Health state utility values in esophageal cancer: a systematic review and meta-analysis

Introduction

Background

Worldwide, esophageal cancer is the sixth leading cause of cancer mortality and the eighth most common cancer (1). There are two types of esophageal cancer based on histological findings: adenocarcinoma and squamous cell carcinoma (2). In high-income countries, the rates of esophageal adenocarcinoma have increased over the past few decades, especially in men (2,3). The rates of esophageal squamous cell carcinoma, which is also more prevalent in men in East Asian and Middle Eastern countries, remain stable (3). The 2020 American Society of Clinical Oncology guidelines recommend multimodal therapy for patients with locally advanced esophageal cancer. Patients with esophageal adenocarcinoma should be offered preoperative chemoradiotherapy or perioperative chemotherapy and surgery, while patients with esophageal squamous cell carcinoma should be offered preoperative chemoradiotherapy and surgery or higher dose chemoradiotherapy without surgery (4). Patients undergoing these treatments often experience an array of symptoms that affect their health-related quality of life (HRQoL), psychological and social functioning, and overall quality of life (QoL) (5,6) at various stages along the treatment pathway.

Treatment for esophageal cancer is complex and expensive, and can require multiple medical visits and hospital stays, as well as indirect costs including travel and accommodation, particularly if patients are remotely located from a treatment center. In the Australian setting in 2011, Gordon et al. (7) estimated overall mean costs of US$49,265 (median US$38,562) for patients with esophageal cancer undergoing surgery and US$22,889 (median US$17,306) for those treated without surgery. When subsequently modelling the cost-effectiveness of strategies for treating esophageal adenocarcinoma, they found a cost of US$50,226 per year for stage IV disease (8). In a more recent study in 2022, the mean cost for esophagectomy was found to be US$54,070 (median US$40,892) (9). Whilst cost may not be a focus for clinicians, it is for funders, and cost and QoL need to be considered and addressed to ensure efficient allocation of resources (10).

Allocation of healthcare resources is routinely informed by economic evaluation studies, particularly cost-utility analysis, where the quality-adjusted life year (QALY) is a measure of outcome. The QALY is a composite measure of both quality and quantity of life in each health state, whereby quality is assessed as HRQoL, and the utility individuals place on that health state. When generating health state utility values (HSUVs), the more preferred or desirable health states receive greater utility (11,12). HSUVs are measured on a cardinal scale from 0 to 1, where 0 denotes death and 1 is full health. Methods of generating HSUVs can be direct, such as time-trade-off (TTO), standard gamble or discrete choice experiments (13), or indirect using utility-based HRQoL measures, also referred to as multi-attribute utility instruments (MAUIs). MAUIs can be generic, such as the Short-Form 6-Dimension (SF-6D) (14) and the EQ-5D-5L and EQ-5D-3L (15) or disease specific such as the cancer-specific European Organization for Research and Treatment of Cancer Quality of Life Utility Core 10-Dimensions (EORTC QLU-C10D) (16) and Functional Assessment of Cancer Therapy 8-Dimension (FACT-8D) (17,18).

Rationale and knowledge gap

Several studies have assessed HRQoL in individuals undergoing care and treatment for esophageal cancer. Jacobs et al. (17) reviewed these studies and identified significant limitations such as low quality of included studies and no application of a MAUI. Therefore, they had no ability to obtain HSUVs for economic analysis (17). Other empirical studies have assessed HSUVs related to esophageal cancer (6), including a systematic review (19). However, the sample sizes in the empirical studies were usually small, and the only previous systematic review did not undertake a meta-analysis to provide more robust estimates of the health state utilities associated with the different stages, treatments and timepoints following treatment for esophageal cancer. Such estimates are critical to informing QALY inputs for the different health states which are required for cost-utility analysis models needed to assess whether new therapies offer clinical benefits at a reasonable cost, and can then be recommended for widespread clinical use (20).

