International Journal of Colorectal Disease

, Volume 19, Issue 6, pp 554–560

Comparison of EORTC Quality of Life Core Questionnaire (EORTC-QLQ-C30) and Gastrointestinal Quality of Life Index (GIQLI) in patients undergoing elective colorectal cancer resection

Authors

    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
  • J. Neudecker
    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
  • O. Haase
    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
  • W. Raue
    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
  • T. Strohm
    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
  • J. M. Müller
    • Universitätsklinik für Allgemein-, Visceral-, Gefäss- und ThoraxchirurgieCharité, Campus Mitte
Original Article

DOI: 10.1007/s00384-004-0609-1

Cite this article as:
Schwenk, W., Neudecker, J., Haase, O. et al. Int J Colorectal Dis (2004) 19: 554. doi:10.1007/s00384-004-0609-1

Abstract

Background

EORTC-QLQ-C30 questionnaires and GIQLI questionnaires are used to evaluate post-operative quality of life (QoL). It was not clear whether results of both instruments are comparable. Therefore, the level of agreement between both QoL questionnaires was evaluated in patients undergoing elective colorectal cancer resection.

Methods

Pre-operatively, 7 and 30 days after surgery 116 patients answered the EORTC-QLQ-C-30 and the GIQLI questionnaires in random order. Individual questions with similar content from each questionnaire were compared. Data for global QoL, physical (PF), emotional (EF) and social function (SF) were linearly transformed to fit a scale from 0 to 100. Data from the two instruments were correlated and the level of agreement between them was calculated according to the method of Bland and Altman.

Results

A total of 308 data sets [(pre-op. n=116; 7th pod n=101; 30th post-operative day (pod) n=91)] were evaluated. Both instruments detected a reversible reduction of QoL after surgery and gave inferior results for patients with conditions known to impair QoL. EORTC-QLQ-C30 was more sensitive than GIQLI. The correlation between the two questionnaires for global QoL, PF and EF was good (r=0.53–0.66, p<0.01), but no correlation for SF was detected (r=−0.44, p=0.44). Linearly transformed scores from the two instruments differed considerably from −13 (95%CI −51 to 24) points (QoL) to 10 (−38 to 58) points (PF).

Conclusion

Although EORTC-QLQ-C30 scores and GIQLI scores from patients undergoing elective colorectal cancer surgery did correlate well, the level of agreement between the two instruments was quite low. Perioperative QoL data from the two instruments cannot be compared with each other.

Keywords

Colorectal resectionQuality of lifeEORTC-QLQ-C30GLQI

Introduction

Post-operative morbidity, mortality, recurrence, and survival rates are classic criteria to evaluate the outcome of patients undergoing surgery for malignant tumors. In the last decade, research has focused on the influence of surgery on patients’ quality of life (QoL) [14]. QoL can be assessed with far more than 50 different instruments [5]. Most of them have been thoroughly evaluated for reliability and validity for different languages and cultural backgrounds [6]. In Germany, the EORTC Quality of Life Core-30-Questionnaire (EORTC-QLQ-C30) [6] and the Gastrointestinal Quality of Life Index (GIQLI) [7] are commonly used to describe QoL after resection of colorectal cancer. Both EORTC-QLQ-C30 and GIQLI determine global QoL as well as physical (PF), emotional (EF), and social function (SF) of selected individuals. It has not been evaluated so far whether the results ofthe two QoL instruments agree with each other. Therefore, the results from patients undergoing QoL surveys within prospective trials concerning elective resections of colorectal carcinoma were investigated to assess the level of agreement between EORTC-QLQ-C30 and GIQLI.

Methods

Aim of the study

The aim of this prospective study was to evaluate the level of agreement between QoL domains measured with the EORTC-QLQ-C30 or the GIQLI in patients undergoing elective resection of colorectal carcinoma.

