A brief instrument to measure health-related quality-of-life in patients with bone metastasis: validation of the German version of Bone Metastases Quality-of-Life-10 (BOMET-QoL-10)

Abstract

Aims: This prospective, epidemiologic study was designed to translate the original Spanish Bone Metastases Quality-of-Life-10 (BOMET-QoL-10) questionnaire and undertake a validation of the translated German version of BOMET-QoL-10 in Germany to assess health-related quality-of-life (HRQoL) in patients with bone metastases (BM).

Methods: The translation process included forward and backward translations, and a linguistic validation. Patients aged ≥18 years with histological confirmation of cancer, diagnosed with BM, life expectancy ≥6 months, and fluency in German were eligible for this study (enrolled consecutively in 33 outpatient centers in Germany). Patients were given the German version of BOMET-QoL-10, together with the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire QLQ-C30 and EORTC QLQ-BM22 questionnaires at inclusion, 6 weeks, 3 months, and 6 months after inclusion. A debriefing questionnaire was administered at inclusion to determine patient acceptability and understanding.

Results: Data include 364 patients with BM (median age = 68 years; females = 71.7%). The BOMET-QoL-10 is brief and clear (median completion time = 5 minutes; >90% of patients completed the questionnaire without assistance). The BOMET-QoL-10 forms only one overall scale. All 10 items showed a substantial correlation with the first factor (factor loading, range = 0.58–0.86). BOMET-QoL-10 exhibits high internal consistency and reproducibility (Cronbach’s alpha = 0.91; intra-class correlation coefficient = 0.76). BOMET-QoL-10 showed significant correlations (range = 0.69–0.79) both with EORTC QLQ-C30 and EORTC QLQ-BM22 within the functioning (physical, social, interference) and symptom (fatigue, pain) scales, displayed significant sensitivity to change in EORTC QLQ-BM22 scores, and proved the potential ability to detect change in HRQoL in patients with different disease status.

Limitations: There was a high proportion of females in this study, which might represent a limitation.

Conclusions: The German version of BOMET-QoL-10 is a valid, reliable, brief, and clear instrument able to measure HRQoL in patients with BM.

Keywords:

• German version of BOMET-QoL-10
• bone metastases
• EORTC QLQ-BM22
• EORTC OLQ-C30
• health-related quality-of-life
• prospective epidemiologic study

JEL classification codes:

• I00
• I31

Introduction

Approximately 480,000 individuals are diagnosed with cancer each year in Germany, with breast (30.8%) and prostate (25.3%) cancers accounting for the most frequent new cases in women and men, respectively¹. Bone is the most common site for metastasis in a wide range of malignancies, resulting in significant morbidity and portending poor survival². The exact incidence of bone metastases (BM) is unknown, although BM are commonly found in cancer patients at the time of post-mortem examination³. Breast and prostate carcinomas are the most common tumors to metastasize to bone^4,5. In addition, thyroid, renal cell, and lung carcinomas have also been found to metastasize to bone⁴. In contrast, gastrointestinal carcinoma seldom produces BM⁶. Metastatic bone disease in patients with advanced cancer is frequently associated with skeletal-related events (SREs). These can be debilitating, causing pain and impaired functioning, resulting in a decreased quality-of-life (QoL). The main aim of management of BM (including bone-targeting therapy) and their according consequences is to alleviate the pain and reduce the risk of SREs having the potential to improve QoL^7,8. QoL, by its very nature, is idiosyncratic to the individual, but intuitively meaningful and understandable to most people⁹. Health-related quality-of-life (HRQoL) measures have become a vital and often required part of health outcomes appraisal. HRQoL is a subjective, multidimensional construct reflecting functional status, psychosocial well-being, and disease-related and treatment-related symptoms¹⁰. In patients with BM, bone pain can be used as a surrogate measure of HRQoL^11,12. Two well established, validated questionnaires are available which are able to detect HRQoL in patients with BM from a patient’s point of view^13–15. These questionnaires comprise a basic questionnaire regarding HRQoL (European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire QLQ-C30, Functional Assessment of Cancer Therapy - General (FACT-G)) and a corresponding bone-specific module (EORTC QLQ-BM22, FACT-Bone Pain)^13–15. However, patients would need to complete 52 (EORTC) and 43 (FACT) items, respectively, which could be burdensome to the patient. Another HRQoL questionnaire, the Bone Metastases Quality-of-Life-10 (BOMET-QoL-10) instrument, was developed to provide a short, brief questionnaire that could be used effortlessly in routine clinical practice^16,17. This instrument was validated in Spain and comprises 10 questions specifically addressing patients with BM. BOMET-QoL-10 is aimed at identifying characteristic bone pain arising from metastases (intensity and localization, 4 items) and general aspects of QoL (6 items) as to enable physicians to gain an understanding of patients’ QoL, and thus to optimize treatment¹¹. There is an unmet need of having the BOMET-QoL-10 questionnaire validated in other countries and in the local language, which would unquestionably be beneficial to the patient. The BOMET-QoL-10 instrument has not been validated in Germany; therefore, the main objective of our study was to translate the original Spanish BOMET-QoL-10 questionnaire into German and undertake a validation of the translated German version of BOMET-QoL-10 in patients with BM in Germany. Secondary objectives were to assess the association of patient characteristics with the outcome of BOMET-QoL-10, to evaluate the sensitivity of BOMET-QoL-10 to change in disease status, and to explore the effect of the sequence in which the questionnaires were filled in (questionnaires sequence effect on the outcome of BOMET-QoL-10).

