Factors Influencing Health Related Quality of Life in Cancer Patients with Bone Metastases

Abstract

Objective:

Health related quality of life (HRQOL) is a multidimensional concept that is especially important for cancer patients with bone metastases, as maintaining and improving HRQOL is often the main focus of treatment. This study aims to determine factors that may influence HRQOL, which may in turn influence treatment and care of patients.

Methods:

Patients (n=396) completed the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ) Bone Metastases module (BM22) at baseline. The EORTC QLQ-BM22 consists of four scales: painful site (PS), pain characteristics (PC), functional interference (FI), and psychosocial aspect (PA) scales. EORTC QLQ-BM22 data, together with sociodemographic and medical factors were analyzed by univariate analysis of variance (ANOVA). Items of significance were determined through backward selection, which were then put through multivariate analysis to determine further significance.

Results:

Through ANOVA analysis, KPS>80 and breast primary histology were predictive of better HRQOL in the PS scale, while KPS>80, female gender, and breast primary histology were predictive of better HRQOL in the PC and FI scales. KPS>80 and prostate primary histology were predictive of better HRQOL in the PA scale. KPS>80 and primary cancer site were confirmed as significant predictive factors in multivariate analysis.

Recommendations:

This study identified baseline factors of gender, performance status, and primary histology as determinants of HRQOL in patients with bone metastases. Further study focusing on current treatment (chemotherapy, bisphosphonates, and radiotherapy) and spiritual well-being may identify additional factors affecting HRQOL. Understanding the influence of these factors will allow health care professionals to provide more effective palliative care.

Introduction

Health related quality of life (HRQOL) is a multidimensional concept that encompasses general health, emotional and social well-being, physical functioning, physical symptoms, cognition, role functioning, sexual functioning, and spirituality.^1,2 In advanced cancer patients, including those with bone metastases, HRQOL is of utmost importance, as one of the main goals of treatment is the improvement of HRQOL, in addition to the reduction of symptoms.^3–6 HRQOL is influenced by multiple factors including the level of performance status, remaining survival time, employment, age, and gender of the patient.^1,2,7 Consequently, the determination of additional factors that may affect HRQOL is worthwhile, as they can influence both treatment decisions and the impact of treatment.⁸

In the current literature, emphasis has been placed on studying the influence of sociodemographic and medical factors on HRQOL in the general population.⁹ Several studies have been conducted on the influence of certain factors on HRQOL in advanced cancer patients, of which four key studies have reported the correlation of sociodemographic and medical factors with HRQOL using both the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ) – Core 30 (EORTC QLQ-C30) and the Functional Assessment of Cancer Therapy – General (FACT-G).^7,10–12 However, these parameters have yet to be investigated for the specific subgroup of cancer patients with bone metastases. The objective of this study is to determine whether sociodemographic and medical factors affect HRQOL in patients with bone metastases as assessed by the EORTC QLQ Bone Metastases module (BM-22) (see Appendix 1), such that health care professionals are better equipped to provide optimal treatment and care.

Methods

Patients

A retrospective analysis was conducted using a database of 396 patients with bone metastases who were recruited prospectively in various countries from March 2010 to January 2011. Treatment decisions were based on physician discretion and were grouped into radiotherapy; bisphosphates; orthopedic stabilization; minimally invasive procedures (vertebroplasty, cementoplasty, and kyphoplasty); and analgesic adjustments. Patients completed the EORTC QLQ-BM22 and EORTC QLQ-C15-PAL at baseline, and sociodemographic and medical information were recorded. Follow-up assessment (EORTC QLQ-BM22 and EORTC QLQ-C15-PAL) took place one week post-treatment for patients with no new treatments or one month post-treatment for patients receiving any type of treatment either by telephone or in clinic.¹³

Data collection

The EORTC QLQ-BM22 is a 22-item questionnaire developed to supplement the EORTC QLQ-C30 in assessing HRQOL of cancer patients with bone metastases. The 22 questions are divided into two subgroups of scales: symptom and functional scales. Symptom scales consist of two scales: painful sites (five questions) and pain characteristics (three questions). For these two scales, the higher the score, the worse the situation. Functional scales include the functional interference (eight questions) and psychosocial aspects (six questions) scales. For these, higher scores indicate better functioning.^5,15