Patients

One hundred sixteen patients who underwent elective resection for colorectal carcinoma in the Department of General, Visceral, Vascular and Thoracic Surgery of the Charité Campus Mitte in Berlin were included in the evaluation. Patients undergoing emergency procedures or surgery without resection (i.e., enterostoma creation only) were excluded.

Study design

The evaluation of QoL was approved by the local ethics committee. Pre-operatively, all patients were informed about the kind of disease and the extent of surgery. In all patients QoL was measured with the EORTC-QLQ-C30 (Version 1.0) [6] and the GIQLI. Laparoscopic surgery was performed using a 5-trocar technique described by Milsom and Böhm in detail [8]. Conventional resections were carried out using a wide midline incision.

Peri-operative treatment

After surgery, all patients were admitted to the surgical intensive care unit and received patient controlled morphine analgesia utilizing electronic pumps (PCA-PACOM-Injectomat [7], Fresenius, Germany). All patients received no oral diet on the day of surgery, tea or water on day 1, soup on day 2, smashed food on day 3 and a regular hospital diet on day 4 after surgery. Intravenous electrolyte solutions were administered as indicated.

Measurement of QoL

QoL was measured 2 days before surgery as well as 7 and 30 days after surgery. All investigations were carried out by a study nurse well educated in the principles and techniques of QoL measurement. All patients were evaluated with the EORTC-QLQ-C30 and the GIQLI handed out to them in random order. All patients were asked to answer the questionnaires by themselves without any external help from physicians, nurses, or relatives. Both QoL instruments have been extensively evaluated; reliability and validity were thoroughly investigated.

EORTC Quality of Life Core 30 Questionnaire [6]

This instrument is a structured questionnaire developed by the Quality of Life research group of the European Organization for Research and Treatment of Cancer (EORTC) especially for patients with malignant tumors. The EORTC-QLQ-C30 consists of 30 questions, which can be answered by the patients using Lickert scales. Twenty-four of the 30 questions can be combined to nine scales that will express different aspects of health-related QoL: global QoL, five functional scales (physical, emotional, cognitive, social function, and role function) and three scales for symptom severity (fatigue, pain, and nausea/vomiting). The remaining six questions describe the intensity of disease-related symptoms (dyspneoa, sleeping disturbances, loss of appetite, constipation, diarrhea, and treatment-related financial problems). All data are calculated following exact algorithms given by the EORTC-Quality of Life Study Group [9]. Within this calculation all data undergo linear transformation and range from 0 (least QoL or function) to 100 (the best QoL or function). The scales for severity of symptoms range from 0 (least severity) to 100 (worst severity).

Gastrointestinal Quality of Life Index [7]

This questionaire was developed by an international board of experts to measure QoL in patients with gastrointestinal diseases and especially in those undergoing surgical treatment of gastrointestinal diseases. The questionnaire consists of 36 questions answered by the patient using Lickert scales. Each question can be assigned to the following aspects of QoL: symptoms, physical function, emotional function, social function, and medical function. The results for each aspect of QoL are calculated by simply summarizing the points given for each question related to the aspect under investigation. Global GIQLI is calculated by summarizing the points from all 36 questions. The maximum score for each dimension of QoL is: physical function 28, emotional function 20, and social function 16. The highest score for QoL is 144 points.

Retrieval of data and statistical analysis

Statistical analysis was performed at the Department of General, Visceral, Vascular, and Thoracic Surgery of the Charité Campus Mitte using the “Statistical Analysis Systems SAS V8 for WindowsXP.” First, individual questions with similar content from the two questionnaires were compared. Then, QoL and its dimensions were calculated according to algorithms given by the EORTC goup and the GIQLI group. Next, results from the EORTC and GIQLI were transformed linearly to numerical values from 0 (worst result) to 100 (best result). Finally, level of agreement for global QoL, physical (PF), emotional (EF), and social function (SF) were calculated using the method described by Bland and Altman [10]. In brief, Pearson’s correlation coefficients for the results of the two instruments were calculated; the unknown “true” values of QoL, PF, EF, and SF were estimated by determining the mean of both measurements EORTC–GIQLI/2. Then, differences between the two scores (EORTC–GIQLI) were calculated for each patient and plotted against the estimated “true” value; standard deviation of the difference between the two instruments (“repeatability coefficient”) and 95% confidence intervals of the differences between the two measurements (“lower and upper level of agreement”) were calculated. To evaluate the sensitivity of both instruments, QoL from patient groups with special conditions known to affect QoL (post-operative morbidity vs no morbidity; enterostomy vs no enterostomy; laparoscopic vs conventional resection) were compared. Continuous parameters following a normal distribution are given as mean (standard deviation); non-normally distributed parameters are given as median (25–75 percentile). Data in figures are shown as box-and-whisker plots.