Methods

Patients

From December 2014 to June 2015, 370 patients diagnosed with BM from breast, kidney, lung, or prostate carcinoma were enrolled in 33 participating Medical Oncology and Medical Uro-Oncology outpatient centers in Germany (Supplementary Table S1).

Patients were enrolled consecutively in the aforementioned outpatient centers if they were aged ≥18 years, had histological confirmation of cancer (breast, lung, prostate, or renal cell carcinoma), were diagnosed with BM, had a life expectancy ≥6 months, and were fluent in German (reading, written, and spoken). Enrolment was possible at any stage of metastatic disease or therapy. Patients were excluded for any of the following reasons: absence of BM, participating in another study involving questionnaires, or unable to comply with study requirements (completion of questionnaires).

Study design

This was a prospective, multi-center, epidemiologic study (ClinicalTrials.gov NCT 02293642).

Study objectives

The primary objectives were to translate the original Spanish BOMET-QoL-10 questionnaire^16,17 and undertake a validation of the translated German version of BOMET-QoL-10 in Germany, and to evaluate its sensitivity to change in EORTC QLQ-BM22 scores. Secondary objectives were to assess the association of patient characteristics with the outcome of BOMET-QoL-10, to evaluate the sensitivity of BOMET-QoL-10 to change in disease status, and to explore the effect of the sequence in which the questionnaires were filled in (questionnaires sequence effect on the outcome of BOMET-QoL-10).

Study procedure

Patients were given the German version of the BOMET-QoL-10 together with EORTC QLQ-C30 and EORTC QLQ-BM22 questionnaires, and a debriefing questionnaire (DBQ; used to determine patient acceptability and understanding)¹⁸. The patients were also asked in the DBQ for the time needed to complete the questionnaire and whether they had encountered any difficulties in filling out the questionnaire (assistance needed). The four different questionnaires were put together into one questionnaire form administered to the patients. In order to assess a possible questionnaires sequence effect, patients received either EORTC QLQ-C30 and EORTC QLQ-BM22 first, followed by the DBQ, and the BOMET-QoL-10 last (version 1), or the BOMET-QoL-10 first, followed by the DBQ, and the EORTC QLQ-C30 and EORTC QLQ-BM22 last (version 2). The DBQ was used at inclusion only. The sequence of the four questionnaires in the form administered to the patients is depicted in Figure 1.

The first questionnaire form was administered by the study site at inclusion (T0). Every patient received the same version of questionnaire form throughout the course of the study. Subsequent questionnaire forms were administered through post mail and followed up (delivery status) by iOMEDICO Site Management Organization at the following time points: 6 weeks (T1), 3 months (T2), and 6 months after inclusion (T3). Each questionnaire form was dated by the patient upon completion and returned to the iOMEDICO AG by post in a provided return envelope.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The project was approved by the German local ethics committee “Ethikkommission der Landesärztekammer Baden-Württemberg” on October 21, 2014.

Figure 1. Sequence of the four questionnaires in the form administered to the patients. *DBQ was only used at baseline (T0).

Informed consent

Written informed consent was obtained from all individual participants included in the study.

Study variables

The following data were captured at inclusion as available from patients’ medical records: demographic parameters, date of diagnosis of metastatic disease, date of diagnosis of BM, comorbidities, Eastern Cooperative Oncology Group (ECOG) performance status, tumor entity, and current status of metastatic disease (disease in general, BM in particular). Six months after inclusion or upon premature study termination, the following data were captured: end of documentation, current status of metastatic disease, and date(s) of progression of metastatic disease during the observational period. All study-related data were transferred to a validated, secure, web-based electronic data capture system (iOstudy office edc), operating as per the guidelines of FDA 21 CFR Part 11.

Translation of the original Spanish BOMET-QoL-10 questionnaire

The original Spanish BOMET-QoL-10 instrument was translated into German using a Spanish translation agency specialized in translation of QoL questionnaires in accordance with international methodology of linguistic validation of QoL questionnaires (Proyectos y Traducciones (PYT), Barcelona, Spain). The translation process included a forward translation, a backward translation, and a linguistic validation. Two independent forward translations were carried out by two professional, bicultural translators who were native in German and fluent in Spanish. Both translations were carefully reviewed by a third independent translator, together with the translation agency, to issue the first intermediary German version. An independent backward translation was then carried out by a professional bicultural translator native in Spanish and fluent in German. This translation was carefully reviewed by the project manager and compared to the original version. Any revision needed was carried out prior to issuing the second intermediary target version. Next, a cognitive debriefing step was performed in which the second translated version was pilot-tested in Germany by five German-speaking patients with BM to assess the level of comprehensibility of the translated wording. The feedback from these five patients was reviewed by the translators to assess as to whether any further changes were needed prior to issuing the third intermediary target version. The final step included careful proofreading of the third intermediary target version, followed by issuing of the final, validated, official German version in February 2014. The German version of the BOMET-QoL-10 questionnaire, with its 10 items and 5-point Likert-type response scale (ranging from point 1 to point 5 with point 5 representing the most positive aspect), is presented in Table 1.

