Measuring health-related quality of life of patients with metastatic colorectal cancer using the Jordanian EQ-5D-3L value set: a cross-sectional observational study

Abstract

Objective

Health-related quality of life (HRQoL) of metastatic colorectal cancer (mCRC) in Jordan has been previously evaluated using disease-specific HRQoL tools. Meanwhile, data on HRQoL utility scores for calculating Quality-Adjusted Life Years for economic evaluations are lacking. In this study, we aim to describe, measure and identify predictors of HRQoL utility scores in patients with mCRC.

Design

This was a cross-sectional, non-interventional, observational study.

Setting

A specialised cancer centre in Jordan.

Participants

A cross-sectional questionnaire survey was conducted on 164 mCRC adult patients.

Outcome measures

Using the five-level EuroQol-3-dimension (EQ-5D-3L) instrument, patients’ health profiles were described and then valued using the EQ-5D-3L value set for Jordan to generate a single utility score. The Kruskal-Wallis test assessed differences in mean utility scores across patient characteristic categories. A Tobit regression model was used to identify potential predictors of HRQoL in mCRC patients.

Results

A total of 164 patients were enrolled with a mean age of 59 years, a mean utility score of 0.78 (SD±0.25) and visual analogue scale score of 68.78 (SD: ±19.9). 19% of patients had a stoma, and most of the patients reported health problems (72%); pain and discomfort were reported by (55%), followed by mobility (32%), usual activities (29%), anxiety/depression and self-care (13%). Analysis revealed that patients with more than one metastatic site, those who received more than one line of systemic treatment, were currently on chemotherapy, received systemic therapy in the last year or had peritoneal metastasis were found to have significantly lower utility scores (p<0.05). In contrast, patients who were employed at the time of the interview had significantly higher utility scores (p<0.05). A multivariate Tobit regression model showed that the number of metastatic sites and number of systemic treatment lines were significant predictors of lower utility scores (p<0.05). Conversely, being employed was a significant predictor of higher utility scores (p<0.05).

Conclusion

Utility scores measured in this study could be valuable for future economic evaluations of mCRC treatments. Pain and discomfort were the most reported problems among patients, highlighting the need for further evaluation to improve pain management strategies. Additionally, our regression analysis identified significant predictors of HRQoL.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study has a large sample size compared with other published studies measuring the HRQoL of metastatic colorectal cancer (mCRC) patients.

• Collecting the EuroQol-3-dimension data through telephone interviews helped avoid missing data, as this mode of administration allowed for real-time clarification and ensured completeness of responses.

• Health-related quality of life (HRQoL) of newly diagnosed mCRC patients could not be captured in our study.

• This study only included mCRC patients, so we could not capture the effect of the cancer stage on HRQoL.

Introduction

One of the most common cancers worldwide is colorectal cancer (CRC); it is the third most commonly diagnosed cancer in males and females and the second leading cause of cancer deaths around the world.¹ The number of metastatic colorectal cancer (mCRC) cases represents around one-third of all diagnosed CRC cases in Jordan.² Furthermore, CRC has the second highest economic burden among cancers,³ where mCRC is associated with the highest direct medical costs among all CRC stages in Jordan.⁴ Moreover, in recent years, the incidence of early-onset CRC (defined as new cases at age <50 years) has significantly increased in multiple countries.5,7 This increase expanded the number of younger individuals affected by CRC.

Over the past two decades, our understanding of the pathogenesis of CRC has advanced considerably. This progress has been translated into the development of biomarker-driven targeted therapies.^{8 9} These therapies include anti-angiogenic agents, epidermal growth factor receptor inhibitors, HER2-directed therapies, BRAF inhibitors, KRAS inhibitors and immune checkpoint inhibitors. The use of biologic agents, new treatment regimens and surgical resection of liver metastasis contributed to the improvement of the mCRC 5-year survival. It increased from 15.7% for patients diagnosed during (2004 to 2006) to 26% for those diagnosed during (2013 to 2015).¹⁰ As these new treatment modalities have improved morbidity and mortality in patients with mCRC, it is essential to measure patients’ health-related quality of life (HRQoL).