Demographics and medical information collected included age, Karnofsky performance status (KPS), gender, marital status, educational level, employment status, cohabitation status, and primary histology. Age, gender, and marital status were divided into binary categories: age older than or younger than 60 years, female or male, and married or unmarried. Educational level was divided into completion of elementary school, high school, university, masters or PhD, and other. Employment status was divided into currently employed, unemployed, and retired. Cohabitation status was divided into spouse, spouse and children, children, alone, and other. Finally, histology was divided into breast, prostate, lung, gastrointestinal, renal cell, and others.

Statistical analysis

Mean and median scores were calculated for each scale, and linear transformation (0–100) was carried out. For painful sites (PS), pain characteristics (PC), functional interference (FI), and psychosocial aspects (PA) scales, if more than three, two, four, and three items were missing respectively, the scale was treated as missing. Simple univariate analysis of variance (ANOVA) was applied to detect significant relationships between each EORTC QLQ-BM22 scale and baseline sociodemographic and medical factors. Natural log-transformation was applied for each of the EORTC QLQ-BM22 scales to normalize the distribution. The independent variables were the eight demographic and medical covariates. After univariate analysis, a backward selection procedure was used to select the most significant factors related to each EORTC QLQ-BM22 scale for inclusion in the multivariate analysis. To adjust for multiple comparisons, Bonferroni-adjusted p-value of <0.006 (0.05/8 demographic covariates) was considered statistically significant. For PS and PC subscales, a negative correlation coefficient indicates that lower PS or PC score is positively related to patients with age>60, KPS>80, of the female gender, married, higher level of education, living with others, and having a primary breast histology. Conversely, a positive correlation coefficient indicates that higher FI and PA score is positively related to patients with age>60, KPS>80, of the female gender, married, higher level of education, living with others, and having a primary breast histology.

For the comparison of baseline to follow-up EORTC QLQ-BM22 results, changes were calculated by subtracting the follow-up score from the baseline score, and the Wilcoxon signed rank test was performed. Furthermore, to compare the change in EORTC QLQ-BM22 scores for different sociodemographic and medical factors, the Wilcoxon signed-rank test (for binary) or the Kruskal-Wallis nonparametric test (for categorical factors) was utilized, with a statistically significant adjusted p-value of 0.006. All analyses were based on the EORTC QLQ manual and conducted by statistical software SAS (SAS version 9.3 for Windows; SAS Institute Inc., Cary, NC).¹⁴

Results

Baseline

Table 1 shows the demographics of the 396 patients. Univariate ANOVA results are in Table 2. For the PS scale, two significant factors were found: KPS>80 (p<0.0001) and primary histology (p=0.0002). KPS>80 was predictive of a better PS score, with a median PS score of 20.0 in patients with KPS>80 and 33.3 in patients with KPS≤80. Breast as a primary cancer site was significantly more predictive of a better PS score than lung or other. The median PS score was 20.0 in patients with primary breast cancer as compared to 33.3 in lung cancer and 33.3 in other. On multivariate analysis, however, only KPS>80 remained significant (p<0.0001).

Table 1.

Baseline Demographics

	(N=396)
Age (years)
n	396
Mean (SD)	58.0 (13.1)
Median (range)	57.0 (26–95)
> 60 years	152 (38.38%)
≤ 60 years	244 (61.62%)
KPS
n	395
Mean (SD)	83.1 (16.8)
Median (range)	80 (30–100)
> 80	179 (45.32%)
≤ 80	216 (54.68%)
Gender
Female	301 (76.01%)
Male	95 (23.99%)
Education level
Elementary school	25 (14.45%)
High school	66 (38.15%)
University	44 (25.43%)
Masters or PhD	12 (6.94%)
Other	26 (15.03%)
Employment status
Retired	89 (50.86%)
Employed	47 (26.86%)
Unemployed	39 (22.28%)
Cohabitants
Spouse	79 (45.14%)
Spouse and child(ren)	38 (21.71%)
Alone	27 (15.43%)
Child(ren)	18 (10.29%)
Other	13 (7.43%)
Married
Yes	135 (77.59%)
No	39 (22.41%)
Primary cancer site
Breast	271 (68.43%)
Prostate	45 (11.36%)
Lung	30 (7.58%)
Gastrointestinal	10 (2.53%)
Renal cell	6 (1.52%)
Other	34 (8.59%)

Table 2.

