3.89 ±2.34 3.22 ±2.07 -0.68 ±2.47 [-8 to 5] 0.11 ±1.03 [-1 to 5] 0.28 0.11 \0.001
Shortness of breath 4.36 ±2.45 2.48 ±2.09 -1.88 ±2.53 [-9to7] -0.29 ±0.81 [-1to
7]
0.74 0.36 \0.001
Disturbed sleep 4.44 ±2.54 2.44 ±2.49 -2.00 ±2.91 [-10 to
7]
-0.29 ±0.95 [-1to
7]
0.69 0.31 \0.001
Values in bold indicate that symptom scores between POD 1 and discharge showed significant difference (p\0.05)
MDASI-LC MD Anderson Symptom Inventory–Lung Cancer Module, POD postoperative day, SD standard deviation, SRM standardized
response mean
a
We excluded patients who reported no symptoms (rated 0) on POD 1
b
Discharge date: a median of 6 (5–7) days after surgery from the development sample
c
Absolute change ¼symptom scores at discharge symptom scores on POD1
d
Relative change %ðÞ¼
ðsymptom scores at discharge symptom scores on POD1Þ
ðsymptom scores on POD1Þ100%
W. Xu et al.
TABLE 3 MANOVA using anchor-based methods to define the MCII of index symptoms from the development sample
a
MCII threshold
b
FpValue
Pain
Relative decrease from POD 1 to discharge
c
10% 1.93 0.08
20% 2.90 0.01
30% 2.90 0.01
40% 2.21 0.04
50% 2.30 0.04
60% 1.45 0.20
70% 1.10 0.36
80% 2.17 0.05
90% 1.16 0.33
Absolute decrease from POD 1 to discharge
d
1 1.86 0.09
2 3.86 0.001
3 2.03 0.06
4 3.45 0.003
5 2.32 0.03
6 2.60 0.02
7 1.56 0.16
8 0.99 0.43
9 2.15 0.05
Fatigue
Relative decrease from POD 1 to discharge
c
10% 2.29 0.04
20% 3.95 \0.001
30% 5.41 \0.001
40% 5.12 \0.001
50% 3.46 0.003
60% 5.15 \0.001
70% 3.08 0.006
80% 2.17 0.05
90% 2.85 0.01
Absolute decrease from POD 1 to discharge
d
1 2.29 0.04
2 4.32 \0.001
3 3.89 0.001
4 2.23 0.04
5 1.99 0.07
6 2.45 0.03
7 2.09 0.06
8 1.71 0.12
9 1.05 0.39
Values in bold indicate that the threshold showed largest Fvalues in nine set
MANOVA multivariate analysis of variance, MCII minimal clinically important improvement, POD postoperative day
a
Patients who scored their index symptoms as 4–10 on POD 1 were included in the MCII determination
b
Thresholds group symptom score decreases into two categories: below the threshold and at or above the threshold
c
Absolutechange ¼symptom scores at discharge symptom scores on POD1
d
Relative change %ðÞ¼
ðsymptom scores at discharge symptom scores on POD1Þ
ðsymptom scores on POD1Þ100%
MCII to Define Symptom Recovery After VATS
in Table 5, compared with patients recovering from pain,
those who did not report pain recovery at discharge
reported poorer physical functional status (mean difference
in score 0.91, 95% CI 0.09–1.72; p=0.03). Similarly,
compared with those who recovered from fatigue, those
who did not recover from fatigue experienced poorer
physical functional status after discharge (mean difference
in score 1.34, 95% CI 0.54–2.14; p=0.001) [Table 5].