Table 1. German version of the BOMET-QoL-10 questionnaire.

BOMET-QoL-10 questionnaire items in German^a	[1] Immer^b	[2] Fast immer^c	[3] Manchmal^d	[4] Selten^e	[5] Nie^f	Factor loading	Cronbach alpha*
1. Ich fühle mich müde.	□	□	□	□	□	0.69	0.91
2. Das Aufstehen aus dem Bett fällt mir sehr schwer.	□	□	□	□	□	0.75	0.90
3. Ich fühle mich allgemein unwohl.	□	□	□	□	□	0.76	0.90
4. Ich fühle mich niedergeschlagen und möchte am liebsten weinen.	□	□	□	□	□	0.58	0.91
5. Ich habe wenig Lust, aus dem Haus zu gehen.	□	□	□	□	□	0.67	0.91
6. Ich vermeide es, gemeinsam mit meiner Familie etwas zu unternehmen.	□	□	□	□	□	0.68	0.91
7. Aufgrund meiner Krankheit habe ich Schmerzen in einigen Körperbereichen wie im Rücken, in den Beinen, in der Hüfte …, die mein Leben beeinträchtigen.	□	□	□	□	□	0.80	0.90
8. Aufgrund meiner Krankheit habe ich Schmerzen, die mein Leben beeinträchtigen.	□	□	□	□	□	0.86	0.90
9. Aufgrund meiner Krankheit habe ich starke Schmerzen, die mich nicht in Ruhe lassen.	□	□	□	□	□	0.85	0.90
10. Aufgrund meiner krankheitsbedingten Schmerzen kann ich das Leben nicht so genieszen wie früher.	□	□	□	□	□	0.82	0.90

^aThe 10 items and the original Likert-type response scale (ranging from point 0 to point 4) in (non-validated) English language are to be found in the original BOMET-QoL-10 publication¹⁷; ^bImmer = Always; ^cFast immer = Almost always; ^dManchmal = Sometimes; ^eSelten = Seldom; ^fNie = Never.

*Cronbach alpha standardized, item deleted, as described in the Results section.

Statistical analysis

The sample size was based on the “rule of thumb” recommendation: for multivariate analysis techniques, the number of observations should be 10-times the number of variables in the model to obtain a reliable estimate¹⁹. Therefore, a minimum sample size of 330 patients was required. Considering an estimated return rate of 85–90% for baseline questionnaires, a total of 370 patients were accrued. All variables were analyzed in a descriptive manner. For continuous variables, the number of observations, mean, standard deviation, 95% confidence interval of the mean, median, 25% and 75% quartile (interquartile range (IQR)), minimum, and maximum were computed. Categorical variables were presented with absolute values and relative frequencies within the single categories. A t-test was performed to evaluate a possible questionnaires sequence effect on the BOMET-QoL-10 score. All statistical analyses were performed according to a pre-specified statistical analysis plan using SAS™ Version 9.4.

BOMET-QoL-10 transformed score

The 10 items of the BOMET-QoL-10 were computed to a global score. The original global score, which ranged from 0 (worst QoL) to 40 (best QoL), was transformed to a standardized scale (transformed score) ranging from 0 (worst QoL) to 100 (best QoL)¹⁷. If not otherwise stated, the transformed score was used for statistical analyses.

Dimensionality

A principal component analysis was performed to investigate the dimensionality (eigenvalues, scree plot), so as to determine whether or not the items in the BOMET-QoL-10 formed only one overall scale (dimension)²⁰.

Internal consistency

Internal consistency is a measure of the extent to which items in a questionnaire (sub)scale are correlated (homogeneous); hence, measuring the same concept²¹. The Cronbach’s alpha coefficient was calculated to assess the internal consistency of the BOMET-QoL-10²². A Cronbach’s alpha value between 0.70–0.95 is expected for a questionnaire with homogeneous items^21,22.

Test–re-test reliability

Test–re-test reliability of the BOMET-QoL-10 was examined by assessing the scores from stable patients (patients with stable disease (SD) or patients in remission) at inclusion and 6 weeks after inclusion. Health stability of the patient was assessed by the treating physician and documented in the electronic data capture system used in this study. Pearson correlation and intra-class correlation (ICC) coefficients were computed. A correlation of ≥0.7 between test and re-test is recommended as a minimum standard of reliability²¹.

Construct validity

Construct validity refers to the degree to which a test measures what it claims^23,24. Pearson correlation coefficients were calculated between the BOMET-QoL-10 transformed score and the scores of EORTC QLQ-C30 (15 variables) and EORTC QLQ-BM22 (4 variables) for all four time points (T0, T1, T2, and T3). The EORTC QLQ-C30 variables included: global health status/QoL, physical functioning, role functioning, emotional functioning, cognitive functioning, social functioning, fatigue, nausea/vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial problems. The EORTC QLQ-BM22 variables included: painful sites, pain characteristics, functional interference, and psychosocial aspects.