Results

Patients, surgical treatment and post-operative course

From January 1999 to July 2001 116 patients undergoing elective colorectal cancer resection were included in this study. In 23 patients, cancer was resected (19.8%) laparoscopically and 93 underwent conventional resection (80.2%). Epidemiological data, tumor stage, and type of resection are given in Table 1. Post-operative hospital stay was 11 (8–24) days.In 27 patients (23.3%) with low anterior rectal resection and total mesorectal excision, a protective loop ileostomy was performed. In 14 patients (12.1%) an abdominoperineal excision of the rectum with permanent colostomy was performed. Eighty-six patients (74.1%) had an uneventful post-operative course, six patients (5.2%) developed general complications (i.e., pneumonia, cardiac complication, and urinary tract infection) and in 23 patients (19.8%) surgical complications (i.e., wound infection, ileus, and anastomotic leakage) were observed. Surgical complications were diagnosed in only one patient (4%) undergoing laparosopic surgery. After conventional resection 22 patients (24%) developed surgical and six patients (8%) developed general complications. There was no mortality.
Table 1

Epidemiological data, UICC stage of tumor and type of resection performed in 116 patients undergoing elective resection of colorectal carcinoma

 

Conventional (n=93)

Laparoscopic (n=23)

n

(%)

n

(%)

Sex

 Female

39

42

12

48

 Male

54

58

11

52

UICC stage

 0 (Adenoma)

15

16

4

17

 I

20

22

6

26

 II

29

31

8

35

 III

29

31

5

22

Type of resection

 Right colectomy

4

5

5

22

 Extended colectomy

16

17

0

 Sigmoidectomy

11

12

12

52

 Rectal resection

45

48

6

26

 Abdom. perin. extirpation

14

15

0

 Other

3

4

0

Age (years), mean (SD)

63.1 (8.0)

63.9 (7.3)

Quality of life

Pre-operatively, all 116 patients answered both questionnaires. On the 7th post-operative day both questionnaires were answered by 101 patients (87%) and after 30 days by 91 patients (78%). A total of 308 EORTC-QLQ-C30 and GIQLI questionnaires were evaluated. Both instruments reflected the expected post-operative reduction of QoL 7 days post-operatively. Thirty days after surgery, QoL, physical, emotional, and social function had almost returned to their pre-operative values (Fig. 1). When compared to the pre-operative score, EORTC-QLQ-C30 QoL index was still reduced 30 days after surgery. Patients with complications expressed a reduced post-operative QoL in both instruments (Table 2). In both instruments, post-operative QoL was better after laparoscopy than after conventional colorectal resection (Table 2). As expected, QoL was decreased in patients who had an enterostoma created during surgery (Table 2). In all comparisons, changes of EORTC-QLQ-C30 QoL scores were larger than changes in GIQLI scores.
Fig. 1

Peri-operative changes in global QoL measured with EORTC-QLQ-C30 and GIQLI in 116 patients undergoing elective resection of colorectal carcinoma

Table 2

Influence of post-operative morbidity, operative technique and enterostoma creation on the results of EORTC-QLQ-C30 and GIQLI in 116 patients undergoing elective resection of colorectal carcinoma (data given as median). Bold figures indicate p<0.05 between groups with or without specific condition (Mann–Whitney-U-test)

 