Univariate Analysis

	Painful scale (PS)			Painful characteristics (PC)			Functional interference (FI)			Psychosocial aspects (PA)
Demographic covariates	CE	SE	p-value	CE	SE	p-value	CE	SE	p-value	CE	SE	p-value
Age>60 years	−0.172	0.118	0.1435	0.014	0.143	0.9228	0.097	0.103	0.3456	0.116	0.050	0.0200
KPS>80	−0.747	0.109	<0.0001	−1.196	0.126	<0.0001	0.698	0.095	<0.0001	0.221	0.048	<0.0001
Gender	−0.322	0.133	0.0162	−0.682	0.159	<.0001	0.393	0.116	0.0008	0.001	0.057	0.9816
Married	−0.126	0.169	0.4587	−0.077	0.171	0.6515	0.103	0.217	0.6363	−0.034	0.100	0.7358
Education			0.3610			0.3897			0.2979			0.8979
Elementary vs. university	0.114	0.233	0.6242	0.303	0.236	0.2001	−0.460	0.298	0.1254	−0.003	0.139	0.9800
High school vs. university	−0.153	0.181	0.3987	−0.079	0.183	0.6668	0.015	0.232	0.9468	0.073	0.108	0.4990
Masters or PhD vs. university	−0.388	0.303	0.2031	−0.280	0.307	0.3621	0.350	0.388	0.3678	0.105	0.181	0.5621
Other vs. university	0.140	0.230	0.5441	−0.043	0.233	0.8523	−0.195	0.295	0.5100	0.105	0.137	0.4471
Employment status			0.1559			0.0990			0.0424			0.1363
Employed vs. unemployed	−0.279	0.200	0.1646	−0.251	0.201	0.2136	0.513	0.254	0.0453	0.042	0.118	0.7192
Retired vs. unemployed	−0.341	0.177	0.0566	−0.386	0.179	0.0320	0.554	0.225	0.0151	0.186	0.104	0.0772
Cohabitant			0.2827			0.5768			0.7233			0.8229
Alone vs. spouse and child(ren)	−0.224	0.233	0.3367	−0.197	0.237	0.4067	0.096	0.301	0.7514	0.117	0.139	0.4001
Child(ren) vs. spouse and child(ren)	0.183	0.265	0.4896	−0.071	0.269	0.7932	−0.270	0.342	0.4307	0.030	0.158	0.8477
Spouse vs. spouse and child(ren)	−0.281	0.183	0.1262	−0.282	0.186	0.1299	0.159	0.236	0.5020	0.121	0.109	0.2691
Other vs. spouse and child(ren)	−0.131	0.297	0.6598	−0.330	0.302	0.2765	0.130	0.384	0.7355	0.105	0.177	0.5521
Primary cancer site			0.0002			<0.0001			<0.0001			<0.0001
Prostate vs. breast	0.310	0.179	0.0839	0.601	0.213	0.0050	−0.187	0.154	0.2239	0.145	0.076	0.0551
Lung vs. breast	0.746	0.214	0.0005	0.966	0.255	0.0002	−0.884	0.184	<0.0001	−0.021	0.090	0.8166
Gastrointestinal vs. breast	0.740	0.358	0.0396	1.343	0.426	0.0018	−1.068	0.308	0.0006	−0.352	0.151	0.0202
Renal cell vs. breast	−0.034	0.459	0.9412	0.293	0.546	0.5925	−0.100	0.395	0.7999	−0.660	0.194	0.0007
Others vs. breast	0.668	0.202	0.0010	0.996	0.241	<0.0001	−0.578	0.174	0.0010	−0.242	0.085	0.0049

Bold indicates statistically significant adjusted p value (<0.006).

For the binary variables: age>60 (1-Y, 0-N); KPS>80 (1-Y, 0-N); gender (1-F, 0-M); married (1-Y, 0-N).