HRQoL of colorectal cancer patients in Jordan has been measured previously.^{11 12} However, these studies used disease-specific HRQoL tools like the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30) and the colorectal cancer-specific module (EORTC QLQ-CR 29) to measure HRQoL. Those disease-specific HRQoL tools are used to understand the impact of a disease and its treatments on HRQoL. The main limitation of these tools is that they are non-preference-based and cannot generate a single index score that reflects the value of patients’ health states. Therefore, we cannot use those measures to calculate QALYs, the primary outcome of cost-utility economic evaluation analysis.¹³

Though there are many preference-based tools to measure HRQoL,¹⁴ the most commonly recommended tool by several Health Technology Assessment (HTA) agencies is the EuroQol Quality of Life 5-Dimensions (EQ-5D)¹⁵; there are two adult versions of the EQ-5D tool that vary based on the severity levels: the three-level version (EQ-5D-3L) and the five-level version (EQ-5D-5L). These tools consist of two parts: a description of the patient’s health profile and a score index called utility reflecting HRQoL.^{15 16} Adjusting patients’ life years using the HRQoL measured as a utility value generates the Quality Adjusted Life Year (QALYs), an outcome used in economic evaluations.¹⁴

HRQoL for mCRC was assessed in many countries using the EQ-5D tool.17,20 In these studies, utility scores were calculated using country-specific utility value sets for China, Iran and Germany.^{17 19 20} It is preferable to use utilities in economic evaluation studies that reflect the countries’ social preferences of health states.¹³ Recently, the EQ-5D-3L value set for Jordan was published.²¹ Therefore, this study aimed primarily to describe and measure the HRQoL (utility scores) in patients with mCRC in Jordan using the EQ-5D-3L and its accompanying national value set.²¹ Our second objective was identifying predictors of HRQoL (utility scores) in patients with mCRC.

Method

Study design and setting

This was a cross-sectional, non-interventional, observational study conducted at King Hussein Cancer Center (KHCC), the only specialised cancer centre in Jordan, treating around (60%) of cancer cases in Jordan.²² The study was approved by the institutional review board (IRB) of KHCC and was waived from patient written consent. Verbal consent was attained from the patients to approve their enrolment in the study (IRB number: 22-KHCC-84) by the interviewers. All data were collected between June 2022 and December 2024.

Study population

The study included all adult patients diagnosed with primary mCRC and receiving oncologic treatment at KHCC. We identified patients using the KHCC hospital registry. A phone call was made to all patients who were alive at the time of the study to invite them to participate. Patients were excluded if they refused to participate or did not answer the phone call three times on three different days.²³ When the participants cannot complete the questionnaire themselves due to mental or physical limitations, caregivers of the patients will answer a proxy version of the EQ-5D-3L on behalf of the patients.²⁴

Outcome measure instrument

The EQ-5D-3L Arabic version was used to evaluate HRQoL and produce both patient health profiles and utility index scores. The EQ-5D is an instrument that measures HRQoL based on five dimensions of health: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension has three possible levels (1=no problem, 2=moderate problem and 3=severe problem). Answering the five questions generates a five-digit number reflecting the patient’s health profile. The EQ-5D-3L included 243 unique possible health states, where 11 111 represents the best health state (no problems), and 33 333 represents the worst (severe problems in all dimensions).¹⁶

In addition, the EQ-5D questionnaire includes a visual analogue scale (EQ-VAS) to measure the patient’s perception of their health on the day of the interview. The EQ-VAS is a scale ranging from zero to 100, with 0 representing the worst possible status and 100 the best possible status.

Data collection

Data was collected through telephone interviews conducted by two interviewers. The principal investigator prepared a telephone interview script that included the study cover letter, the Arabic version of the EQ-5D-3L script for telephone administration,²⁵ and questions related to socio-demographic information (employment, education, marital status). Responses were recorded on a copy of the EQ-5D telephone script and then single-entered into Microsoft Excel. Before starting data collection, interviewers received a training session by an experienced pharmacist in conducting HRQoL studies using the EQ-5D tool. Patients’ and disease characteristics were collected from medical records using a standardised data collection form; only study members had access to the Excel sheet by the same interviewers. Prior to analysis, we applied post-entry data checks (data ranges and cross-field logic rules) and cleaned flagged records accordingly.