Objective

The primary aim of this review was to critically appraise and summarize the up-to-date literature on HSUVs for esophageal cancer at different disease stages, and for various treatments, to then inform QALY inputs for cost-utility analysis. A secondary aim was to pool data through a meta-analysis to provide more accurate estimates of HSUVs for the health states involved.

Review questions

• What methods have been used to obtain utility values for patients undergoing care and treatment for esophageal cancer?
• What are the utility values associated with the different health states for patients undergoing care and treatment for esophageal cancer?

We present this article in accordance with the PRISMA reporting checklist (available at https://aoe.amegroups.com/article/view/10.21037/aoe-25-13/rc) (21).

Methods

This review was conducted according to the Good Practices Report by the International Society for Pharmacoeconomics and Outcomes Research (22), and a guide published for the systematic review and meta-analysis of health utility state values (23). The review was registered with PROSPERO (CRD42024440276).

Search strategy

Databases searched were MEDLINE (Ovid), Embase (OVID), Scopus, Web of Science Core Collections, Cochrane Register of Controlled Trials, CINAHL (EBSCOhost) and ProQuest. We also searched EconLit, NHS Economic Evaluation Database (NHS EED) and Cost-Effectiveness Analysis Registry source for additional resources. Unpublished studies and grey literature were also searched, including WorldCat, Dissertations and Theses Global, and the Health and Psychosocial Instruments database. The search included studies from database inception until September 2023 with guidance from a specialist research librarian. Finally, the reference lists of all included studies were searched for additional resources. The search strategy for MEDLINE (OVID) is provided in Appendix 1.

Study selection

All identified esophageal cancer studies that collected data for HRQoL at various stages of disease and various treatment regimens were collated in EndNote and uploaded into Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) for de-duplication, and title and abstract screening. Review authors who authored studies that were eligible for inclusion in this review had no part in screening, critically appraising, extracting, or making decisions about those studies.

Inclusion criteria

We included empirical studies that generated HSUVs using direct elicitation methods such as TTO and standard gamble, or indirect methods using a preference-based measure or a non-preference based measure but generated utility values using a validated scoring algorithm such as SF-36 mapped onto the SF-6D (22). Study designs selected included experimental studies, observational studies including prospective and retrospective cohort studies, case-control and cross-sectional studies across all stages of esophageal cancer and all interventions, including esophagectomy. The inclusion criteria for individual studies were considered and analyzed as a part of the quality assessment.

Exclusion criteria

Studies reported in languages other than English were excluded due to lack of language interpretive services, along with studies on children or animals. We also excluded studies focusing on patients with Barrett’s esophagus as these patients had a precancerous state, and patients who had a cervical or upper-third of esophagus cancer, as they often have a different treatment pathway which does not include esophagectomy, which was one of our key treatment foci. Descriptive observational studies, including case reports and case series, qualitative studies, and studies that focused on diagnosis, staging, and imaging modalities, and studies that focused on tumor marker identification and association with outcomes, were excluded.

Following screening, relevant studies progressed to full-text screening. Screening and the subsequent stages of the review were undertaken by two independent reviewers with disagreements and inconsistencies resolved through discussions with a third senior reviewer (21).

Data extraction

Data was extracted using overview tables and templates from the JBI Manual for Evidence Synthesis (24), with modifications tailored to this study following recommendations from Petrou et al. (23) for systematic reviews and meta-analysis of HSUVs. Extracted data included sample sizes, participant characteristics including cancer stage, the reported health states, methods used to obtain HSUVs, and descriptive statistics of the HSUV, such as mean, median, and standard deviations.

