Development of the scale
Forming alternative items and initial scales. Through literature researches, case reviews, collecting opinions of patients, and soliciting personal experience of experts, the research group collected alternative items. Then the core group formed the initial scale which consisted of three subscales with a total of 25 items. There were 14 items in signs and symptoms subscale (scale A), 7 items in psychological and psychiatric subscale (scale B) and 4 items in social–life subscale (scale C) (annex Table 1). The score of each item ranges from 0 to 3, indicating that the severity of corresponding symptoms or conditions are from none to extremely serious. The total score refers to the sum of the scores for all items, and the subscale score is the sum of the scores for each subscale item.
Preliminary investigation. Three NSCLC patients and 10 healthy people were investigated with the primary NSCLC–PQOL scale during the pilot survey. Results showed that no item was ambiguous, added or deleted.
Developing the final scale. (1) Demographic characteristics of the study subjects: The subjects of this study are 347 patients who received operation and treatment in the Department of Lung Cancer Surgery at Tianjin Medical University Cancer Institute and Hospital from November 2019 to May 2021, and then we used primary NSCLC–PQOL scale to evaluate the severity of the patients’ signs and symptoms, social adaptation and mental well-being six months after their surgery. The demographic characteristics are shown in Table 1. (2) Screening scale items by correlation coefficient method: The correlation coefficient method is to calculate the correlation coefficient between each item and the secondary variable (the subscale) to which the item belongs, and then delete the items with the absolute value of the correlation coefficient less than 0.4. The data collected in this study did not obey the normal distribution, so the correlation coefficient was calculated by Spearman test. The results are shown in Table 2. The items which could be eliminated are “weight loss”, “hair loss”, “poor appetite”, “pain of the surgical wound”, “chest pain”, “other pain”, “vomiting”, “insomnia” and “difficulty in remembering”. (3) Analysis of item distribution inspection method: The method inspects the distinguishability of scale items from the central tendency of item answers’ distribution. Generally, the rate of item choices over 80% can be deleted, which indicates that the poor specificity of the items. The results are shown in Table 3. It proposed to delete the items: “hair loss”, “vomiting”, “difficulty in remembering”, “The illness interferes with my family life”, “The illness interferes with my social activities”, “I can't do my previous leisure activities”, “I can't take care of myself, such as eating and dressing”. (4) Analysis of discrete tendency: This method is used to examine the sensitivity of the items. The higher the degree of dispersion is, the stronger its ability to distinguish different symptoms shows. Items with standard deviation less than 0.7 should be deleted. The results are shown in Table 4. The items to be excluded are: “weight loss”, “ expectoration”, “hair loss”, “poor appetite”, “pain of the surgical wound”, “chest pain”, “other pain”, “vomiting”, “Difficulty in remembering”, “difficulty in concentrating”, “I'm disappointed in my struggle with the illness”, “I feel very depressed”, “Irritability”, “The illness interferes with my social activities”, “I can't do my previous leisure activities”, “The illness interferes with my family life”, “I can't take care of myself, such as eating and dressing”. (5) Analysis of multiple stepwise linear regression: This method takes the original items in the scale as independent variables, and takes the total score as dependent variables. Then we used different independent variables to estimate the dependent variables. It can calculate the contribution of independent variables to the dependent variables. It is considered to include the items with P < 0.05 into the equation. According to the results of multiple stepwise linear regression analysis (Table 5), no item should be deleted from the scale. (6) Eliminating items by the Cronbach's α coefficient and examining the scale: We respectively calculated the Cronbach's α coefficient of each item and the total scale. The overall Cronbach's α coefficient of the previous total scale is 0.837. After removal of a certain item, we compared the recalculated Cronbach's α coefficient of the total scale and the previous one. If the recalculated value of the scale increases after removing an item, the item should be deleted. The results are shown in Table 6. Thus we considered deleting items such as “vomiting”, “pain of the surgical wound”, “chest pain”, “other pain”, “insomnia”, and “difficulty in remembering”.