CE, coefficient; SE, standard error.

For the PC scale, KPS>80 (p<0.0001), female gender (p<0.0001), and primary cancer site (p<0.0001) were significant on univariate analysis. KPS>80 was predictive of a better PC score, with a median PC score of 16.7 in patients with KPS>80 and 44.4 in patients with KPS≤80. Female gender was predictive of a better PC score, with a median score of 22.2 in comparison to 44.4 in males. Breast as a primary cancer site was predictive of a better PC score, with a median score of 22.2 versus lung (55.6), gastrointestinal (61.1), prostate (44.4), and other (44.4). Significant difference was not found between breast and renal cell primary (27.8). On multivariate analysis, only KPS>80 remained significant (p<0.0001).

For the FI scale, three significant factors were found on univariate analysis: KPS>80 (p<0.0001), female gender (p=0.0008), and primary cancer site (p<0.0001). KPS>80 was predictive of a better FI score, with a median FI score of 79.2 in patients with KPS>80 and 50.0 in patients with KPS≤80. Female gender was predictive of a better FI score, with a median score of 66.7 in comparison to 54.2 in males. Breast as a primary cancer site was predictive of a better FI score (median 70.8) than lung (45.8), gastrointestinal (35.4), or other (41.7). Only KPS>80 (p<0.0001) and primary histology (p=0.0031) remained significantly related on multivariate analysis. No significant interaction between KPS and primary histology was found.

For the PA scale, two significant factors were found on univariate analysis: KPS>80 (p<0.0001) and primary cancer site (p<0.0001). KPS>80 was predictive of a better PA score, with a median PA score of 61 in patients with KPS>80 and 50 in patients with KPS≤80. Breast as a primary cancer site was more predictive of a better PA score (median 61.1) than renal (50.0) or other primary sites (50.0). On multivariate analysis, both factors remained significantly related to PA score. No significant interaction between KPS and primary histology was found. The multivariate analysis also found that patients with primary prostate cancer were more likely to score higher on the PA scale (median 66.7) compared to patients with primary breast cancer (61.1).

Changes between baseline and follow-up

Of the 396 patients recruited initially, 346 patients completed the follow-up EORTC QLQ-BM22. Overall, three EORTC QLQ-BM22 scales showed significant changes from baseline: PS, PC, and FI. Patients demonstrated significant improvement on the PS scale (median of 27 at baseline and 20 at follow-up, p=0.0023); PC scale (median of 33 at baseline and 22 at follow-up, p<0.0001); and FI scale (median of 67 at baseline increased to 71 at follow up, p=0.0002).

For the PS and PC scale, KPS and primary histology also correlated with specific changes with time; however no such correlation was found for the FI scale (see Table 3). On the PS scale, patients with KPS>80 had a median change score of zero, but those with KPS≤80 had a median change score of −6.7, which indicated that patients with poorer performance status experienced a greater improvement in HRQOL. Similarly, on the PC scale, patients with KPS>80 had an average change score of −2, while those with lower performance status had an average change score of −10, signifying that patients with KPS≤80 also experienced greater improvement in HRQOL.

Table 3.

Changes in the Four BM22 Scale Scores Over Time

Change in score (follow-up—baseline) scales	Demographic covariates	p-value
Painful site (PS) (N=346)	Age>60 years	0.6993
	KPS>80	0.0037
	Gender	0.0758
	Married	0.1103
	Education	0.5615
	Employment status	0.8422
	Cohabitant	0.1869
	Primary cancer site	0.1360
Painful characteristic (PC) (N=345)	Age>60 years	0.7659
	KPS>80	0.0036
	Gender	0.0102
	Married	0.2939
	Education	0.7856
	Employment status	0.6089
	Cohabitant	0.2822
	Primary cancer site	0.0038
Functional interference (FI) (N=346)	Age>60 years	0.9173
	KPS>80	0.0068
	Gender	0.0246
	Married	0.1833
	Education	0.8734
	Employment status	0.6714
	Cohabitant	0.3173
	Primary cancer site	0.0621
Psychosocial aspects (PA) (N=345)	Age>60 years	0.8371
	KPS>80	0.9458
	Gender	0.3000
	Married	0.6627
	Education	0.5300
	Employment status	0.0682
	Cohabitant	0.6737
	Primary cancer site	0.4034

Bold indicates statistically significant adjusted p value (<0.006).