Pre-operative

Day 7

Day 30

EORTC/GLQI

EORTC/GLQI

EORTC/GLQI

Morbidity (all patients)

 No (n=86)

  QL

67/72

50/68

67/69

  PF

100/75

60/52

80/54

  SF

83/50

50/75

83/63

  EF

67/55

75/55

75/60

 Yes (n=30)

  QL

67/72

41/65

50/68

  PF

100/71

40/46

60/50

  SF

67/50

33/75

67/69

  EF

56/50

58/60

67/65

Operative technique (patients without morbidity only, n=86)

 Conventional (n=65)

  QL

67/72

50/69

58/69

  PF

100/75

60/50

80/54

  SF

83/50

58/78

83/63

  EF

67/55

75/55

76/60

 Laparoscopic (n=21)

  QL

63/72

54/64

67/70

  PF

100/71

70/54

90/61

  SF

83/50

50/75

92/63

  EF

54/50

58/50

71/60

Enterostomy (patients after conventional surgery only, n=93)

 No (n=52)

  QL

67/72

50/73

67/70

  PF

100/79

60/61

80/59

  SF

92/50

67/75

67/63

  EF

67/55

79/60

83/65

 Yes (n=41)

  QL

67/71

42/66

50/68

  PF

100/71

40/39

62/46

  SF

83/50

50/75

67/69

  EF

67/55

67/55

67/60

Level of agreement between EORTC-QLQ-C30 and GIQLI

Questions concerning pain perception, sleeping disorders, fatigue, nausea, constipation/diarrhea, sleepiness, personal concerns, and family life from both questionnaires allowed direct comparison of data. Although most of the questions yielded comparable results, there were relevant differences between the two questionnaires for questions concerning sleeping disorders, fatigue, nausea, and sleepiness (Table 3).
Table 3

Results of individual questions from EORTC-QLQ-C30 and GIQLI in 116 patients undergoing elective resection of colorectal carcinoma. (Individual questions from the GIQLI converted to the Lickert scale used in the EORTC-QLQ-C30: 1= not at all to 4= very much. Data given as median and 25–75 percentile)

EORTC-QLQ-C30

Pre-operative

Day 7

Day 30

GIQLI

Pre–operative

Day 7

Day 30

Have you had pain?

1 (1–2)

3 (2–3)

2 (1–3)

How often in the past 2 weeks have you had pain in the abdomen?

1 (1–2)

2 (2–3)

1 (1–2)

Have you had trouble sleeping?

2 (1–3)

3 (2–4)

2.5 (2–4)

Over the past week, have you woken up in the night?

2 (0–3)

3 (3–4)

2 (2–4)

Have you felt weak?

1 (1–2)

3 (2–3)

3 (2–3)

Because of your illness, how much physical strength have you lost?

2 (2–4)

3 (1–3)

2 (1–3)

Have you felt nauseated?

1 (1–1)

2 (1–2)

1 (1–2)

How often during the past 2 weeks have you been troubled by nausea?

1 (1–1)

2 (1–2)

1 (1–2)

Have you been constipated?

1 (1–2)

1 (1–1)

1 (1–1)

How often during the past 2 weeks have you been troubled by constipation?

1 (1–2)

1 (1–2)

1 (1–2)

Have you had diarrhea?

1 (1–2)

1 (1–2)

1 (1–2)

How often during the past 2 weeks have you been troubled by diarrhea?

1 (1–2)

1 (1–2)

1 (1–2)

Have you been tired?

2 (1–2)

3 (2–3)

2 (2–3)

How often during the past 2 weeks have you been tired or fatigued?

3 (2–4)

2 (2–3)

2 (2–2)

Have you been worried?

3 (2–4)

2 (2–3)

2 (1–3)

How often during the past 2 weeks have you been nervous or anxious about your illness?

2 (1–3)

2 (2–2)

2 (1–2)

Has your physical condition or medical treatment interfered with your family life?