Table 1
The Baseline Patient Data
|
Sublobar Resection
|
Lobectomy
|
|
Item
|
No.
|
Rate (%)
|
No.
|
Rate(%)
|
P
|
Age groups
|
|
|
|
|
|
<40 years
|
10
|
8.1
|
5
|
2.2
|
<0.05
|
40–49 years
|
20
|
16.1
|
20
|
9.0
|
|
50–59 years
|
39
|
31.5
|
76
|
34.1
|
|
60–69 years
|
43
|
34.7
|
96
|
43.0
|
|
70–79 years
|
12
|
9.7
|
26
|
11.7
|
|
Marital status
|
|
|
|
|
|
Married
|
122
|
98.4
|
221
|
99.1
|
0.549
|
Widowed
|
2
|
1.6
|
2
|
0.9
|
|
Educational Backgroung
|
|
|
|
|
|
Primary School
|
49
|
39.5
|
88
|
39.5
|
0.796
|
Junior High School
|
31
|
25.0
|
66
|
29.6
|
|
High School/Technical
secondary school
|
18
|
14.5
|
26
|
11.7
|
|
Associate College
|
13
|
10.5
|
25
|
11.2
|
|
University
|
13
|
10.5
|
18
|
9.1
|
|
Gender
|
|
|
|
|
|
Male
|
15
|
12.1
|
99
|
44.4
|
<0.05
|
Female
|
109
|
87.9
|
124
|
55.6
|
|
Pathologic Classification
|
|
|
|
|
|
Squmous cell cancer
|
9
|
7.3
|
50
|
22.4
|
<0.05
|
Adenocancer
|
115
|
92.7
|
173
|
77.6
|
|
TNM Classification
|
|
|
|
|
|
IA1
|
59
|
47.6
|
22
|
9.9
|
<0.05
|
IA2
|
34
|
27.4
|
66
|
29.6
|
|
IA3
|
6
|
4.8
|
46
|
20.6
|
|
IB
|
1
|
0.8
|
16
|
7.2
|
|
IIA
|
0
|
0.0
|
8
|
3.6
|
|
IIB
|
8
|
6.5
|
29
|
13.0
|
|
IIIA
|
6
|
4.8
|
23
|
10.3
|
|
IIIB
|
0
|
0.0
|
3
|
1.3
|
|
IIIC
|
0
|
0.0
|
0
|
0.0
|
|
IVA
|
10
|
8.1
|
10
|
4.5
|
|
Smoking pack–year history
|
|
|
|
|
|
Negative
|
84
|
67.7
|
111
|
49.8
|
<0.05
|
Positive
|
40
|
32.3
|
112
|
50.2
|
|
Profession
|
|
|
|
|
|
Worker
|
44
|
35.5
|
36
|
16.1
|
<0.05
|
Farmer
|
13
|
10.5
|
29
|
13.0
|
|
Staff
|
14
|
11.3
|
27
|
12.1
|
|
Civil servant
|
3
|
2.4
|
2
|
0.9
|
|
Teacher
|
7
|
5.6
|
18
|
8.1
|
|
OTH
|
43
|
34.7
|
111
|
49.8
|
|
Nationality
|
|
|
|
|
|
Han Nationality
|
120
|
96.8
|
223
|
100.0
|
0.064
|
Manchu
|
1
|
0.8
|
0
|
0.0
|
|
Hui Nationality
|
2
|
1.6
|
0
|
0.0
|
|
Daur
|
1
|
0.8
|
0
|
0.0
|
|
Notes: P: Sublobar Resection group vs. Lobectomy none group
|
Table 2
Statistical results of Spearman correlation coefficient method
Item
|
Correlation coefficient
|
Item
|
Correlation coefficient
|
Breath shortness
|
0.753
|
Difficulty in remembering 🞺
|
0.285🞺
|
Chest tightness
|
0.74
|
Mental stress
|
0.839
|
Breathlessness
|
0.759
|
Difficulty in concentrating (reading books and newspapers, etc.)
|
0.748
|
Weight loss
|
0.307
|
I'm disappointed in my struggle with the illness
|
0.793
|
Cough
|
0.536
|
I feel very depressed
|
0.586
|
Expectoration
|
0.495
|
Irritability
|
0.793
|
Hair loss🞺
|
0.327🞺
|
The illness interferes with my social activities.