Quality assessment

There is no validated tool to assess the quality of studies generating HSUVs. However, several commentators have provided guidance on key characteristics to be considered. Cooper et al. (25) developed a ranking system, Papaioannou et al. (26) describe specific criteria to be considered, such as response rate, while Brazier and Deverill (27) suggest assessing the empirical validity of the MAUIs used in the study. In this review, we applied the hierarchical ranking system and the assessment criteria by both Cooper et al. (25) and Papaioannou et al. (26). As recommended by Petrou et al. (23), a holistic description of the quality of included studies was provided, which involved a comment about the relevant features of multiple checklists and relevance of the studies applying the utility measure or derived values. In addition, as recommended by Cooper et al. (25), studies were ranked from A to E. An ‘A’ ranking was given if the study reported clinical effect sizes, adverse events and complications. A ‘B’ ranking was given for studies with baseline clinical data, whilst ‘C’ was based on resource use, and ‘D’ was assigned to costing studies. Finally, an ‘E’ ranking was given to utility studies, with subcategorization into those which used (I) a direct utility assessment for the specific study, (II) indirect utility assessment without a validated tool, (III) direct assessment from a previous study, (IV) unsourced utility data from a previous study, (V) patient preference values obtained using the visual analogue scale and (VI) utility values from Delphi panels/expert opinion.

Additionally, a number from 1–6 was assigned based on the hierarchy of evidence: 1= meta-analysis of randomized controlled trials (RCTs), 2= single RCT with direct comparison between therapies, 3= single placebo-controlled RCT, 4= case control or cohort studies, 5= non-analytic studies such as case reports and case series, and 6= expert opinion. As recommended by Papaioannou et al. (26), we also applied key criteria of the HSUV studies. These key criteria included a consideration of sample size, respondent selection and recruitment, inclusion/exclusion criteria, response rates, loss to follow-up, missing data, and appropriateness of measure. The rankings and assessments are provided in Table S1.

Data synthesis

The studies were synthesized into a narrative review, with grouping based on the stage of disease or timepoint at which HSUVs were obtained. In doing so, we identified studies that reported HSUVs for the different health states using (I) similar methods, or (II) the same MAUI and scored with the same population algorithm.

Statistical analysis

To generate estimates for the different health states, studies reporting HSUVs for the same health state or disease stage with the same MAUI at the same timepoint were included in a proportional meta-analysis using random-effects models in R version 4.3.2 for Windows (R Core Team, R Foundation for Statistical Computing, Vienna, Austria). Due to significant variability in treatment, further stratification based on intervention/treatment type was not possible. Between-study heterogeneity was assessed using the χ² (Q) test, with the proportion of true heterogeneity estimated using the I² statistic. Sensitivity analyses were conducted to examine baseline health utility scores measured using the EQ-5D-5L instrument for patients with different stages of esophageal cancer.

Results

Study selection

A total of 5,929 records were identified for title and abstract screening. Of these, 446 were included for full-text screening, and 13 were then determined to be eligible for inclusion in the qualitative evidence synthesis. This is summarized in the PRISMA flow diagram (Figure 1).

Figure 1 PRISMA diagram describing the results of the literature search and the reasons for study exclusion. EAC, esophageal adenocarcinoma; HSUV, health state utility value; QoL, quality of life.

Study characteristics

Of the 13 studies, six were RCTs (28-33) and seven were prospective observational studies (6,34-39). Three studies were conducted in China and two each in the UK, the Netherlands, and Australia. There was one study each from Canada, France and the US. Table 1 summarizes the included studies.

Study quality assessment

The majority of the studies were ranked as A1 or E1(c) (29) according to the hierarchies of data sources modified by Cooper et al. (25). The prospectively designed studies were classified as the most appropriate source in most cases and also used the most appropriate outcome measure (25,26). Most of the studies were also very clear, with a detailed description of the inclusion criteria (6,29-35,37-39), exclusion criteria (28-30,33,35,36,39) and response rates (6,28-30,32,33,35-39). Some studies were explicit in their handling of missing data (28,29,32-34,36,37) and loss-to-follow up (32,36,38), but the majority did not outline these steps (6,30,31,35,39). All studies had a reliable patient data source with patients already randomized to a treatment of interest (28,30-33,38), or they included prospectively collected data that focused on HRQoL (6,34-37,39). There was a clear description of HRQoL generation with direct evidence collection in each study. Results of the quality assessment are provided in Table S1.