1.5 (1–2.5)

2 (1–3)

1 (1–3)

To what extent have your personal relations with people close to you (family or friends) worsened because of your illness?

1 (1–2)

1 (1–1)

1 (1–2)

Moderate to good correlation for global QoL scores (r=0.49–0.55; Fig. 2), PF (r=0.35–0.64), and EF (r=0.59–0.71) between both instruments were detected (Table 4). Correlations between SF scores from EORTC-QLQ-C30 and GIQLI were only weak or not detected at all (r=−0.16 to 0.43; Table 4). No significant influence of the covariates age or sex on correlations between EORTC-QLQ-C30 and GIQLI were detected (data not shown). Figure 3 shows the differences in QoL between EORTC-QLQ-C30 and GIQLI when plotted against the estimated “true” QoL value. For global QoL the mean difference between the two questionnaires was −13 points and the difference increased with increasing QoL (Fig. 3). The lower and upper level of agreement for both questionaires ranged from −51 to 24 points and the repeatability coefficient was 19 points. Results from physical, emotional, and social function are given in Table 4. Separate analysis of all values for the individual points of measurement (pre-operative, 7th and 30th post-operative day) revealed results similar to the overall analysis (Table 4).
Fig. 2

Correlation of scores for global QoL measured with EORTC-QLQ-C30 and GIQLI in 116 patients undergoing elective resection of colorectal carcinoma

Fig. 3

Differences in scores for QoL measured with EORTC-QLQ-C30 and GIQLI in relation to the estimated “true” value of QoL in 116 patients undergoing elective resection of colorectal carcinoma

Table 4

Correlation, mean difference (repeatability coefficient), lower and upper level of agreement of scores for QoL, physical, emotional and social function in EORTC-QLQ-C30 and GIQLI of 116 patients undergoing elective resection of colorectal carcinoma

 

Pre-operative

Day 7

Day 30

Total

Quality of life

 Mean difference (repeatability coefficient)

−5 (18)

−20 (17)

−10 (15)

−13 (19)

 Estimated (“true“ value)

68

57

64

62

 Correlation coefficient

r=0.49; p<0.001

r=0.52; p<0.001

r=0.53; p<0.001

r=0.55; p<0.001

 Level of agreement

−40 to 29

−54 to 14

−40 to 20

−51 to 24

Physical function

 Mean difference

18 (23)

5 (22)

17 (21)

10 (25)

 Estimated (“true” value)

80

53

63

62

 Correlation coefficient

r=0.35; p<0.001

r=0.64; p<0.001

r=0.48; p<0.001

r=0.63; p<0.001

 Level of agreement

−27 to 63

−39 to 49

−25 to 59

−38 to 58

Emotional function

 Mean difference

13 (18)

8 (21)

9 (16)

9 (20)

 Estimated (“true” value)

58

60

64

60

 Correlation coefficient

r=0.59, p=0.01

r=0.61;p<0.001

r=0.71; p<0.001

r=0.58; p<0.001

 Level of agreement

−23 to 49

−33 to 48

−22 to 41

−29 to 47

Social function

 Mean difference

20 (34)

−25 (36)

1 (35)

−4 (41)

 Estimated (“true” value)

63

67

65

65

 Correlation coefficient

r=−0.16, p<0.01

r=0.43; p<0.01

r=0.10; p=0.08

r=−0.04; p=0.44

 Level of agreement

−47 to 87

−95 to 46

−67 to 69

−83 to 76

Discussion

More than 50 instruments are available to evaluate QoL in different patient groups. However, only a few of those instruments have been used to investigate QoL in patients undergoing surgery [11]. The EORTC-QLQ-C30 was invented for patients with malignant diseases [6]. Disease-related modules can be connected to the Core-30 questionaire to investigate special conditions unique to different types of cancer [12, 13]. The GIQLI was especially created to investigate QoL in patients undergoing surgical procedures [7]. Therefore, GIQLI is not specific to cancer patients but may also be used in patients undergoing surgery for benign diseases like cholecystolithiasis [7]. It is important, however, to realize that neither instrument has ever been tested for reliability or validity in patients who undergo elective colorectal cancer resection.