DISCUSSION
MCIIs are recognized as context-specific,
26
and to the
best of our knowledge, this is the first study to describe
postoperative recovery in patients who underwent VATS
using MCIIs established using patient-reported symptoms
induced by the surgery, rather than symptoms whose relief
was targeted by the surgery. We identified pain and fatigue
Composite Physical Interference Score after discharge
Composite Physical Interference Score after discharge
(defined by 30% MCII of pain)
(defined by 30% MCII of fatigue)
Unrecovered
Recovered
Unrecovered
Recovered
Interference Mean ScoreInterference Mean Score
Discharge
Weeks since Discharge
6
5
4
3
2
1
0
1234
P=0.006
Discharge
Weeks since Discharge
6
5
4
3
2
1
0
1234
P=0.006
Composite Affective Interference Score after discharge
(defined by 30% MCII of pain)
Unrecovered
Recovered
Interference Mean Score
Discharge
Weeks since Discharge
6
5
4
3
2
1
0
1234
P=0.01
(a) (b)
(c) (d) Composite Affective Interference Score after discharge
(defined by 30% MCII of fatigue)
Unrecovered
Recovered
Interference Mean Score
Discharge
Weeks since Discharge
6
5
4
3
2
1
0
1234
P<0.001
FIG. 1 Trajectory of recovery of daily functioning (composite physical and affective interference scores) from the development cohort at
discharge, defined by 30% MCII for pain and fatigue. MCII minimal clinically important improvement.
TABLE 4 Multivariate Cox regression analysis of risk factors in unrecovered patients defined by 30% MCII on index symptoms from the
development sample
Estimate (SE) HR 95% CI pValue
30% MCII for pain [n =332]
CCI score (C1 vs. 0) 0.31 (0.14) 1.36 1.04–1.77 0.02
Preoperative neoadjuvant therapy (no vs. yes) 1.02 (0.46) 2.78 1.13–6.83 0.02
30% MCII for fatigue [n =228]
Postoperative pathological type (malignant vs. benign) 0.39 (0.19) 1.47 1.02–2.13 0.04
Values in bold indicate that patients showed significant difference in this factor (p\0.05)
CI confidence interval, CCI Charlson Comorbidity Index, HR hazard ratio, SE standard error, MCII minimal clinically important improvement
W. Xu et al.
as index symptoms and identified 30% and 2-point
decreases in symptom scores as the relative and absolute
MCIIs, respectively, that define the postoperative recovery
from the patients’ perspective. Furthermore, a 30%
threshold for MCII could identify patients who may need
extensive care after discharge. As a sensitive, valid, and
clinically interpretable PRO instrument, MCIIs can be used
to effectively trace postoperative recovery, to measure
patients’ responses to the insult of surgical intervention,
and to support the identification of those who might need
extensive care after discharge.
26
In this study, we established MCIIs for lung cancer
patients who underwent VATS. The National Compre-
hensive Cancer Network guidelines recommend that VATS
or minimally invasive surgery (including robotic-assisted
approaches) should be strongly considered for patients with
no anatomic or surgical contraindications.
27
In previous
studies, VATS was associated with lower postoperative
pain and fatigue, especially on POD 1, compared with
thoracotomy.
2,6,28
MCII is considered a ‘variable con-
cept’
29
because its start point (e.g., POD 1 after surgery)
and setting (e.g., VATS) affect the consistency of esti-
mates.
30
Therefore, there is an urgent need for tailored
MCIIs to precisely assess the recovery status after VATS.
The identified index symptoms emphasized the impor-
tance of routine monitoring and management of pain and
fatigue. Thoracic surgery is considered one of the most
painful surgeries
20,31
and patients have reported that
reduced physical functioning after surgery lasts up to 2
years.
32
A number of studies use pain or fatigue as major
targets for symptom monitoring after thoracic sur-
gery,
18,24,33,34
even with minimally invasive
techniques.
12,35,36
These findings suggested that pain and
fatigue were associated with a large overall symptom
burden after surgery. Our study provided real-world
evidence to support pain and fatigue as outcomes important
to patients
37
and sensitive to the transition of the recovery
status.
38
Rather than establishing the MCIIs for symptom-relief
surgery, e.g., 30% for neck and arm pain after cervical
spine surgery
39
or 3 points for leg pain and 33% for back
pain after lumbar spinal surgery,
40
our study focused on the
recovery of symptoms induced by the procedure. We used
MCIIs to address trends in improvement rather than
worsening in condition, which is usually addressed by
minimal clinically important differences (MCIDs).
11
Few
MCIDs for fatigue have been reported on an 11-point
numerical rating scale, while 0.8–1.1 points were reported
for rheumatoid arthritis.
41
Compared with other fatigue
MCIIs, our 30% and 2-point MCIIs may define recovery
from surgery more conservatively (with larger differences),
which would impact discharge decisions and symptom
management after discharge, thus benefiting patients more,
especially in the era of continuously shortened LOS after
surgery.