HSUV elicitation methods

Most of the studies elicited utilities using a MAUI, with 9 (69.2%) using EQ-5D-3L (6,28,29,32-34,36,37,39), 3 (23.1%) using EQ-5D-5L (31,35,38), two using the QLU-C10D (28,35), and one study using the SF-6D (35). Direct elicitation was applied in two studies, using the standard gamble (30) and the TTO (37). Three studies applied two or more instruments or methods (28,35,37).

Health states and type of treatment received

Stage I–III: resectable disease

Active surveillance (no surgery) with neoadjuvant chemotherapy (nCRT)

One study by Hermus et al. (38) included patients with both squamous cell cancer and adenocarcinoma of the esophagus under active surveillance with no surgery. nCRT was administered according to the Chemoradiotherapy for Esophageal Cancer followed by Esophagectomy Surgery Study (CROSS) protocol (five cycles of neoadjuvant intravenous chemotherapy, carboplatin and paclitaxel with concurrent radiotherapy). Hermus et al. (38) reported a high utility score of 0.95 (0.08) and 0.92 (0.07) at three months and twelve months post-treatment respectively using the EQ-5D-5L.

Surgery after nCRT

Two studies included individuals who received surgery following nCRT, using cisplatin and 5-fluorouracil chemotherapy plus radiotherapy (29) and the CROSS protocol (38). Using the EQ-5D-5L, Hermus et al. (38) reported a high score of 0.82 (0.13) and 0.86 (0.13) at three months and twelve months post-surgery, respectively. Similarly, Malthaner et al. (29) reported HSUV of 0.90 (0.18) at baseline and 0.87 (0.15) twelve months post-treatment using the EQ-5D-3L.

Adjuvant chemotherapy with extended volume radiotherapy and surgery

For postoperative patients receiving adjuvant CRT with extended volume radiation treatment that included the anastomosis (29), high HSUVs of 0.84 (0.15) and 0.86 (0.19) using EQ-5D-3L were reported at baseline and 12 months post-treatment, respectively.

Surgical resection

One study reported standard gamble values when comparing two different surgical techniques. De Boer et al. (30) compared transhiatal resection with extended transthoracic resection, and found that the postoperative HRQoL was high with HSUV of 0.96 (0.07) for the at-home, disease-free state.

Stage IV: locally advanced unresectable or metastatic carcinoma of the esophagus, including recurrence

Immunotherapy and chemotherapy

Qu et al. (31) obtained data from the KEYNOTE-590 multi-center RCT which applied the US-specific EQ-5D-5L algorithm to obtain HSUV. The treatments compared were pembrolizumab plus cisplatin and fluorouracil vs. fluorouracil and cisplatin alone vs. a blended chemotherapy regimen. They reported a high HSUV of 0.90 for ≥360 (time to death, days) and adverse event disutility value of 0.04 for the combined data.

Continuation and discontinuation of first-line chemotherapy

Marguet et al. (32) defined three health states for patients allocated to first-line fluorouracil/platinum-based chemotherapy continuation (CT-CONT) and to a chemotherapy discontinuation (CT-DISC): the first instance of disease progression, after the first instance of disease progression, and death. HSUVs reported for the three health states using EQ-5D-3L were 0.75, 0.67 and 0, respectively, for the CT-CONT group, and 0.83, 0.73, and, 0 for the CT-DISC group.

Self-expanding metal stents (SEMS) and non-SEMS therapies

Shenfine et al. (33) used the EQ-5D-3L questionnaire in a UK setting. Reported HSUVs for patients with SEMS were 0.56 (0.35) at baseline and 0.49 (0.36) at six weeks. For patients receiving non-SEMS therapies, HSUVs were 0.56 (0.30) at baseline and 0.45 (0.32) at six weeks.