Spearman correlation coefficients were calculated between ECOG performance status and the BOMET-QoL-10 transformed score and between ECOG performance status and the four EORTC QLQ-BM22 scores at inclusion.

Interpretation of correlation

Interpretation or assessing correlations are context-dependent²⁵. The guide presented in Table 2 was used for interpretation of correlation coefficients, as adopted from Mukaka²⁶.

Table 2. Interpretation of correlation coefficients.

Size of correlation	Interpretation
0.90–1.00 (−0.90 to −1.00)	Very high positive (negative) correlation
0.70–0.90 (−0.70 to −0.90)	High positive (negative) correlation
0.50–0.70 (−0.50 to −0.70)	Moderate positive (negative) correlation
0.30–0.50 (−0.30 to −0.50)	Low positive (negative) correlation
0.00–0.30 (0.00 to −0.30)	Negligible correlation

For correlations between variables from different domains such as QoL and ECOG performance status, the guide presented in Table 3 was used as adopted from Cohen²⁷.

Table 3. Effect size index (r) and respective value for small, medium, and large effects.

Effect size	r
Small	0.10
Medium	0.30
Large	0.50

Sensitivity to change and clinical validity

The sensitivity of the BOMET-QoL-10 to change in EORTC QLQ-BM22 scores was tested in patients with increased, stable (unchanged scores), and decreased EORTC QLQ-BM22 scores between the last available assessment and T0. An increased EORTC QLQ-BM22 score was defined as ≥20 points higher score than T0 and a decreased score was defined as ≥10 points lower score as compared to T0. This modified definition of minimal clinically important difference (MCID) in our study was based on a previous study evaluating MCID in the EORTC QLQ-BM22 and QLQ-C30 in patients with BM²⁸. For each of the four scores of the EORTC QLQ-BM22, patients were accordingly grouped into an EORTC QLQ-BM22-increased group, an EORTC QLQ-BM22-stable group, and an EORTC QLQ-BM22-decreased group, to which the change in BOMET-QoL-10 mean score was compared (same time points as used for the aforementioned grouping).

In order to analyze the association of tumor progression with the BOMET-QoL-10 outcome, the BOMET-QoL-10 score (T0) was compared in patients stratified according to disease status (SD, remission, progressive disease (PD)).

To explore the sensitivity of the BOMET-QoL-10 to change in tumor progression over time, the BOMET-QoL-10 difference score between T0 and respective subsequent time points (T1, T2, T3) was computed for each subsequent time point for the following three patient groups: (i) favorable & unchanged: patients in remission or with SD both at inclusion and any subsequent time point, (ii) unfavorable & unchanged: patients with PD either at inclusion or between inclusion and any subsequent time point, and (iii) favorable & changed: patients in remission or with SD at earlier time points with PD at subsequent time point. For patients with multiple tumor progressions during the observational period, only the first documented progression was considered and analyzed.

Univariate and multiple linear regression analyses

Univariate and multiple linear regression analyses (type 3 tests of fixed effects) were performed to assess the potential association of patient characteristics with BOMET-QoL-10 scores (T0). A p-value <.05 was considered significant, although, due to the exploratory nature of the analyses, all p-values must be interpreted in a descriptive manner.

Results

Patient characteristics

Out of the 370 enrolled patients, six patients were omitted from final statistical analysis; one patient presented with primary tumor entity other than study indication and five patients had not returned any questionnaire at all, resulting in a total of 364 evaluable patients, including 236 patients with breast cancer, 67 with prostate cancer, 46 with lung cancer, and 15 with renal cell cancer. Median age at enrolment was 68 years, and the proportion of female patients was 71.7%. Two thirds (65.4%) of the patients were documented as having SD, and 44.2% of the patient population were reported as having metastases confined to bone only. One or more than one comorbidity were reported in 66.2% of the patients. A Charlson comorbidity index²⁹ was calculated where a Charlson score of 0 was documented in 73.9% of the patients. An ECOG performance status of 0 or 1 was reported in 70% of the patients. The patient characteristics at inclusion are detailed in Table 4.

Table 4. Characteristics of patients at inclusion.