|
0.850
|
Poor appetite🞺
|
0.397🞺
|
I can't do my previous leisure activities
|
0.929
|
Pain of the surgical wound🞺
|
0.097🞺
|
I can't accept my illness
|
0.723
|
Chest pain🞺
|
0.227🞺
|
I can't take care of myself, such as eating and dressing
|
0.494
|
Other pain🞺
|
0.232🞺
|
The illness interferes with my family life
|
0.914
|
Vomiting🞺
|
0.211🞺
|
|
|
Fatigue
|
0.609
|
|
|
Insomnia🞺
|
0.266🞺
|
|
|
Notes: 🞺 ,Items to be deleted with Correlation coefficient less than 0.4
|
Table 3
Statistical results of item distribution
Item
|
Frequency of option
|
0
|
1
|
2
|
3
|
Breath shortness
|
0.57
|
0.26
|
0.10
|
0.08
|
Chest tightness
|
0.56
|
0.27
|
0.10
|
0.08
|
Breathlessness
|
0.55
|
0.28
|
0.09
|
0.08
|
Weight loss
|
0.79
|
0.15
|
0.04
|
0.02
|
Cough
|
0.46
|
0.39
|
0.07
|
0.07
|
Expectoration
|
0.54
|
0.37
|
0.09
|
0.00
|
Hair loss🞺
|
0.82🞺
|
0.10
|
0.05
|
0.02
|
Poor appetite
|
0.84
|
0.12
|
0.03
|
0.01
|
Pain of the surgical wound
|
0.89
|
0.11
|
0.00
|
0.00
|
Chest pain
|
0.83
|
0.09
|
0.06
|
0.01
|
Other pain
|
0.81
|
0.15
|
0.03
|
0.01
|
Vomiting🞺
|
0.93🞺
|
0.05
|
0.01
|
0.01
|
Fatigue
|
0.61
|
0.27
|
0.10
|
0.03
|
Insomnia
|
0.67
|
0.20
|
0.07
|
0.06
|
Difficulty in remembering 🞺
|
0.90🞺
|
0.10
|
0.01
|
0.00
|
Mental stress
|
0.59
|
0.28
|
0.11
|
0.02
|
Difficulty in concentrating (reading books and newspapers, etc.)
|
0.63
|
0.31
|
0.05
|
0.01
|
I'm disappointed in my struggle with the illness
|
0.67
|
0.26
|
0.07
|
0.01
|
I feel very depressed
|
0.71
|
0.22
|
0.07
|
0.00
|
The illness interferes with my family life🞺
|
0.81🞺
|
0.16
|
0.03
|
0.00
|
The illness interferes with my social activities🞺
|
0.84🞺
|
0.14
|
0.02
|
0.00
|
I can't do my previous leisure activities🞺
|
0.81🞺
|
0.16
|
0.03
|
0.00
|
I can't accept my illness
|
0.62
|
0.26
|
0.11
|
0.02
|
I can't take care of myself, such as eating and dressing🞺
|
0.90🞺
|
0.08
|
0.01
|
0.00
|
Irritability
|
0.67
|
0.26
|
0.07
|
0.01
|
Notes: 🞺, Items to be deleted with frequency of option greater than 0.8
|
Table 4
Statistics of scale data description
Item
|
Standard deviation
|
Item
|
Standard deviation
|
Breath shortness
|
0.936
|
Difficulty in remembering 🞺
|
0.327🞺
|
Chest tightness
|
0.942
|
Mental stress
|
0.767
|
Breathlessness
|
0.930
|
Difficulty in concentrating (reading books and newspapers, etc.) 🞺
|
0.626🞺
|
Weight loss🞺
|
0.630🞺
|
I'm disappointed in my struggle with the illness🞺
|
0.655🞺
|
Cough
|
0.883
|
I feel very depressed🞺
|
0.624🞺
|
Expectoration🞺
|
0.667🞺
|
Irritability🞺
|
0.655🞺
|
Hair loss🞺
|
0.650🞺
|
The illness interferes with my social activities🞺
|
0.433🞺
|
Poor appetite🞺
|
0.535🞺
|
I can't do my previous leisure activities🞺
|
0.477🞺
|
Pain of the surgical wound🞺
|
0.309🞺
|
I can't accept my illness
|
0.754
|
Chest pain🞺
|
0.632🞺
|
I can't take care of myself, such as eating and dressing🞺
|
0.364🞺
|
Other pain🞺
|
0.547🞺
|
The illness interferes with my family life🞺
|
0.477🞺
|
Vomiting🞺
|
0.395🞺
|
|
|
Fatigue
|
0.783
|
|
|
Insomnia
|
0.878
|
|
|
Notes: 🞺, Items to be deleted with standard deviation less than 0.7
|
Table 5
Analysis of multiple stepwise linear regression
Item
|
standardized coefficient
|
P
|
Breathlessness
|
0.130
|
🞲🞲🞲
|
Mental stress
|
0.046
|
🞲🞲🞲
|
The illness interferes with my family life
|
0.067
|
🞲🞲🞲
|
Cough
|
0.124
|
🞲🞲🞲
|
Insomnia
|
0.123
|
🞲🞲🞲
|
I’m disappointed in my struggle with the illness
|
0.088
|
🞲🞲🞲
|
I can’t accept my illness
|
0.106
|
🞲🞲🞲
|
Weight loss
|
0.088
|
🞲🞲🞲
|
Fatigue
|
0.110
|
🞲🞲🞲
|
Chest pain
|
0.089
|
🞲🞲🞲
|
I feel very depressed
|
0.087
|
🞲🞲🞲
|
Hair loss
|
0.091
|
🞲🞲🞲
|
Difficulty in concentrating (reading books and newspapers, etc.)