All interventions, all disease-stages

Six studies reported HSUVs across different pathological stages of esophageal cancer and included all types of treatments received at various timepoints (6,34-37,39). The mean HSUVs across all of these ranged from 0.60 to 0.89. Five studies used the EQ-5D-3L only (6,34,36,39), one used EQ-5D-5L (35) and one study additionally used both EQ-5D-3L and TTO valuation (37).

EQ-5D-3L

Wildi et al. (37) generated EQ-5D-3L HSUVs from patients with newly diagnosed esophageal cancer following treatment, which consisted mainly of surgery for local disease, chemoradiotherapy and surgery for regional disease, and palliative care for metastatic disease (37). Mean HSUVs for the different surveillance, epidemiology, and end results (SEER) stages were 0.93 [standard deviation (SD) 0.12, 95% confidence interval (CI): 0.65–1.20] for stage 0, 0.60 (SD 0.29, 95% CI: 0.41–0.80) for stage I, 0.71 (SD 0.21, 95% CI: 0.62–0.80) for stage II, and 0.69 (SD 0.35, 95% CI: 0.47–0.91) for stage III. Ding et al. (39) generated HSUVs from EQ-5D-3L in a Chinese population and reported a HSUV median (range) of 0.86 (0.76, 0.88).

Doherty et al. (36) reported pre-treatment EQ-5D-3L HSUVs from Canadian patients. Patients under post-treatment surveillance had a combined score of 0.88 (0.21), with 0.87 (0.15) and 0.78 (0.77) for stage I and stage II/III disease, respectively. Patients who either underwent definitive chemoradiotherapy without surgery, or neoadjuvant chemoradiotherapy and surgery, had a combined score of 0.78 (0.15) pre-treatment and 0.75 (0.16) post-treatment. When stage I and stage II/III disease were separated, the scores were 0.82 (0.13) and 0.73 (0.13) for pre-treatment and post-treatment, respectively, for stage II/III. Stage I disease only had one patient with a score 0.71 pre-treatment and 0.94 post treatment. The palliative chemotherapy patients had a pre-treatment score of 0.72 (0.18) and post-treatment score of 0.74 (0.19) (36).

Wang et al. (34) similarly separated the EQ-5D-3L-generated HSUV scores for patients with esophageal cancer based on stage of disease and treatment in a Chinese cohort. Most patients had stage II or III disease and underwent surgery with or without adjuvant or nCRT. The HSUV for patients who underwent surgery only was 0.66 (95% CI: 0.64–0.68), and the HSUV for patients who underwent both surgery and postoperative chemotherapy, was 0.75 (95% CI: 0.71–0.78). Regardless of treatment, the scores for stages I, II, III and IV diseases were 0.69, 0.75, 0.76 and 0.75, respectively.

The final study by Liu et al. (6) reported HSUVs from EQ-5D-5L for Chinese esophageal cancer patients who underwent surgery or radical resection as a monotherapy. Postoperative utility scores were 0.90 (0.16) for severe dysphagia/carcinoma in situ, 0.82 (0.23) for stage I, 0.81 (0.21) for stage II, 0.80 (0.24) for stage III and 0.66 (0.34) for stage IV disease.

Six studies met the inclusion criteria for the meta-analysis. These studies applied the EQ-5D-3L with HSUVs obtained at baseline (28,32-34,36,37). The pooled utility values with EQ-5D-3L at baseline was 0.66 (95% CI: 0.56–0.75, I²=83%) for stage I (Figure 2) with considerable heterogeneity, 0.84 (95% CI: 0.81–0.87, I²=48%) for stages II/III (Figure 3) with moderate heterogeneity, and 0.67 (95% CI: 0.57–0.78, I²=85%) for stage IV (Figure 4) with considerable heterogeneity.