Using MCIIs to interpret changes in PRO scores, we
demonstrated that one in five patients may not recover by
discharge, with a median postoperative LOS of 6 days.
Although studies on ERAS usually reported a shortened
postoperative LOS (e.g., from 6 to 4 days),
42
our data
showed that nearly half of the patients did not recover by
POD 4. Together with the patient-reported symptom bur-
den, MCIIs can provide clinical evidence for patient
characteristics that may be detrimental to a fast recovery,
such as comorbidity and pathological type, in the practice
of VATS.
The 30% MCII identified patients at the time of dis-
charge who might need extensive care after discharge.
43
Even with a median 6-day LOS after surgery, those who
did not recover from postoperative pain or fatigue
Defined by 30% MCII of Pain Defined by 30% MCII of Pain Defined by 30% MCII of Fatigue
Probability of Recovery
Log-rank P=0.009 Log-rank P=0.005 Log-rank P=0.02
Postoperative days to recovery Postoperative days to recovery
No. at risk No. at risk No. at risk
CCI score=0
CCI score≥1
CCI score=0
CCI score≥1
(a) (b) (c)
100
80
60
40
20
0
Probability of Recovery
100
80
60
40
20
0
Probability of Recovery
100
80
60
40
20
0
07
14 21 28 35 42 49 0 7 14 21 28 35 42 49
Postoperative days to recovery
07
14 21 28 35 42 49
88 9 7 620 0 0
186 40 23 18 14 4 1 0
500 0 0000
269 49 30 24 15 4 10
with Neoadjuvant
without Neoadjuvant
with Preoperative Neoadjuvant Therapy
without Preoperative Neoadjuvant Therapy
Benign
Malignant
Benign
Malignant
35 5 3 20000
193 36 26 24 13 1 0 0
FIG. 2 Pain and fatigue recovery defined using MCIIs in the development cohort. For pain recovery, by CCI score and neoadjuvant therapy;
fatigue recovery, by pathological type; pvalue by log-rank test. CCI Charlson Comorbidity Index, MCII minimal clinically important
improvement
MCII to Define Symptom Recovery After VATS
TABLE 5 Comparison of functional status between recovered and unrecovered patients at discharge per the 30% MCII in the development and validation datasets
Development cohort [n= 366]
30% MCII of pain
a
30% MCII of fatigue
a
Daily functioning Unrecovered
b
[n=80]
Recovered
c
[n=275] Mean difference
(95% CI)
pValue Unrecovered
[n=94]
Recovered
[n=232]
Mean difference
(95% CI)
pValue
Composite physical interference
score
d
3.69 ±2.94 2.54 ±2.17 1.15 (0.56–1.74) \0.001 3.80 ±2.59 2.52 ±2.26 1.27 (0.71–1.84) \0.001
Composite affective interference
score
e
2.60 ±2.44 1.71 ±1.85 0.90 (0.40–1.40) \0.001 2.95 ±2.31 1.61 ±1.84 1.34 (0.86–1.82) \0.001
Validation cohort [n= 155]
30% MCII of pain 30% MCII of fatigue
Daily functioning Unrecovered [n=40] Recovered [n=100] Mean difference
(95% CI)
pValue Unrecovered
[n=39]
Recovered
[n=97]
Mean difference
(95% CI)
pValue
Composite physical interference score
d
3.14±2.33 2.24±2.15 0.91 (0.09–1.72) 0.03 3.55 ±2.51 2.21 ±1.97 1.34 (0.54–2.14) 0.001
Composite affective interference score
e
2.50±2.22 1.78±1.89 0.72 (-0.01 to 1.46) 0.05 2.49 ±1.87 1.91 ±2.08 0.57 (-0.19 to 1.33) 0.14
Values in bold indicate that patients between recovered and unrecovered showed significant difference (p\0.05)
CI confidence interval, MCII minimal clinically important improvement, MDASI-LC MD Anderson Symptom Inventory–Lung Cancer Module
a
We excluded patients with a score of 0 on POD 1 for relative MCII analysis
b
Unrecovered: a relative change of \30% in the index symptom on discharge
c
Recovered: a relative change of C30% in the index symptom on discharge
d
Composite physical interference score is the mean of walking, general activity, and work scores in MDASI-LC
e
Composite affective interference score is the mean of scores for mood, enjoyment of life, and relations with others in MDASI-LC
W. Xu et al.
experienced significantly worse daily functioning for 1
month after discharge. From the patient’s perspective,
returning to the preoperative functional status in daily life
was the most important target for recovery from surgery.