Characteristic	n (%)
Number of patients	364
Age
Mean, years	66.4
StD, years	10.4
Median, years	68
Min – Max, years	36–88
Sex, n (%)
Female sex	261 (71.7%)
Male sex	103 (28.3%)
BMI
Mean, kg/m^2	26.6
95% CI lower–upper, kg/m^2	26.1–27.1
StD, kg/m^2	4.8
Median, kg/m^2	26
Min–Max, kg/m^2	16–44
Tumor entity, n (%)
Breast	236 (64.8%)
Lung	46 (12.6%)
Renal	15 (4.1%)
Prostate	67 (18.4%)
Disease status, n (%)
In remission	45 (12.4%)
Stable	238 (65.4%)
Progressive	81 (22.3%)
Status of metastatic disease, n (%)
Bone only	161 (44.2%)
Additional distant sites	203 (55.8%)
Time from diagnosis of metastatic disease to study inclusion
Mean, years	2.9
95% CI lower–upper, years	2.6–3.3
StD, years	3.4
Median, years	2
Min–Max, years	0–21
Time from diagnosis of BM to study inclusion
Mean, years	2.7
95% CI lower–upper, years	2.4–3.0
StD, years	3.3
Median, years	2
Min–Max, years	0–21
Comorbidities, n (%)
≥1 comorbidities	241 (66.2%)
Most frequent comorbidities (>5%), n (%)
Hypertension	142 (39.0%)
Chronic pulmonary disease	23 (6.3%)
Renal disease	21 (5.8%)
Diabetes with end organ damage	21 (5.8%)
Heart failure	19 (5.2%)
Charlson score, n (%)
0	269 (73.9%)
1	50 (13.7%)
2	33 (9.1%)
≥3	12 (3.3%)
ECOG performance status, n (%)
0	114 (31.3%)
1	141 (38.7%)
2	27 (7.4%)
3	1 (0.3%)
4	0 (0.0%)
Missing data records	81 (22.3%)

Abbreviations. StD, Standard deviation; BMI, Body mass index; CI, Confidence interval of the mean.

Questionnaires return rate and comprehensibility

In total, 1,436 questionnaires were distributed to the enroled patients, of these, 1,228 (85.5%) were returned. Out of the returned questionnaires, 98.5% (n = 1,210) were evaluable, resulting in 349 questionnaires evaluable at T0, 316 at T1, 289 at T2, and 256 at T3. Version 1 and version 2 of the questionnaire were evenly distributed between participating patients (version 1: 171 patients (49.0%); version 2: 178 patients (51.0%)). The order (version 1 vs version 2) in which the questionnaire was filled in had no influence on the BOMET-QoL-10 score (p = .21). Median completion time reported by patients (n = 158) was 5 min (IQR = 3–15 min) and no assistance was required in 157 patients (90.8%), as assessed in patients having received the BOMET-QoL-10 version 2. In comparison, median completion time (n = 164) of the EORTC QLQ-C30 and EORTC QLQ-BM22 questionnaires was 11 min (IQR = 10–15 min), and 147 patients (89.6%) patients did not require any assistance, as evaluated in patients having received the version 1 questionnaire.

Transformed score of BOMET-QoL-10

The BOMET-QoL-10 mean transformed score ± StD was 61.3 ± 20.6 at T0, 58.9 ± 20.9 at T1, 58.7 ± 22.0 at T2, and 59.0 ± 20.3 at T3. Figure 2 presents the mean transformed score, together with (although, not directly comparable) a depiction of the EORTC QLQ-BM22 scores (four variables).

Figure 2. The BOMET-QoL-10 mean score and the scores of EORTC QLQ-BM22 (four different scales) depicted at inclusion, T1, T2, and T3.

Item analysis and validation of BOMET-QoL-10

In order to evaluate the psychometric properties of the BOMET-QoL-10, standard analyses of item structure, construct validity, and reliability were performed only for patients for whom all 10 items were answered to at inclusion (n = 309). A full range of responses to all items was observed, and none had any floor or ceiling effects. In general, the items were skewed toward better health. Three items were stronger skewed, with response rates of 31.7% to 44.7% (items 4, 6, and 8) in the category “never”.

Dimensionality

The data show that BOMET-QoL-10 forms only one overall scale (one dimension), as the first factor represented 56.5% of the variance of the items. All 10 items showed a substantial correlation with the first factor (factor loading, range = 0.58–0.86), as detailed in Table 1.

Internal consistency

A Cronbach’s coefficient alpha was calculated in order to assess the internal consistency of the BOMET-QoL-10. The BOMET-QoL-10 questionnaire was determined to be homogenous, as Cronbach’s coefficient alpha came to 0.91 (high internal consistency). Each of the 10 items provided high internal consistency, as Cronbach’s coefficient alpha remained between 0.90–0.91 when deleting one item at a time, as detailed in Table 1.

Test–re-test reliability

Test–re-test reliability of the BOMET-QoL-10 was examined by assessing the scores from patients in remission or with SD at T0 and T1 in all patients with evaluable questionnaires at both time points (n = 239). Pearson correlation and ICC coefficients were computed, which showed high test–re-test reliability considering the 6-week time span between test and re-test. The Pearson correlation coefficient came to 0.77 (p < .0001). The ICC coefficient was determined to be 0.76 when considering the time effect and 0.77 (95% CI = 0.71–0.81) without consideration of the time effect^30,31. Thus, the ICC was ≥0.7, which is recommended as a minimum standard for reliability²¹.