|
0.108
|
🞲🞲🞲
|
Breath shortness
|
0.131
|
🞲🞲🞲
|
Expectoration
|
0.093
|
🞲🞲🞲
|
Other pain
|
0.077
|
🞲🞲🞲
|
Poor appetite
|
0.075
|
🞲🞲🞲
|
I can’t do my previous leisure activities
|
0.067
|
🞲🞲🞲
|
Vomiting
|
0.055
|
🞲🞲🞲
|
Pain of the surgical wound
|
0.043
|
🞲🞲🞲
|
The illness interferes with my social activities
|
0.061
|
🞲🞲🞲
|
Difficulty in remembering
|
0.027
|
🞲🞲🞲
|
Chest tightness
|
0.132
|
🞲🞲🞲
|
I can’t take care of myself, such as eating and dressing
|
0.051
|
🞲🞲🞲
|
Irritability
|
0.092
|
🞲🞲🞲
|
Notes: 🞲🞲🞲: P < 0.005
|
Table 6
Analysis of Cronbach's α coefficient
Item
|
Cronbach's α after deleting the item
|
Item
|
Cronbach's α after deleting the item
|
Breath shortness
|
0.819
|
Difficulty in remembering 🞺
|
0.839🞺
|
Chest tightness
|
0.819
|
Mental stress
|
0.825
|
Breathlessness
|
0.818
|
Difficulty in concentrating (reading books and newspapers, etc.)
|
0.831
|
Weight loss🞺
|
0.833🞺
|
I'm disappointed in my struggle with the illness
|
0.827
|
Cough
|
0.833
|
I feel very depressed
|
0.827
|
Expectoration
|
0.835
|
Irritability
|
0.827
|
Hair loss🞺
|
0.835🞺
|
The illness interferes with my social activities
|
0.832
|
Poor appetite🞺
|
0.832🞺
|
I can't do my previous leisure activities
|
0.831
|
Pain of the surgical wound🞺
|
0.839🞺
|
I can't accept my illness
|
0.831
|
Chest pain🞺
|
0.840🞺
|
I can't take care of myself, such as eating and dressing
|
0.832
|
Other pain🞺
|
0.839🞺
|
The illness interferes with my family life
|
0.830
|
Vomiting🞺
|
0.838🞺
|
|
|
Fatigue
|
0.828
|
|
|
Insomnia🞺
|
0.842🞺
|
|
|
Notes:🞺: The overall Cronbach's α coefficient increased after the entry is deleted,
|
In combination with the above steps, it was considered that items which met three or more of the five screening methods should be deleted. As pain is an important part of the evaluation of PQOL, relevant items were retained after consulting experts reference group. Finally, three items would be deleted to form the final scale (Annex Table 2) including “hair loss”, “vomiting” and “difficulty in remembering”.
Evaluation Of The Final Scale
Feasibility. The return rate and effective response rate of the clinical questionnaire were both 97.5%. All the questionnaires returned were filled completely without missing items. The vast majority of patients understood the contents of each item. The mean filling time was 11.25 ± 2.54 minutes which showed the patients accepted the scale well.