For the binary variables: age>60 (1-Y, 0-N); KPS>80 (1-Y, 0-N); gender (1-F, 0-M); married (1-Y, 0-N).

Patients with different primary cancer sites also had significantly different PC changes from baseline (see Table 3). Patients with primary breast or renal cell carcinoma had median change scores of zero, those with gastrointestinal cancers had the highest decrease with a median change of −22, primary lung or prostate cancer patients had a median change of −11, and those with other cancer sites had a median change of −6. These decreases in the PC scale scores correspond to improved HRQOL.

Discussion

This study is the first to examine the relationship between sociodemographic and medical factors and HRQOL as determined by the EORTC QLQ-BM22 in cancer patients with bone metastases. We report that the greatest determinants of HRQOL in our cohort are Karnofsky performance status and primary histology. Zimmerman and colleagues, using the FACT-G on a general advanced cancer population, also found that performance status was a significant determinant of HRQOL.⁷ Cassie and colleagues, utilizing the EORTC QLQ-C15-PAL, described significant associations between KPS and HRQOL in patients with lung, brain, and bone metastases from any primary histology.¹⁵ In a study in which patients and health care professionals described issues thought to affect HRQOL in the context of bone metastases, items such as ‘able to perform self-care’ and ‘able to perform role functioning’ were important for better HRQOL.² These two items are part of the description of KPS, which further confirms its importance in HRQOL.¹⁶

We also found certain primary histologies correlated with better HRQOL (certain scales). Patients with primary breast cancer were more likely to score higher on the FI scale of the EORTC QLQ-BM22 than patients with lung cancer. Those with prostate cancer were more likely to report better PA scores than patients with breast cancer. Gender difference may be influential in this finding, as it has been previously reported that females tend to approach and describe cancer in a more emotional and psychosocial way. A study of 700 patients was conducted to determine the perception and explanation of cancer, which found that males tend to use more neutral terminology such as ‘abnormal cells,’ while females tend to describe cancer using emotional terms such as ‘an eating away disease’ or ‘a dreaded disease.’¹⁷ However, the influence of primary histology in comparison to gender on HRQOL is still inconclusive in current literature.

Similar differences based on histology have been found in other studies in an advanced cancer population. Zimmerman and colleagues reported significantly worse emotional well-being when the primary cancer site was breast and gynecological as opposed to gastrointestinal, lung, or genitourinary, and worse physical and functional HRQOL in lung primary patients.⁷ Ludh and colleagues found no significant differences relating to primary cancer site, but found that lung cancer patients did consistently score the lowest on the functioning scales of the EORTC QLQ-C30.¹¹

In terms of demographic factors, gender was significant on univariate analysis in two of the four EORTC QLQ-BM22 scales, with female gender correlating with a better outcome on the PC and FI scales. The influence of gender on HRQOL may be confounded by primary cancer site. However, in support of the influence of gender, even within the general population there was a difference in HRQOL by gender as seen in Hjermstad and colleagues' study.¹⁸ Comparable conclusions were also reached by Zimmerman and colleagues and Lundh and colleagues.^7,11

In the four previous studies of sociodemographic and medical factors affecting HRQOL in advanced cancer patients, three determined that age was an important factor;^7,11,12 however, it was not found to be significant in the current study. Koo and colleagues reported age related differences in HRQOL, in advanced cancer patients, based on the EORTC QLQ-C15-PAL.¹⁹ Age difference in HRQOL was also demonstrated in newly diagnosed cancer patients and was relevant to treatment for elderly patients.^20,21 In the general population, older age has been correlated with worse HRQOL scores, opposite to what is seen in advanced cancer patients.^9,18,22 In this cancer population, older age may be associated with better HRQOL because of the increased traumatization of being diagnosed with advanced cancer in younger patients. Younger patients usually have greater family and community responsibilities; a diagnosis of advanced cancer usually comes unexpectedly, and for which they are unprepared.^7,23–25 Further exploration is recommended.