Construct validity

Pearson correlation coefficients were calculated between the BOMET-QoL-10 score and the scores of EORTC QLQ-C30 (15 variables) and EORTC QLQ-BM22 (four variables) at inclusion. Close to high positive correlations were observed between BOMET-QoL-10 and the EORTC QLQ-C30 physical functioning (r = 0.69; p < .0001) and social functioning (r = 0.69; p < .0001) scales. High positive correlation was found between BOMET-QoL-10 and the EORTC QLQ-C30 global health status (r = 0.75; p < .0001), whereas the correlation with emotional functioning (r = 0.64; p < .0001) and cognitive functioning (r = 0.51; p < .0001) scales were moderate. With regards to symptoms scales, high, negative correlations were found between BOMET-QoL-10 and the EORTC QLQ-C30 fatigue (r = −0.73; p < .0001) and pain (r = −0.73; p < .0001) scales. As for comparison with EORTC QLQ-BM22, a high positive correlation was observed between BOMET-QoL-10 and the EORTC QLQ-BM22 functional interference sub-scale (r = 0.79; p < .0001) and high and close to high, negative correlations were found with the EORTC QLQ-BM22 pain characteristics (r = −0.72; p < .0001) and painful sites (r = −0.69; p < .0001) sub-scales, respectively. The correlation between BOMET-QoL-10 and the EORTC QLQ-BM22 psychosocial aspects scale was found to be moderate (r = 0.53; p < .0001). Similar data were obtained for T1, T2, and T3, as displayed in Supplementary Table S2.

Spearman correlation coefficients were calculated between ECOG performance status and the BOMET-QoL-10 score and between ECOG performance status and the EORTC QLQ-BM22 scores at inclusion. The correlation between ECOG performance status and BOMET-QoL-10 was similar (r = −0.27; p < .0001) to the correlations between ECOG and the four sub-scores of the EORTC QLQ-BM22 (r range = 0.16 to −0.29).

Sensitivity of BOMET-QoL-10 to change in EORTC QLQ-BM22 scores

The ability of BOMET-QoL-10 to detect change was assessed in patient groups who showed increased, stable, or decreased EORTC QLQ-BM22 scores between the last available assessment and T0 (Figure 3). The BOMET-QoL-10 mean transformed scores increased in all four dimensions in the EORTC QLQ-BM22-increased group, and the increases ranged from 9.0 points (EORTC QLQ-BM22 psychosocial aspects) to 13.8 points (EORTC QLQ-BM22 functional interference). The BOMET-QoL-10 mean transformed scores decreased in all four dimensions in the EORTC QLQ-BM22-decreased group, and the decreases ranged from −8.4 points (EORTC QLQ-BM22 psychosocial aspects) to −11.9 points (EORTC QLQ-BM22 functional interference). Statistical trend tests showed significant effects of the change in all four EORTC QLQ-BM22 scores (p < .0001) on the change in the BOMET-QoL-10 score.

Figure 3. Sensitivity of BOMET-QoL-10 to change in EORTC QLQ-BM22 scores. Corresponding changes in BOMET-QoL-10 mean scores are shown for patient groups who showed increased, stable (unchanged scores), or decreased EORTC QLQ-BM22 scores between the last available assessment and T0. An increased EORTC QLQ-BM22 score was defined as a ≥ 20 points higher score than T0, and a decreased score was defined as a ≥10 points lower score as compared to T0.

Sensitivity of BOMET-QoL-10 to change in status of metastatic disease

The BOMET-QoL-10 score at T0 for patients documented with different disease status at inclusion was compared by summary statistics (n = 349). This analysis showed slight differences between groups (patients in remission or with SD: mean score at 63.1 ± 19.4; patients with PD with progressive BM: mean score at 55.5 ± 25.7; patients with progressive metastases at distant sites other than bone: mean score at 56.5 ± 19.2). In order to explore the sensitivity of BOMET-QoL-10 to change in disease status over time, the BOMET-QoL-10 difference scores between inclusion (T0) and respective later time points (T1, T2, T3) were calculated for each patient and each time point. Next, the mean difference in scores was compared between patients in remission or with SD (unchanged compared to previous time point), patients currently with PD (earlier in remission or with SD), and patients with PD (unchanged compared to previous time point) across respective time points. The BOMET-QoL-10 score showed a slight decrease both in patients in remission or with SD (unchanged) (mean values range = −1.5 to −2.8 points) and in patients with PD (unchanged) (mean values range = −1.0 to −2.7 points) during the course of this study, whereas the fall in scores was prominent in patients with a worsening in disease status across the three time points (mean values range = −7.1 to −13.4 points). Table 5 details the data (for graphic illustration, please see Supplementary Figure S1).

Table 5. Difference in BOMET-QoL-10 score compared to earlier time points.