32
Additionally, although no threshold was established for
discharge in terms of symptom recovery, the substantial
proportion of unrecovered patients on the day of discharge
suggests that healthcare professionals should collaborate
more with patients in making the discharge decision and in
after-discharge care.
This study has some limitations. First, using symptom
scores on the 14th day as the discharge score for patients
discharged [14 days after surgery may overestimate the
average symptom burden at discharge. Although only 5%
of all patients were hospitalized for longer than 14 days
after surgery, the overestimation may benefit those who
need more comprehensive care after discharge since the
LOS has continued to decrease with the implementation of
ERAS. Second, only five patients were treated with pre-
operative neoadjuvant therapy, which was beneficial to
recovery; further validation is needed to interpret this result
and to provide a solid reference in surgical practice.
CONCLUSION
Pain and fatigue were identified as index PRO measures
for recovering from VATS. The 30% decreases in index
symptom severity indicated an MCII as the threshold rep-
resenting a meaningful recovery after VATS and could
identify patients who may need extensive care after dis-
charge. Combined with clinical outcomes, well-defined
PRO information could be valuable to patients and refer-
ring physicians to design the best suitable follow-up plan
for the post-discharge period.
Supplementary Information The online version contains
supplementary material available at https://doi.org/10.1245/s10434-
022-11629-7.
ACKNOWLEDGMENT The authors would like to thank Xing
Wei, Shaohua Xie, Yunfei Mu, Wenhong Feng, Yuanqiang Zhang,
Rui Zhang, and Xiaoqing Liao for their contributions to the data
collection in this study. This manuscript was edited by Sarah Bron-
son, ELS, of the Research Medical Library at The University of Texas
MD Anderson Cancer Center.
REFERENCES
1. Villamizar NR, Darrabie M, Hanna J, et al. Impact of T status and
N status on perioperative outcomes after thoracoscopic lobec-
tomy for lung cancer. J Thorac Cardiovasc Surg.
2013;145(2):514–20.
2. Bendixen M, Jørgensen OD, Kronborg C, Andersen C, Licht PB.
Postoperative pain and quality of life after lobectomy via video-
assisted thoracoscopic surgery or anterolateral thoracotomy for
early stage lung cancer: a randomised controlled trial. Lancet
Oncol. 2016;17(6):836–44.
3. Khullar OV, Rajaei MH, Force SD, et al. Pilot study to integrate
patient reported outcomes after lung cancer operations into the
society of thoracic surgeons database. Ann Thorac Surg.
2017;104(1):245–53.
4. Yang CJ, Kumar A, Deng JZ, et al. A national analysis of short-
term outcomes and long-term survival following thoracoscopic
versus open lobectomy for clinical stage II non-small-cell lung
cancer. Ann Surg. 2021;273(3):595–605.
5. Dziedzic R, Marjanski T, Binczyk F, Polanska J, Sawicka W,
Rzyman W. Favourable outcomes in patients with early-stage
non-small-cell lung cancer operated on by video-assisted thora-
coscopic surgery: a propensity score-matched analysis. Eur J
Cardiothorac Surg. 2018;54(3):547–53.
6. Wei X, Yu H, Dai W, et al. Patient-reported outcomes of video-
assisted thoracoscopic surgery versus thoracotomy for locally
advanced lung cancer: a longitudinal cohort study. Ann Surg
Oncol. 2021;28(13):8358–71.
7. Lee L, Tran T, Mayo NE, Carli F, Feldman LS. What does it
really mean to ‘‘recover’’ from an operation? Surgery.
2014;155(2):211–6.
8. Neville A, Lee L, Antonescu I, et al. Systematic review of out-
comes used to evaluate enhanced recovery after surgery. Br J
Surg. 2014;101(3):159–70.
9. Fiore JF Jr, Figueiredo S, Balvardi S, et al. How do we value
postoperative recovery?: a systematic review of the measurement
properties of patient-reported outcomes after abdominal surgery.
Ann Surg. 2018;267(4):656–69.