Other factors previously found to be determinants of HRQOL were either not investigated or did not reach significance in our study.^7,10–12 Zimmerman and colleagues found that the type of chemotherapy greatly influenced HRQOL mainly with respect to emotional and existential well-being.⁷ In contrast to our current study, another publication found that only spiritual well-being had a relationship with HRQOL in advanced cancer patients.¹² Contradictory findings on marital status have been reported. Lundh and colleagues found that being married may relate to lower HRQOL, while Jordhoy and colleagues found no influence from a live-in partner.^10,11

KPS and primary histology were also found to correlate significantly with degree of change with time. Of these two factors, patients with a lower KPS had a greater change in HRQOL scores from baseline. Patients with gastrointestinal, lung, and other primary sites also demonstrated greater changes in HRQOL scores in comparison to breast and renal cell. Since KPS<80 and having a primary histology of gastrointestinal, lung, and other sites were predictive of poorer HRQOL, the larger change can be attributed to having more room for the improvement of HRQOL.²⁶

Limitations of this study include the retrospective nature of the analysis and the 12.6% of patients lost at follow-up. A large majority of patients were female (76%) or had primary breast cancer (68.43%). A further limitation was the use of only the EORTC QLQ-BM22. The inclusion of the EORTC QLQ-C30 may have demonstrated additional significant relationships as seen in other studies. Finally, differences in HRQOL may arise due to the symptoms of bone metastases (which were not measured in the present study) masking the effects of other factors impacting on HRQOL.²⁷

Palliative care for patients with bone metastases encompasses not only the pain and symptom management but also a combination of spiritual care, social support, and psychological care to maximize the improvements to quality of life.¹⁹ As such, the findings from this analysis will aid health care professionals, especially palliative care professionals, in identifying patients who are more likely to experience a lower HRQOL to encourage earlier and more efficient use of appropriate resources such as psychological or social support.

Recommendations

Karnofsky performance status>80 and primary histology of breast and prostate positively related to HRQOL in patients with bone metastases as measured by the EORTC QLQ-BM22. In addition, gender affected the experience of HRQOL, with females more often reporting better HRQOL than males. Further study in this area with factors such as treatment and spiritual well-being is recommended to identify other relevant factors, which may help health care professionals recognize vulnerable individuals and offer other necessary social supports and resources.

Footnotes

Acknowledgments

We acknowledge the generous support of the Bratty Family Fund, the Michael and Karyn Goldstein Cancer Research Fund, the Joseph and Silvana Melara Cancer Research Fund, and the Ofelia Cancer Research Fund.

Author Disclosure Statement

No conflicting financial interests exist.

Appendix 1. EORTC QLQ-BM22

Patients sometimes report that they have the following symptoms or problems. Please indicate the extent to which you have experienced these symptoms or problems during the past week. Please answer by circling the number that best applies to you.

References

Movsas

. Quality of life in oncology trials: A clinical guide. Semin Radiat Oncol, 2003; 13,3:235–247.

Harris

, Chow

, Zhang

, Velikova

, Bezjak

, Wu

et al. Patients' and health care professionals' evaluation of health-related quality of life issues in bone metastases. Eur J Cancer, 2009; 45,14:2510–2518.

Culleton

, Kwok

, Chow

. Radiotherapy for pain. Clin Oncol (R Coll Radiol), 2011; 23,6:399–406.

Giesinger

, Wintner

, Oberguggenberger

, Gamper

, Fiegl

, Denz

et al. Quality of life trajectory in patients with advanced cancer during the last year of life. J Palliat Med, 2011; 14,8:904–912.

Chow

, Hird

, Velikova

, Johnson

, Dewolf

, Bezjak

et al. The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire for patients with bone metastases: The EORTC QLQ-BM22. Eur J Cancer, 2009; 45,7:1146–1152.

Kaasa

, Loge

. Quality of life in palliative care: Principles and practice. Palliat Med, 2003; 17,1:11–20.

Zimmermann

, Burman

, Swami

, Krzyzanowska

, Leighl

, Moore

et al. Determinants of quality of life in patients with advanced cancer. Support Care Cancer, 2011; 19,5:621–629.