Overall status of metastatic disease compared to earlier time points	Total n	Mean	StD	Min	Median	Max	95% CI (lower)	95% CI (upper)
Time point 1, n	304
In remission or SD (unchanged)	231	−2.5	13.5	−40	−2	50	−4.2	−0.7
Now PD (earlier in remission or SD)	8	−13.4	16.0	−35	−14	5	−26.8	0.0
PD (unchanged)	65	−1.0	12.9	−35	−3	38	−4.2	2.2
Unknown status	0	—	—	—	—	—	—	—
Time point 2, n	281
In remission or SD (unchanged)	199	−2.8	15.3	−70	−1	43	−5.0	−0.7
Now PD (earlier in remission or SD)	13	−7.1	16.1	−43	0	18	−16.8	2.7
PD (unchanged)	57	−2.7	13.5	−35	−3	33	−6.3	0.8
Unknown status	12	−5.8	15.1	−38	−4	18	−15.4	3.8
Time point 3, n	248
In remission or SD (unchanged)	142	−1.5	14.7	−38	−3	55	−4.0	0.9
Now PD (earlier in remission or SD)	27	−7.3	15.8	−50	−6	23	−13.5	−1.0
PD (unchanged)	59	−1.7	15.5	−35	−3	35	−5.7	2.4
Unknown status	20	−7.5	16.0	−45	−8	40	−15.0	0.0

A patient could fall into the group “Now PD (earlier in remission or SD)” only once, whereas a patient could fall into either of the other two groups multiple times.

Abbreviations. SD, stable disease; PD, progressive disease.

Association of demographic and clinical parameters with BOMET-QoL-10 score

The potential association of patient characteristics (T0) with the score of BOMET-QoL-10 was analyzed by univariate and multiple linear regression analysis. The univariate regression analysis revealed that only ECOG performance status exhibited a significant (p < .0001) association with the outcome of the BOMET-QoL-10 (272 patients with known ECOG status included in the analysis), whereas the other factors tested did not. Next, the multiple regression analysis showed that several factors independently exhibited a significant association with the BOMET-QoL-10 score, including (i) time from diagnosis of BM to study inclusion (p = .0009), (ii) time from diagnosis of metastatic disease to study inclusion (p = .0022), and (iii) ECOG performance status (p < .0001). Tumor entity and disease status were not found to be independently associated with the BOMET-QoL-10 score. As a limiting factor in this analysis, complete data was available for 269 patients only. The results of both univariate and multiple regression analyses are detailed in Table 6.

Table 6. Linear regression analysis: association of patient characteristics with BOMET-QoL-10 at inclusion.

Variable	Total n	Effect estimate^a	95% CI (lower)^a	95% CI (upper)^a	p-value
Univariate regression analysis
Age	348	0.018	−0.19	0.22	.87
Sex	348	−3.72	−8.55	1.10	.13
BMI	344	−0.14	−0.60	0.31	.54
Time from diagnosis of BM to enrolment	347	−0.0013	−0.0031	0.00045	.14
Time from diagnosis of metastatic disease to enrolment	348	−0.00019	−0.0019	0.0015	.83
Charlson score (score 0–score >0)	348	2.13	−2.78	7.05	.39
ECOG performance status	272	−9.13	−12.75	−5.51	<.0001*
Tumor entity overall	348				.26
Breast – Lung	345	−1.04	−7.72	5.65	.76
Breast – Prostate	345	−4.41	−10.11	1.30	.13
Breast – Renal	345	6.49	−4.67	17.66	.25
Lung – Prostate	345	−3.37	−11.29	4.55	.40
Lung – Renal	345	7.53	−4.91	19.97	.23
Prostate – Renal	345	10.90	−1.05	22.85	.07
Current disease status overall	348				.07
In remission – PD	346	4.56	−3.06	12.18	.24
In remission – SD	346	−1.62	−8.28	5.04	.63
PD – SD	346	−6.19	−11.49	−0.88	.02*
Current status of metastatic disease (BM only – other locations)	348	−0.43	−4.82	3.96	.85
Multiple regression analysis
Age	269	0.14	−0.10	0.38	.26
Sex	269	−0.48	−12.41	11.45	.94
BMI	269	−0.22	−0.71	0.27	.38
Time from diagnosis of BM to enrolment	269	−0.0073	−0.012	−0.0030	.0009*
Time from diagnosis of metastatic disease to enrolment	269	0.0063	0.0023	0.010	.0022*
Charlson score	269	−1.54	−7.09	4.008	.58
ECOG performance status	269	−10.05	−13.84	−6.26	<.0001*
Tumor entity overall	269				.80
Breast – Lung	269	−1.74	−11.42	7.95	.72
Breast – Prostate	269	−2.05	−15.47	11.36	.76
Breast – Renal	269	4.49	−10.61	19.58	.56
Lung – Prostate	269	−0.32	−11.58	10.94	.96
Lung – Renal	269	6.22	−8.10	20.55	.39
Prostate – Renal	269	6.54	−7.13	20.21	.35
Current disease status overall	269				.85
In remission – PD	269	1.69	−6.85	10.22	.70
In remission – SD	269	−0.28	−7.14	6.58	.94
PD – SD	269	−1.97	−8.68	4.75	.56
Current status of metastatic disease (BM only – other locations)	269	−2.18	−7.07	2.70	.38

*Significant results (p < .05).

^aEffect estimates refer to change in BOMET-QoL-10 for one unit increase in variable. Solution for fixed effects (age, BMI, time from diagnosis of BM to enrolment, time from diagnosis of metastatic disease to enrolment, ECOG) or differences of least squares means (sex, Charlson score, tumor entity, current disease status, current status of metastatic disease) was used to determine the effect estimate.