Study’s dependent variable

The dependent variable in our study was the mean utility score, which was calculated using the Jordanian value set for EQ-5D-3L.²¹ This value set was developed using the EuroQol Group valuation protocol (EQ-VT v2.1).^{26 27}

Study’s independent variable

The independent variables (potential predictors) included sociodemographic (marital status, education, current employment status), patients’ and disease characteristics (age, gender, comorbidities other than cancer primary site of cancer (colon vs rectum), diagnosis date, sites of metastasis, previous surgeries related to the primary disease, having a stoma, the date of the last systemic therapy cycle and the lines of systemic therapies the patient received including the current line of treatment).

Statistical analysis

All patients with EQ-5D-3L data were included in the final analysis. For continuous variables, results were presented as means and SD. For binary outcomes, results were presented as frequencies. The normal distribution of the utility scores was tested using the Kolmogorov-Smirnov test. The results showed that respondents’ utility scores did not follow a normal distribution. Therefore, the Kruskal-Wallis statistical test was used to identify differences in mean utility scores based on patient characteristics. A p-value of less than 0.05 was considered statistically significant. A ceiling effect frequently occurs in HRQoL studies, where many participants achieve the highest possible score^{28 29}; this effect is notably prevalent with the EQ-5D tool, leading to utility scores being capped at 1. Therefore, the Tobit regression model is recommended for handling censored data like EQ-5D utility data. Univariate and multivariate Tobit regression models were employed to identify relationships between utility score and independent variables and identify HRQoL (utility) predictors. We started with univariate Tobit regressions; statistically significant variables (p<0.1) were included in the multivariate Tobit model. Data was collected in real time to avoid any missing data. All data analyses and figure production were performed using the R programming language, V. 2023.12.1.402.³⁰

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Results

A total of 637 eligible, alive mCRC patients were identified using the KHCC registry. Of these, 408 patients could not be reached, and 47 died during the study implementation. 182 alive patients were contacted to participate in our study. 164 patients agreed to participate and completed the EQ-5D questionnaire with a response rate of 90% (164/182). Figure 1 presents the patients’ recruitment flow chart.

Figure 1. Patient flow chart. mCRC, metastatic colorectal cancer.

The patients’ population mean age was 59 years (SD±11.4). Most enrolled patients were male (59%), and approximately 61% had completed college or a higher level of education. Furthermore, 73% of patients were not employed at the time of the interview. Two-thirds of the patients were diagnosed with colon cancer (63%). Moreover, 38% had a metastatic disease in two or more separate areas within the body. Most participants (55%) have been diagnosed with the disease for more than 2 years at the time of the interview. Additionally, 62% of the patients were off systemic treatment, and 81% did not have a stoma at the time of the interview. 38% of the patients received at least two lines of treatment, and 44% used targeted systemic therapy as part of their treatment plan. Patients’ characteristics are summarised in table 1.

Table 1. Sociodemographic, clinical characteristics and utility score in patients with different characteristics.

	Frequency (n)	Frequency (%)	Mean utility	SD	CI	P values
All	164	100%	0.78	0.25	0.70 to 1.00
Age
Below 60 years	86	52.44	0.77	0.25	0.63 to 0.92	0.44
Above 60 years	78	47.56	0.79	0.26	0.71 to 0.88
Gender
Female	68	41.46	0.77	0.28	0.61 to 1.00	0.74
Male	96	58.54	0.79	0.23	0.70 to 1.00
Marital status
Married	141	85.98	0.80	0.24	0.71 to 1.00	0.39
Single/widowed/divorced	23	14.02	0.72	0.3	0.55 to 0.96
Education level
High school	64	39.02	0.75	0.26	0.60 to 0.91	0.16
College/vocational	39	23.78	0.79	0.28	0.70 to 1.00
University/higher	61	37.20	0.81	0.26	0.75 to 1.00
Currently employed
No	119	72.56	0.76	0.26	0.61 to 0.91	0.02
Yes	45	27.44	0.84	0.23	0.80 to 1.00
Other comorbidities
None	89	54.27	0.78	0.25	0.66 to 1.00	0.81
More than 1	75	45.73	0.79	0.25	0.70 to 1.00
Site of primary cancer
Colon	104	63.41	0.77	0.27	0.61 to 1.00	0.92
Rectal	60	36.59	0.80	0.21	0.71 to 0.92
Number of metastases
One	102	62.20	0.82	0.23	0.73 to 1.00	0.01
More than one	62	37.80	0.72	0.27	0.60 to 0.92
Duration since diagnosis
Less than 2 years	74	45.12	0.77	0.24	0.71 to 1.00	0.36
More than 2 years	90	54.88	0.79	0.26	0.66 to 0.92
Presence of stoma
No	132	80.49	0.78	0.26	0.71 to 1.00	0.95
Yes	32	19.51	0.77	0.24	0.65 to 1.00
Number of treatment lines
Less than or equal to 1	102	62.20	0.83	0.19	0.73 to 1.00	0.003
More than 1	62	37.80	0.70	0.31	0.60 to 0.91
Duration since last chemotherapy
Less than 1 year	106	64.63	0.74	0.28	0.61 to 0.92	0.003
More than 1 year	58	35.37	0.85	0.16	0.78 to 1.00
Targeted therapy
No	92	56.10	0.78	0.26	0.70 to 1.00	0.98
Yes	72	43.90	0.78	0.24	0.70 to 1.00
Currently on chemotherapy
No	101	61.59	0.81	0.26	0.74 to 1.00	0.01
Yes	63	38.41	0.75	0.21	0.60 to 0.91
Presence of peritoneal metastasis
No	146	89.02	0.88	0.25	0.70 to 1.00	0.01
Yes	18	10.98	0.71	0.24	0.49 to 0.87