Figure 2 Baseline health utility scores measured using the EQ-5D-3L instrument for patients with stage I esophageal cancer. CI, confidence interval; SD, standard deviation.

Figure 3 Baseline health utility scores measured using the EQ-5D-3L instrument for patients with stage II and III esophageal cancers. CI, confidence interval; SD, standard deviation.

Figure 4 Baseline health utility scores measured using the EQ-5D-3L instrument for patients with stage IV esophageal cancer. CI, confidence interval; SD, standard deviation.

EQ-5D-5L and SF-6D

Applying the EQ-5D-5L in an Australian population, Bulamu et al. (35) generated HSUVs, with esophageal cancer patients (all stages with various treatments) scoring 0.82 (0.18), and patients undergoing surgery scoring 0.81 (0.15). This same study also reported SF-6D HSUVs of 0.62 (0.25) for patients with esophageal cancer, and 0.57 (0.27) for patients undergoing surgery. Results from the sensitivity analysis of baseline health utility score are presented in Figure 5, with considerable heterogeneity (I²=89%). This was conducted to assess the effect of stage of disease on the health state utility outcomes but showed considerable heterogeneity.

Figure 5 Baseline health utility scores measured using the EQ-5D-5L instrument for patients with different stages of esophageal cancer. CI, confidence interval; SD, standard deviation.

QLU-C10D

Two Australian studies applied the QLU-C10D, with HSUVs for patients with esophageal cancer (all stages with various treatments) scoring 0.75 (0.17), and patients undergoing surgery scoring 0.72 (0.17) (35). Bulamu et al. (28) reported a preoperative score of 0.81 (0.16) and postoperative score of 0.52 (0.22).

Discussion

To the best of our knowledge, this is the first comprehensive systematic review and meta-analysis of HSUVs across all stages of esophageal cancer. Most studies used EQ-5D-3L to obtain HSUVs, followed by EQ-5D-5L in more recent studies. This is not surprising as the EQ-5D, initially developed in 1990, is one of the most widely used and validated MAUIs, and it has been translated into several languages (15,40). It is also recommended for use in economic evaluation submissions for decision-making bodies such as the National Institute for Health and Care Excellence in the UK (41) and the Medical Services Advisory Committee in Australia (42). Initially designed as a three-level answer structure (EQ-5D-3L), it was revised to the five-level version (EQ-5D-5L) with increased feasibility, sensitivity and reliability (43).

One of the cancer-specific health utility measures is the QLU-C10D, which, when compared to the EQ-5D-3L in a multicenter RCT among esophageal cancer patients, was found to be more sensitive to short-term utility changes and more reliably able to distinguish between health states (28). This measure showed good comparative validity but showed a higher detection rate and measurement efficiency when compared to EQ-5D-3L (44). However, compared to EQ-5D-5L in Chinese gastric cancer patients, although possessing similar validity, the EQ-5D-5L was postulated to have better discriminative power than QLU-C10D (45). Conversely, in another more recent and larger German cross-sectional study of a general oncology cohort, the QLU-C10D was found to be better at distinguishing health states compared to the EQ-5D-5L (46).

When trying to compare already available data across published studies to assess corresponding improvements in HSUVs with changing treatment pathways, the EQ-5D-5L should be considered in future studies. However, given the specific focus on oncology and potential to better distinguish health states compared to the EQ-5D-5L, the QLU-C10D may be the preferred method to collect the HSUVs for esophageal cancer patients and inform economic analyses. Furthermore, in studies with inherently small sample sizes, we recommend direct utility elicitation studies, such as using standard gamble or TTO, as these can be robust even with small sample sizes.