Abbreviations. BMI, Body mass index; BM, bone metastases; SD, stable disease; PD, progressive disease; CI, Confidence interval of the mean.

Discussion

For patients with advanced cancer, the treatment intention usually changes from curative to palliative in nature. Measurement of QoL provides a meaningful way to determine the impact of healthcare, especially when cure is not possible. Questionnaires for these patients should be brief and comprehensible so as not to burden the patient, while still including the most essential QoL aspects and maintaining the intrinsic measurement properties^32,33. The original BOMET-QoL-10 was developed for patients with BM and designed to be administered on its own. The longer BM-specific EORTC QLQ-BM22 module was designed to function as a supplement to EORTC QLQ-C30^{14–17,32,34,35}. Here, we successfully translated the original Spanish BOMET-QoL-10 into German (Table 1) and demonstrate that this instrument is a valid, reliable tool for patients with BM in Germany. Indeed, the data on internal consistency and test–re-test reliability confirm that the BOMET-QoL-10 exhibits both high homogeneity and reproducibility. Furthermore, significant, strong correlations with the EORTC QLQ-BM22 scores with the exception of psychosocial aspects support the construct validity of the BOMET-QoL-10. Unexpectedly, similar high correlations with EORTC QLQ-C30 global health and the EORTC QLQ-C30 symptom scales fatigue and pain were observed. Fatigue and pain are the symptoms most relevant for QoL scales. These traits of BOMET-QoL-10 are in line with the original validation study¹⁷. Empirical data about bone-related QoL questionnaires and their correlations with other constructs are very rare. At this point it was not possible to derive hypothesis driven by data. There was a more generic expectation that the correlation of BOMET-QoL-10 with EORTC QLQ-BM22 scores would be higher than with the more general EORTC QLQ-C30 scores, especially global health status. This expectation was not fulfilled. However, this result supports the intention of the BOMET-QoL-10 to measure QoL in patients with BM. EORTC QLQ-BM22 show much lower correlations with QLQ-C30 scores¹⁵, but we believe that the EORTC QLQ-BM22 was constructed in this direction. The univariate and multiple regression analyses revealed that ECOG performance status overall exhibited a significant association with the outcome of the BOMET-QoL-10 (Table 6), whereas neither sex nor age was a significant predictor of QoL. In the multiple regression analysis, time from diagnosis of metastatic disease/BM to enrolment were both significant predictors of the BOMET-QoL-10 outcome; this is probably due to the very high correlation between BOMET-QoL-10 and respective time-related variable. It needs to be stressed that the sample size between the univariate analyses and the multivariate analysis is quite different. It is also possible that the exclusion of patients without documented ECOG performance status (∼70 patients) may have a systematic effect on the results. Importantly, the BOMET-QoL-10 is brief and clear, as reflected by a short completion time, and >90% of the patients completing the questionnaire on their own. This feature of BOMET-QoL-10 may lessen the burden on patients while still providing valuable health-related information to the physician. The BOMET-QoL-10 was able to detect change in QoL during the course of the disease, as demonstrated by correlation to changed EORTC QLQ-BM22 scores and markedly reduced BOMET-QoL-10 scores (decreased QoL) in patients with worsened disease status over time (Table 5). The latter analysis corresponded to a secondary objective in our study; thus, the QoL scores were of limited comparability between patient groups, due to limited data and, consequently, the results are not interpretable any further. The BOMET-QoL-10 was validated in patients with several different tumor entities. The most common tumor entity was breast cancer, resulting in a high proportion of females in our study. This might represent a limitation to the study, although sex was not a significant predictor of QoL.

Conclusion

In conclusion, the German version of BOMET-QoL-10 is a valid, reliable, brief, and clear instrument able to measure HRQoL in patients with BM.

Transparency

Declaration of funding

This project received financial support from Amgen GmbH, Germany.

Declaration of financial/other relationships

CS received an institutional research grant from iOMEDICO (Freiburg, Germany) for statistical analysis. All other authors have disclosed that they have no significant relationships with or financial interests in any commercial companies related to this study or article. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Previous presentations

The data were partly presented at the symposium “Akademie Knochen und Krebs” in Munich, Germany on May 5–6, 2017.

Acknowledgments

The authors thank all patients, physicians and study teams participating in this study. The authors thank Dr Holger Hartmann and PD Dr Jens Hasskarl (former employees at iOMEDICO) for support during project design, Marian Carbonell (PYT Proyectos y Traducciones, Barcelona, Spain) for translation and linguistic validation of the BOMET-QoL-10), Angelika Gerlach (Clinical Trials Unit of the Medical Centre, University of Freiburg) for support during data analysis, Dr Martina Jänicke (iOMEDICO) and Dr med. Leonora Houet (iOMEDICO) for support during project design, data analysis, and critical review of the manuscript, Dr Andrea Kiemen (iOMEDICO) for translation of the project plan from German to English, and Dr Christian Johansson (iOMEDICO) for preparation of the manuscript. This study was designed, managed, analyzed, and sponsored by iOMEDICO. The BOMET-QoL-10 is the intellectual property of Dr Francisco Javier Badia Llach, author of the original validation study. The rights of the German version remain with this author.