Reliability evaluation. This study adopted the reliability evaluation of calculating the Cronbach's α coefficient. The Cronbach's α coefficient of the total scale, the signs and symptoms subscale, the Psychological and Psychiatric subscale and the social life subscale were 0.838, 0.758, 0.849 and 0.879 (Table 7). All the coefficients were greater than 0.7 (13). Thus we can conclude that the internal consistency of the scale was good.
Table 7
Analysis of Cronbach's α coefficient
Item
|
Cronbach's α coefficient
|
Spearman–Brown coefficient
|
Total Scale
|
0.84
|
0.62
|
Signs and Symptoms Domain
|
0.76
|
0.63
|
Psychological and Psychiatric Domain
|
0.85
|
0.87
|
Social Life Domain
|
0.88
|
0.88
|
The Spearman–Brown coefficient of the total scale, the signs and symptoms domain, Psychological and Psychiatric Scale and social life scale were 0.616, 0.628, 0.871 and 0.882 (Table 7). As the Spearman–Brown coefficients of the signs and symptoms and the total scale were close to 0.7, the consistency of the total scale and the cross–item consistency of each dimension were acceptable, and the contents reflected by the scale were relatively same and stable.
In conclusion, we can conclude that the reliability of NSCLC–PQOL scale was good.
Validity. (1) Content validity: The NSCLC–PQOL scale combined FACT–L scale, LCSS scale and the SF–36 Short Form Health Measurement Scale. The development process was based on the principles of the international QOL scale development. The item pool of the scale was created combining the medical records reviews, the literature analysis and other sources. Then the items were screened at multiple levels to form the final scale, so the NSCLC–PQOL scale had good content validity. (2) Structural validity: The Spearman coefficient was calculated to determine the correlation between the subscale and the total scale (Table 8). The correlation coefficients between each subscale and the total scale were over 0.4, which showed that the subscales have good correlation with the total scale and the contents of the scales were independent of each other but coming from the same unity.
Table 8
Correlation coefficient between each subscale and the total scale
Subscale
|
Total scale
|
Signs and symptoms scale
|
0.843
|
Psychological and Psychiatric scale
|
0.727
|
Social life scale
|
0.441
|
Application of scale scoring: The research group evaluated the scale data to analyze the difference in QOL of NSCLC patients after lobectomy and sublobar resection. The total score was the indicator of overall declines in QOL six months after surgery. The average scale scores of the two groups were shown in Table 9. And the Mann–Whitney U test was performed to determine whether there was any difference in the rank average of the scale score between the two groups of patients.
Table 9
Comparisons of Scale Score of two groups
Item
|
Scale Scores
|
Sublobar Resection Group
|
8.07 ± 6.02
|
Lobectomy Group
|
9.55 ± 7.68
|
Notes: Data was expressed as mean ± standard deviation and was reserved for second decimal place.
|
It can be seen that the respiratory symptoms of patients in sublobar resection group were significantly better than those in lobectomy group on the indicators of “breath shortness”, “chest tightness”, “breathlessness”, “cough” and “expectoration” (P < 0.05) (Table 10). There was no significant difference between the two groups in postoperative pain, such as “pain of the surgical wound”, “chest pain” and “other pain”.
Table 10
Comparison of Post–operative Symptoms of two groups
Item
|
Post–operative Scale Score
|
P
|
Lobectomy group
|
sublobar resection group
|
Breath shortness
|
186.90
|
150.80
|
🞲🞲🞲
|
Chest tightness
|
184.33
|
155.42
|
🞲🞲🞲
|
Breathlessness
|
183.83
|
156.33
|
🞲🞲🞲
|
Cough
|
184.49
|
155.13
|
🞲🞲🞲
|
Expectoration
|
183.94
|
156.12
|
🞲🞲🞲
|
Notes:🞲🞲🞲: P < 0.005; Data was reserved for second decimal place.
|
We also evaluated several factors that might be the confounding factors of the PQOL with univariate and multivariate analysis, these factors included TNM classification, age, gender, morphology and smoking pack–year history. The results are shown in Table 11.