The pooled utility values were relatively low at 0.66 for stage I, 0.84 for stages II/III and 0.67 for stage IV. The higher HRQoL scores for stage II/III compared to stage IV were also reported in a narrative literature review by Chung et al. (47), where stage II/III disease had a mean EQ-5D baseline utility score of 0.82 (0.13) and stage IV had a score of 0.72 (0.18). Pourrahmat et al. (48) also reported similar trends with a range of 0.46–0.81 in stage II, 0.15–0.80 in stage III, and 0.66 in stage IV disease.

When assessing HRQoL in long-term esophageal cancer survivors who had undergone potentially curative treatment, Courrech Staal et al. (49) found no statistically significant differences between utilities from patients at different disease stages. This may account for the similar HSUV scores that were generated from our analysis of the data. Interestingly, in a Swedish nationwide, prospective, population-based study, Rutegard et al. indicated that extensive surgery was not associated with lower HRQoL than less invasive operations (50). However, using the disease specific Functional Assessment of Cancer Therapy-Esophagus to assess HRQoL in esophageal cancer, Kidane et al. (51) found that HRQoL declined as the T-stage progressed, with a significant difference between T1 and T2/T3 disease, though this trend was not significant for all instruments. A recent narrative literature review found that HRQoL in a range of cancers, including pancreatic and stomach, generally declines from early to late-stage disease. However, this may not be the case for esophageal cancer (47).

There may be other factors potentially associated with HRQoL other than the stage of disease, such as the physical and emotional demands of diagnosis and treatment that should be considered (52). Additionally, it is possible that the EQ-5D is not a sensitive HRQoL measure at the less severe stages of disease, and therefore, a disease specific measure may be preferred (28,46). It can also be argued that the higher HSUVs in the later stages of disease are a result of response-shift, where patients’ conceptualization of HRQoL has changed following a major life event, such as treatment for late-stage cancer (53,54). The patient is more likely to have adapted to match their new circumstances (55), which may account for the higher HRQoL values observed in stages II/III compared to stage I disease in our review.

A strength of our study is the extensive search strategy implemented to identify studies of interest. However, the meta-analysis is limited by the heterogeneity of the treatments, disease stage and HSUV assessment time-points reported in the included studies. Although three studies utilized EQ-5D-5L, they were spread across different stages of disease; one study included patients across all stages, one with resectable stages II/III and one with unresectable stage IV disease. Therefore, none could be pooled in the meta-analysis. Only one study applied direct elicitation methods, which could not be included in the meta-analysis. In their 2015 systematic review of HSUVs for advanced gastric, esophageal and gastroesophageal junction cancers, Carter et al. (19) similarly faced challenges in making cross-study comparisons due to limited data consistency and were not able to perform a meta-analysis. Most studies that we identified had at least one utility score, which was usually at pre-treatment baseline. For studies that provided a pre-treatment and post-treatment health utility score, the postoperative timepoints were also variable, ranging from seven or 42 days, two, three or six months, and even up to five years. As such, the generalizability of the data to patients outside of the reported unique oncological settings and treatments was very limited (19) and meta-analysis was not feasible. However, results from six studies assessing HRQoL at baseline using the EQ-5D-3L were pooled in the meta-analysis. Our findings are therefore unique in that we were able to combine a few studies, despite the limitation of small sample sizes and different treatments applied in each esophageal cancer stage.