The mean reported EQ-VAS score was 68.8 (SD: ±19.9), and most of the patients reported having a health problem (72%). Pain or discomfort was reported by 55% of the individuals, followed by mobility (32%), performing usual activities (29%), anxiety/depression (28%) and self-care (13%). Figure 2 presents the distribution of health problems with each dimension of the EQ-5D-3L for mCRC patients. The overall mean utility score for mCRC is 0.78 (SD: ±0.25), ranging from −0.32 to 1. Figure 3 shows the distribution of utility scores for the total population.

Figure 2. Problems reported by patients in the EQ-5D-3L.

Figure 3. Distribution of utility score. EQ-5D-3L 2, five-level EuroQol-3-dimension.

Utility scores did not show statistically significant differences across most of the patients’ baseline characteristics (table 1). However, patients with more than one metastatic site, those who received more than one line of systemic treatment, were currently on chemotherapy, received systemic therapy in the last year or had peritoneal metastasis were found to have significantly lower utility scores (p<0.05). In contrast, patients who were employed at the time of the interview had significantly higher utility scores (p<0.05). Meanwhile, patients of different ages, genders, marital statuses and education levels did not have significantly different utility scores (table 1).

The univariate Tobit regression model identified the number of metastatic sites, number of systemic treatment lines, recent chemotherapy, peritoneal metastasis and employment status as significant predictors of utility scores (table 2). A subsequent multivariate Tobit regression model including these predictors (table 3) confirmed that the number of metastatic sites and number of systemic treatment lines were significant predictors of lower utility scores. Conversely, being employed was a significant predictor of higher utility scores. However, recent chemotherapy, which was initially a significant predictor, became non-significant in the multivariate model.

Table 2. Results of univariate Tobit regression model on EQ-5D-3L index scores.

	Frequency (n)	Frequency (%)	Coefficient	SE	P values
All	164	100%
Age
<60	86	52.44	Ref		0.42
≥60	78	47.56	−0.04	0.05
Gender
Female	68	41.46	Ref		0.61
Male	96	58.54	0.03	0.05
Marital status
Married	141	85.98	Ref		0.22
Single/widowed/divorced	23	14.02	−0.09	0.007
Education level
College/vocational	39	23.78	Ref
High school	64	39.02	−0.06	0.07	0.4
University/higher	61	37.20	0.026	0.07	0.7
Employment status
No	119	72.56	Ref		0.03
Yes	45	27.44	0.13	0.06
Other comorbidities
None	89	54.27	Ref		0.8
More than 1	75	45.73	0.01	0.05
Site of primary cancer
Colon	104	63.41	Ref		0.9
Rectal	60	36.59	0.01	0.05
Number of metastases
More than one	62	37.8	Ref		0.01
One	102	62.20	0.13	0.05
Duration since diagnosis
More than 2 years	90	54.88	Ref		0.4
Less than 2 years	74	45.12	−0.04	0.05
Presence of stoma
No	132	80.49	Ref		0.78
Yes	32	19.51	0.02	0.07
Number of lines of treatment
>1	62	37.8	Ref		0.0004
	102	62.20	0.18	0.05
Duration since last chemotherapy
More than 1 year	58	35.37	Ref		0.005
Less than 1 year	106	64.63	−0.15	0.05
Targeted therapy
No	92	56.10	Ref		0.9
Yes	72	43.90	0.01	0.05
Currently on chemotherapy
No	101	61.59	Ref		0.05
Yes	63	38.41	−0.1	0.05
Presence of peritoneal metastasis
No	146	89.02	Ref	0	0.04
Yes	18	10.98	−0.17	0.08