10. US Department of Health and Human Services FDA Center for
Drug Evaluation and Research; US Department of Health and
Human Services FDA Center for Biologics Evaluation and
Research; US Department of Health and Human Services FDA
Center for Devices and Radiological Health. Guidance for
industry: patient-reported outcome measures: use in medical
product development to support labeling claims: draft guidance.
Health Qual Life Outcomes. 2006;4:79.
11. Shi Q, Wang XS, Vaporciyan AA, Rice DC, Popat KU, Cleeland
CS. Patient-reported symptom interference as a measure of
postsurgery functional recovery in lung cancer. J Pain Symptom
Manag. 2016;52(6):822–31.
12. Fagundes CP, Shi Q, Vaporciyan AA, et al. Symptom recovery
after thoracic surgery: Measuring patient-reported outcomes with
the MD Anderson Symptom Inventory. J Thorac Cardiovasc
Surg. 2015;150(3):613-9.e2.
13. Reeve BB, Wyrwich KW, Wu AW, et al. ISOQOL recommends
minimum standards for patient-reported outcome measures used
in patient-centered outcomes and comparative effectiveness
research. Qual Life Res. 2013;22(8):1889–905.
14. McLeod LD, Coon CD, Martin SA, Fehnel SE, Hays RD.
Interpreting patient-reported outcome results: US FDA guidance
and emerging methods. Expert Rev Pharmacoecon Outcomes
Res. 2011;11(2):163–9.
15. Kvien TK, Heiberg T, Hagen KB. Minimal clinically important
improvement/difference (MCII/MCID) and patient accept-
able symptom state (PASS): what do these concepts mean? Ann
Rheum Dis. 2007;66(Suppl 3):iii40–1.
16. Tubach F, Giraudeau B, Ravaud P. The variability in minimal
clinically important difference and patient acceptable symp-
tomatic state values did not have an impact on treatment effect
estimates. J Clin Epidemiol. 2009;62(7):725–8.
17. Rosen RC, Allen KR, Ni X, Araujo AB. Minimal clinically
important differences in the erectile function domain of the
International Index of Erectile Function scale. Eur Urol.
2011;60(5):1010–6.
18. Rebelo P, Oliveira A, Andrade L, Valente C, Marques A. Mini-
mal clinically important differences for patient-reported outcome
MCII to Define Symptom Recovery After VATS
measures of fatigue in patients with COPD following pulmonary
rehabilitation. Chest. 2020;158(2):550–61.
19. Clinical Outcome Assessment (COA): Frequently Asked Ques-
tions. Available at: https://www.fda.gov/about-fda/clinical-outc
ome-assessment-coa-frequently-asked-questions. Accessed from
2 Dec 2020.
20. Dai W, Xie S, Zhang R, et al. Developing and validating utility
parameters to establish patient-reported outcome-based periop-
erative symptom management in patients with lung cancer: a
multicentre, prospective, obs. BMJ Open. 2019;9(10):e030726.
21. Wang XS, Wang Y, Guo H, Mendoza TR, Hao XS, Cleeland CS.
Chinese version of the M. D. Anderson Symptom Inventory:
validation and application of symptom measurement in cancer
patients. Cancer. 2004;101(8):1890–901.
22. Mendoza TR, Wang XS, Lu C, et al. Measuring the symptom
burden of lung cancer: the validity and utility of the lung cancer
module of the M. D. Anderson Symptom Inventory. Oncologist.
2011;16(2):217–27.
23. Ward MM, Guthrie LC, Alba MI. Clinically important changes in
individual and composite measures of rheumatoid arthritis
activity: thresholds applicable in clinical trials. Ann Rheum Dis.
2015;74(9):1691–6.
24. Cleeland CS, Wang XS, Shi Q, et al. Automated symptom alerts
reduce postoperative symptom severity after cancer surgery: a
randomized controlled clinical trial. J Clin Oncol.
2011;29(8):994–1000.
25. Reni M, Braverman J, Hendifar A, et al. Evaluation of minimal
important difference and responder definition in the EORTC
QLQ-PAN26 module for assessing health-related quality of life
in patients with surgically resected pancreatic adenocarcinoma.
Ann Surg Oncol. 2021;28(12):7545–54.
26. Wright A, Hannon J, Hegedus EJ, Kavchak AE. Clinimetrics
corner: a closer look at the minimal clinically important differ-
ence (MCID). J Man Manip Ther. 2012;20(3):160–6.
27. NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell
Lung Cancer. Version 6. Available at: https://www.nccn.org/pro
fessionals/physician_gls/f_guidelines.asp. Accessed 30 June
2020.
28. Abrams BA, Murray KA, Mahoney K, et al. Postdischarge pain
management after thoracic surgery: a patient-centered approach.
Ann Thorac Surg. 2020;110(5):1714–21.
29. de Vet HCW, Terwee CB, Mokkink LB, Knol DL. Measurement
in medicine: a practical guide. Cambridge: Cambridge University
Press; 2011.
30. Alma H, de Jong C, Jelusic D, et al. Baseline health status and
setting impacted minimal clinically important differences in
COPD: an exploratory study. J Clin Epidemiol. 2019;116:49–61.
31. Thompson C, French DG, Costache I. Pain management within
an enhanced recovery program after thoracic surgery. J Thorac
Dis. 2018;10(Suppl 32):S3773-s3780.
32. Poghosyan H, Sheldon LK, Leveille SG, Cooley ME. Health-
related quality of life after surgical treatment in patients with non-
small cell lung cancer: a systematic review. Lung Cancer.
2013;81(1):11–26.
33. Hugoy T, Lerdal A, Rustoen T, Oksholm T. Predicting postop-
erative fatigue in surgically treated lung cancer patients in
Norway: a longitudinal 5-month follow-up study. BMJ Open.
2019;9(9):e028192.
34. Van Herck M, Antons J, Vercoulen JH, et al. Pulmonary reha-
bilitation reduces subjective fatigue in COPD: a responder
analysis. J Clin Med. 2019;8(8):1264.
35. Yao Y, Jiang J, Lin W, Yu Y, Guo Y, Zheng X. Efficacy of
systemic lidocaine on postoperative quality of recovery and
analgesia after video-assisted thoracic surgery: a randomized
controlled trial. J Clin Anesth. 2017;71:110223.
36. Yao Y, Fu S, Dai S, et al. Impact of ultrasound-guided erector
spinae plane block on postoperative quality of recovery in video-
assisted thoracic surgery: a prospective, randomized, controlled
trial. J Clin Anesth. 2020;63:109783.
37. Principles for Selecting, Developing, Modifying, and Adapting
Patient-Reported Outcome Instruments for Use in Medical
Device Evaluation. Available at: https://www.fda.gov/regulatory-
information/search-fda-guidance-documents/principles-selecting-
developing-modifying-and-adapting-patient-reported-outcome-in
struments-use. Accessed from 28 Aug 2020.
38. Patient-Reported Outcome Measures: Use in Medical Product
Development to Support Labeling Claims. Available at: https://
www.fda.gov/regulatory-information/search-fda-guidance-docu
ments/patient-reported-outcome-measures-use-medical-product-d
evelopment-support-labeling-claims. Accessed from 17 Oct
2009.
39. Khan I, Pennings JS, Devin CJ, et al. Clinically meaningful
improvement following cervical spine surgery: 30% reduction
versus absolute point-change MCID values. Spine (Phila Pa
1976). 2021;46(11):717–25.
40. Austevoll IM, Gjestad R, Grotle M, et al. Follow-up score,
change score or percentage change score for determining clinical
important outcome following surgery? An observational study
from the Norwegian registry for Spine surgery evaluating patient
reported outcome measures in lumbar spinal stenosis and lumbar
degenerative spondylolisthesis. BMC Musculoskelet Disord.
2019;20(1):31.
41. Khanna D, Pope JE, Khanna PP, et al. The minimally important
difference for the fatigue visual analog scale in patients with
rheumatoid arthritis followed in an academic clinical practice. J
Rheumatol. 2008;35(12):2339–43.
42. Gonzalez M, Abdelnour-Berchtold E, Perentes JY, et al. An
enhanced recovery after surgery program for video-assisted tho-
racoscopic surgery anatomical lung resections is cost-effective. J
Thorac Dis. 2018;10(10):5879–88.
43. Sung MR, Patel MV, Djalalov S, et al. Evolution of symptom
burden of advanced lung cancer over a decade. Clin Lung Can-
cer. 2017;18(3):274-280.e6.
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
W. Xu et al.