Ferrans

. Advances in measuring quality-of-life outcomes in cancer care. Semin Oncol Nurs, 2010; 26,1:2–11.

Holzner

, Kemmler

, Cella

, De Paoli

, Meraner

, Kopp

et al. Normative data for functional assessment of cancer therapy: General scale and its use for the interpretation of quality of life scores in cancer survivors. Acta Oncol, 2004; 43,2:153–160.

10.

Jordhoy

, Fayers

, Loge

, Saltnes

, Ahlner-Elmqvist

, Kaasa

. Quality of life in advanced cancer patients: The impact of sociodemographic and medical characteristics. Br J Cancer, 2001; 85,10:1478–1485.

11.

Lundh Hagelin

, Seiger

, Furst

. Quality of life in terminal care—with special reference to age, gender and marital status. Support Care Cancer, 2006; 14,4:320–328.

12.

Fisch

, Titzer

, Kristeller

, Shen

, Loehrer

, Jung

et al. Assessment of quality of life in outpatients with advanced cancer: The accuracy of clinician estimations and the relevance of spiritual well-being—a Hoosier Oncology Group Study. J Clin Oncol, 2003; 21,14:2754–2759.

13.

Chow

, Nguyen

, Zhang

, Tseng

, Hou

, Fairchild

et al. International field testing of the reliability and validity of the EORTC QLQ-BM22 module to assess health-related quality of life in patients with bone metastases. Cancer, 2012; 118,5:1457–1465.

14.

Fayers

, Aaronson

, Bjordal

, Groenvold

, Curran

, Bottomley

. EORTC QLQ-C30 Scoring Manual, 3rd. Brussels: EORTC Publications, 2001.

15.

Caissie

, Culleton

, Nguyen

, Zhang

, Zeng

, Holden

et al. EORTC QLQ-C15-PAL quality of life scores in patients with advanced cancer referred for palliative radiotherapy. Support Care Cancer, 2012; 20,4:841–848.

16.

Yates

, Chalmer

, McKegney

. Evaluation of patients with advanced cancer using the Karnofsky performance status. Cancer, 1980; 45,8:2220–2224.

17.

Murray

, McMillan

. Gender differences in perceptions of cancer. J Cancer Educ, 1993; 8,1:53–62.

18.

Hjermstad

, Fayers

, Bjordal

, Kaasa

. Using reference data on quality of life: The importance of adjusting for age and gender, exemplified by the EORTC QLQ-C30 (+3) Eur J Cancer, 1998; 34,9:1381–1389.

19.

Koo

, Zeng

, Chen

, Zhang

, Culleton

, Dennis

et al.

Do elderly patients with metastatic cancer have worse quality of life scores?

Support Care Cancer, 2012; 20,9:2121–2127.

20.

Rustoen

, Moum

, Wiklund

, Hanestad

. Quality of life in newly diagnosed cancer patients. J Adv Nurs, 1999; 29,2:490–498.

21.

Pasetto

, Falci

, Compostella

, Sinigaglia

, Rossi

, Monfardini

. Quality of life in elderly cancer patients. Eur J Cancer, 2007; 43,10:1508–1513.

22.

Michelson

, Bolund

, Nilsson

, Brandberg

. Health-related quality of life measured by the EORTC QLQ-C30: Reference values from a large sample of Swedish population. Acta Oncol, 2000; 39,4:477–484.

23.

Mor

, Allen

, Malin

. The psychosocial impact of cancer on older versus younger patients and their families. Cancer, 1994; 74,S7:2118–2127.

24.

Exley

, Letherby

. Managing a disrupted lifecourse: Issues of identity and emotion work. Health, 2001; 5,1:112–132.

25.

Wan

, Counte

, Cella

. The influence of personal expectations on cancer patients' reports of health-related quality of life. Psychooncology, 1997; 6,1:1–11.

26.

Cella

, Hahn

, Dineen

. Meaningful change in cancer-specific quality of life scores: Differences between improvement and worsening. Qual Life Res, 2002; 11,3:207–221.

27.

Guyatt

, Feeny

, Patrick

. Measuring health-related quality of life. Ann Intern Med, 1993; 118,8:622–629.