Table 3. Results of multivariate Tobit regression model.

Variables	Regression coefficient	SE	P values
Currently employed (ref=No)
yes	0.13	0.055
Number of treatment line (ref=more than one line)
one line of treatment	0.12	0.05
Number of metastases (ref=more than one site)
one site of metastases	0.10	0.05
Currently on chemotherapy (ref=No)
Yes	−0.02	−0.05	0.72
Assessment of model performance AIC	136.40
BIC	154.86

Discussion

In recent years, patient-reported outcome measures, including HRQoL, have gained an increasing focus, particularly in oncology.³¹ HRQoL is a multidimensional concept that captures patients’ subjective evaluations of their health, which are influenced not only by their medical condition but also by family and social circumstances in their environment.³² This study aimed to measure and describe the HRQoL in patients with mCRC in Jordan. To our knowledge, this is the first study to measure the HRQoL of mCRC in Jordan using the EQ-5D-3L questionnaire and its accompanying national value set.²¹ By quantitatively measuring HRQoL, our findings provide valuable data that could be incorporated into future cost-effectiveness models, offering insights from the Jordanian population.

Our results illustrated that the mean utility score and the EQ-VAS for mCRC patients in our study were 0.78 and 68.78, respectively, lower than the Jordanian general population EQ-VAS of (81.24).²¹ mCRC patients in our study appear to have lower HRQoL than those reported in Finland (0.85–0.87)^{18 33} and Germany (0.82).¹⁹ However, our mCRC patients have significantly higher HRQoL than those reported in Iran (0.45), Vietnam (0.45), China (0.49) and Indonesia (0.68).^{17 20 34 35} These findings could be due to using different value sets for each country, informed by different health states. However, our study results are lower than those reported in high-income countries but higher than HRQoL results in some low-middle-income countries. This might be explained by the level of development of healthcare systems and the availability of newer innovative medications.³⁶ Furthermore, it is important to consider whether the EQ-5D-3L or EQ-5D-5L was used to generate utility scores, as the EQ-5D-5L tends to have a lower ceiling effect than the EQ-5D-3L,³⁷ which may have influenced the results. Notably, studies from Vietnam, Indonesia, Iran and China used the EQ-5D-5L,^{17 20 34 35} whereas studies from the United Kingdom, the Netherlands and Finland used the EQ-5D-3L.^{18 33 38}

Despite advancements in pain management and the development of comprehensive guidelines over the past two decades, pain remains a prevalent issue, affecting 70% of CRC patients.³⁹ In many cases, this pain is still poorly controlled, indicating a gap in pain management, especially in patients with metastatic disease.³⁹ In our study, pain or discomfort was the most reported problem among individuals, affecting 55% of mCRC patients. This was followed by mobility problems (32%), difficulties in performing usual activities, anxiety or depression and problems with self-care. These findings align with previous studies, where pain or discomfort is typically the most commonly reported problems, while self-care problems were the least frequently reported.^{17 20 34 35 38} A previously conducted study in Germany showed a significant negative relationship between the severity of pain and utility score.¹⁹ Similarly, the results of our study suggest a need to investigate the underlying causes of pain and discomfort in patients with mCRC in future studies to improve management strategies.