Table 11
The Results of the Single Factor Analysis of Respiratory Symptoms Developed after Surgery
Factor
|
Rank Average of Breath Shortness
|
p
|
Rank Average of
Chest Tightness
|
p
|
Rank Average of
Breathlessness
|
p
|
Rank Average of Cough
|
p
|
Rank Average of Expectoration
|
p
|
Smoking pack–year history
|
|
0.14
|
|
0.14
|
|
0.18
|
|
0.60
|
|
0.61
|
Smokers
|
182.02
|
|
182.15
|
|
181.32
|
|
176.91
|
|
171.20
|
|
Nonsmokers
|
167.75
|
|
167.64
|
|
168.29
|
|
171.73
|
|
176.18
|
|
Pathologic Classification
|
|
0.41
|
|
0.65
|
|
0.58
|
|
0.18
|
|
0.29
|
Squmous Cell denocancer
|
182.69
|
|
178.78
|
|
179.96
|
|
188.72
|
|
185.31
|
|
Adenocancer
|
172.22
|
|
173.02
|
|
172.78
|
|
170.98
|
|
171.68
|
|
Gender
|
|
0.44
|
|
0.55
|
|
0.62
|
|
0.36
|
|
0.96
|
Male
|
170.46
|
|
171.30
|
|
171.72
|
|
169.77
|
|
173.79
|
|
Female
|
177.99
|
|
177.05
|
|
176.57
|
|
178.87
|
|
174.24
|
|
Morphology
|
|
0.11
|
|
0.11
|
|
0.18
|
|
0.52
|
|
0.45
|
lung nodules
|
169.32
|
|
169.37
|
|
170.10
|
|
172.08
|
|
171.79
|
|
pulmonary mass
|
186.78
|
|
186.64
|
|
184.64
|
|
179.24
|
|
180.05
|
|
TNM Classification
|
|
0.31
|
|
0.36
|
|
0.42
|
|
0.31
|
|
0.39
|
IA1
|
158.42
|
|
164.19
|
|
164.81
|
|
152.57
|
|
158.69
|
|
IA2
|
179.82
|
|
173.36
|
|
177.51
|
|
175.09
|
|
170.91
|
|
IA3
|
176.18
|
|
178.69
|
|
176.17
|
|
188.41
|
|
191.44
|
|
IB
|
192.97
|
|
191.50
|
|
185.35
|
|
171.97
|
|
164.00
|
|
IIA
|
189.88
|
|
200.06
|
|
191.50
|
|
166.69
|
|
143.75
|
|
IIB
|
170.73
|
|
166.09
|
|
164.30
|
|
198.09
|
|
192.88
|
|
IIIA
|
172.47
|
|
173.86
|
|
173.19
|
|
184.09
|
|
189.57
|
|
IIIB
|
293.50
|
|
302.50
|
|
302.67
|
|
128.67
|
|
147.00
|
|
IIIC
|
0.00
|
|
0.00
|
|
0.00
|
|
0.00
|
|
0.00
|
|
IVA
|
170.20
|
|
174.98
|
|
171.20
|
|
170.15
|
|
173.30
|
|
Age groups
|
|
0.19
|
|
0.16
|
|
0.12
|
|
0.16
|
|
0.18
|
<40 years
|
128.33
|
|
126.93
|
|
125.77
|
|
118.83
|
|
126.00
|
|
40–49 years
|
167.88
|
|
173.51
|
|
171.08
|
|
183.18
|
|
181.15
|
|
50–59 years
|
184.84
|
|
184.34
|
|
184.91
|
|
175.79
|
|
175.33
|
|
60–69 years
|
170.27
|
|
168.15
|
|
167.89
|
|
172.22
|
|
171.12
|
|
70–79 years
|
179.30
|
|
183.21
|
|
185.46
|
|
187.21
|
|
191.95
|
|
Notes: Data was reserved for second decimal place.
|
Mann–Whitney U test showed that gender, morphology and smoking pack–year history were not the factors of the respiratory symptoms. Then we used Kruskal–Wallis test to identify whether TNM classification and age had impact on PQOL. The results also showed that there was no difference in “breath shortness”, “chest tightness”, “breathlessness”, “cough” and “expectoration” between different age groups and TNM classifications. Considering the great impact of these facts on respiratory symptoms, we also included these five factors in the multivariate analyses. After evaluated the model by Chi–square test, we figured out that the difference was not statistically significant (χ² = 21.266, P = 0.169). Thus, surgical procedure is an important factor which could infect the PQOL, and TNM classification, age, gender, morphology and smoking pack–year history were proved to be the notable confounder of PQOL.