In both instruments specific questions concerning certain aspects of QoL are answered by the patients using Lickert scales (Table 3). However, both questionnaires follow different approaches to assess QoL on the level of individual questions as well as in the algorithms to calculate QoL scores and function indices. While EORTC-QLQ-C30 questions focus on the presence or absence of a specific condition (i.e., “Have you had pain?”), GIQLI questions always incorporate a quantitative approach to the condition under question (“How often in the past 2 weeks have you had pain in the abdomen?”). This difference may explain why different answers are given to individual questions in the two instruments. Furthermore, the algorithms to calculate QoL indices show striking differences between the two instruments which may also be responsible for the different results. EORTC-QLQ-C30 QoL scores are calculated from only two of 30 questions [9], but the GIQLI-QoL score is the sum of all 36 questions [7]. A change of one point in the Lickert scale of one EORTC question may change the result of the EORTC-QoL score, while a change of one point on the Lickert scale in one GIQLI question only has a small influence on the GIQLI-QoL score. Therefore, the GIQLI may be rather “robust” to changes in QoL caused by certain conditions known to impair QoL (i.e., post-operative complications or stoma creation). Another explanation for the GIQLI’s lack of sensitivity in detecting differences between these subgroups may be the fact that the GIQLI focuses on upper gastrointestinal tract symptoms like belching or bloating and is therefore less suited to identify QoL impairments caused by other symptoms typical for colorectal resections.

Several ways to describe the level of agreement between two instruments have been described in the literature but because of its unique advantages, the approach published by Bland and Altman has been chosen for this study [10]. When the results of EORTC-QLQ-C-30 questionnaires and GIQLI questionnaires were compared in 116 patients undergoing elective resection of colorectal cancer, correlation was good for global QoL (r=0.55), PF (r=0.63), and EF (r=0.58) suggesting that higher scores in the EORTC-QLQ-C30 will be accompanied by higher scores in the GIQLI. Surprisingly, no correlation of SF scores on the two instruments (r=−0.04) were found when the results of all 306 interviews were analyzed. By estimating the unknown “true” QoL value, calculating the difference between the two instruments and plotting both values in relation to each other, the following information can be obtained from our data: (1) QoL scores from EORTC-QLQ-C30 are 13 points lower on average than GIQLI scores; (2) the differences between the two instruments increase with increasing QoL scores; (3) with a mean “true” QoL value of 62 points, the standard deviation of the difference between the two instruments is quite high (19 points=31% of the mean “true” value); (4) the 95% confidence interval of the mean difference between the two instruments is very large (−51 to 24 points = −84 to 39% of the mean “true” value). Although correlation coefficients of greater than 0.5 are considered excellent in sociographic or psychological research, analysis of the level of agreement between the two instruments revealed large differences between the two instruments for QoL, PF, EF, and SF. These differences prohibit direct comparison of EORTC-QLQ-C30 data and GIQLI data.

Conclusion

The comparison of QoL data obtained with the EORTC-QLQ-C30 and the GIQLI after elective colorectal cancer resection is problematic. In general, higher EORTC-QLQ-C30 scores will be associated with higher GIQLI indices. However, detailed analysis revealed that individual QoL scores from the two questionnaires may differ by as much as 30% (95%CI: −84 to 39%). At the same time the EORTC-QLQ-C30 was more sensitive in detecting the impairment of QoL in the early post-operative period. This greater sensitivity is the main reason for us to continue using the EORTC-QLQ-C30 as a standard instrument to assess peri-operative changes of QoL in patients undergoing colorectal surgery. However, this may not hold true for other surgical procedures. Especially after upper gastrointestinal surgery the GIQLI might have an increased sensitivity. In general, there is a need for more research in this field to develop symptom-related instruments for reliable and valid assessment of short-term post-operative QoL.

Copyright information

© Springer-Verlag 2004