The complexity of mCRC and the heterogeneity in treatment modalities can lead to varying HRQoL across patient populations.^{18 33 35 36 40 41} Some patients have resectable disease and are treated with curative intent, while others receive palliative care. Additionally, patients may be on systemic therapy, taking a treatment holiday, undergoing multiple surgeries or receiving only the best supportive care. These different treatment paths may influence patient HRQoL.^{18 33 36 40} As a result, the presence of different subgroups within a study can substantially impact the overall mean utility score, making comparisons between studies challenging. Previous studies that observed patients shortly after their initial diagnosis reported lower HRQoL^{17 33} compared with those that observed mCRC patients within 1 year after diagnosis.^{18 33 38 42 43}

Consistent with our findings, a cross-sectional study by Borchert et al found that patients undergoing three or more lines of palliative treatment had worse HRQoL. Additionally, age and gender were not significant predictors of utility scores in mCRC patients.¹⁹ On the contrary, the study did not find a relationship between the number of metastatic sites, the presence of peritoneal metastases and HRQoL. A review by Marventano et al reached a similar conclusion regarding the relationship between gender and HRQoL; it found that gender was not a significant predictor of HRQoL. Furthermore, age and having a stoma were controversial determinants of HRQoL across the literature.⁴¹ The review also reported no relationship between educational level or marital status and higher HRQoL. However, wide social networks positively influenced patients’ HRQoL.⁴¹ Our findings also showed no statistically significant relationship between surgical procedures and HRQoL, possibly because surgical procedures only affect short-term HRQoL, which patients typically restore about 3 months post-surgery.⁴¹

Our study encountered some limitations. While our study was carried out at a single centre, it is worth mentioning that KHCC is the only specialised cancer centre in Jordan, treating approximately 60% of cancer patients in the country.²² Therefore, our result can be considered generalisable to Jordan. As a cross-sectional study, our results reflect a snapshot of mCRC patients’ HRQoL without information about how HRQoL changes with time. Another area for improvement of our study is that we could not capture the HRQoL of newly diagnosed mCRC patients, as patients’ data is only available 6 months after their first contact with our institution. As well as we could not assess the association between HRQoL score and surgical procedures, date of surgery was not collected and therefore time from surgery to HRQoL was not one of our independent variables. HRQoL following surgery represents a transient health state and would likely yield a different HRQoL profile. Capturing this postoperative trajectory would require a different study design with a longitudinal follow-up to assess HRQoL across these distinct phases. Additionally, our study only included mCRC patients, so we could not capture the effect of the cancer stage on HRQoL. And finally, collecting data through telephone interviews has hindered collecting highly sensitive personal data, especially sexual function-related data.

A strength of this study is its relatively large sample size compared with other published studies measuring the HRQoL of mCRC patients.^{17 20 33 34 38} Despite focusing only on mCRC, the larger sample allows for more robust conclusions. Additionally, collecting the EQ-5D-3L data through telephone interviews helped us avoid missing data, as this mode of administration allowed for real-time clarification and ensured completeness of responses. Another strength of our study is using the Jordanian national value set,²¹ and this will encourage health economists to conduct national cost-utility analysis studies that take into consideration Jordanian national preferences of health states. Finally, this study represents an important milestone towards implementing a patient-reported outcomes (PROs) system in the future. Collecting PROs in a regular and systematic way will improve the detection of changes in quality of life and enable earlier interventions before symptoms worsen. Symptom monitoring through PROs offers an evidence-based approach to identifying emerging issues, providing clinicians with critical information to guide patient care. This approach has the potential to enhance clinical management, while reducing emergency visits and hospital admissions.⁴⁴

Future studies may incorporate qualitative data to better understand the factors contributing to lower reported HRQoL scores. Research could explore pain management practices including the use of analgesics, type of pain and barriers to effective pain and discomfort management. In addition, studying the role of supplements, herbs and dietary modifications (eg, reducing sugar or protein intake) that some patients adopt with no medical advice in shaping HRQoL, along with the influence of media on patients’ beliefs and expectations.

Conclusion

HRQoL is considered one of the most important outcomes in cancer studies alongside OS. To our knowledge, this is the first study to measure the HRQoL of mCRC patients in Jordan using EQ-5D-3L accompanying the national value set. The study has confirmed previous findings that pain/discomfort is the most associated problem reported by mCRC patients. Therefore, there is a need to investigate the cause of pain in this population to improve management. Employment status, having metastases in more than one site and receiving multiple lines of systemic therapy were significant predictors of HRQoL of mCRC patients.

Data availability statement

Data are available upon reasonable request.