A limitation of our study, however, is the lack of stage-specific data for sensitivity analyses: the studies available for meta-analysis either grouped patients together who had various treatments, assessed QoL using different instruments or assessed HRQoL at different timepoints. Whilst EQ-5D-3L data were available across disease stages for a limited analysis, EQ-5D-5L and QLU-C10D data were only available for baseline scores and thus could not be used in meta-analysis. We were only able to identify one study that reported EQ-5D-5L scores for stage I, further highlighting this gap in the literature (6). In addition, pooled utility values using the standard gamble were only assessed postoperatively in stage II-III disease, with a mean utility of 0.92. These studies, however, were over 20 years old, with several key changes in treatment modalities occurring since their publication. The meta-analysis also included studies with considerable heterogeneity that was not able to be modified in that we were limited by the small number of studies available. The heterogeneity was particularly high for stage I and stage IV studies, whereby there was substantial variability in the types of treatments received and across countries. For the studies included in stage IV disease, in particular, the Shenfine et al. study (33) included a wide variety of treatments including palliative stents in the UK and was published in 2005, which was 16 years before the subsequent major study in the analysis by Marguet et al. (32), who only focused on chemotherapy and in France, which may account for the differences in health state utilities. As a result, interpreting the meta-analysis results should be done with caution. Therefore, results are largely presented as a narrative review, as there is a significant paucity of congruent stage-specific health state utilities, and in particular, a lack of comparison before and after a treatment modality.

Conclusions

Despite the heterogeneity in the reported health states in the setting of esophageal cancer treatment identified in this review, it is evident that a variety of methods can be used to obtain HSUVs to inform a cost-utility analysis. This review provides insight into the most frequently applied utility instruments, as well as some estimates for esophageal cancer states. In the future, there is a need for a systematic approach to define health states and assessment timepoints for the collection of health utility data in the setting of esophageal cancer, to better inform cost-utility analysis models that can then optimize resource allocation in this population. Additionally, large-scale studies comparing the different MAUIs to identify which one is most sensitive to change at the different stages, different treatments and timepoints in patients with esophageal cancer are recommended. However, due to the inherently small sample sizes of studies in this population, direct utility elicitation methods may be preferred to obtain robust estimates.

Acknowledgments

The authors would like to thank Ms. Nikki May, who is a Library Services Manager for Southern Adelaide Local Health Network and Mental Health, for her help with the creation of search terms for the literature review.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://aoe.amegroups.com/article/view/10.21037/aoe-25-13/rc

Peer Review File: Available at https://aoe.amegroups.com/article/view/10.21037/aoe-25-13/prf

Funding: This work was supported by a National Health and Medical Research Council of Australia Postgraduate Scholarship (to J.P.), a Cancer Council South Australia Beat Cancer grant (to M.A.), and a Cancer Council South Australia Beat Cancer Early Career Researcher Fellowship (to N.B.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://aoe.amegroups.com/article/view/10.21037/aoe-25-13/coif). J.P. was supported by a National Health and Medical Research Council of Australia Postgraduate Scholarship. M.A. was supported by a Cancer Council South Australia Beat Cancer grant. N.B. was supported by a Cancer Council South Australia Beat Cancer Early Career Researcher Fellowship. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Pennathur A, Gibson MK, Jobe BA, et al. Oesophageal carcinoma. Lancet 2013;381:400-12. [Crossref] [PubMed]
2. Siewert JR, Stein HJ, Feith M, et al. Histologic tumor type is an independent prognostic parameter in esophageal cancer: lessons from more than 1,000 consecutive resections at a single center in the Western world. Ann Surg 2001;234:360-7; discussion 368-9. [Crossref] [PubMed]
3. Arnold M, Soerjomataram I, Ferlay J, et al. Global incidence of oesophageal cancer by histological subtype in 2012. Gut 2015;64:381-7. [Crossref] [PubMed]
4. Shah MA, Kennedy EB, Catenacci DV, et al. Treatment of Locally Advanced Esophageal Carcinoma: ASCO Guideline. J Clin Oncol 2020;38:2677-94. [Crossref] [PubMed]
5. Lv Y, Zhang J, Qiao L. Quality of life in patients with esophageal cancer receiving definitive chemoradiotherapy or esophagectomy. Mol Clin Oncol 2014;2:870-4. [Crossref] [PubMed]
6. Liu Q, Zeng H, Xia R, et al. Health-related quality of life of esophageal cancer patients in daily life after treatment: A multicenter cross-sectional study in China. Cancer Med 2018;7:5803-11. [Crossref] [PubMed]
7. Gordon LG, Eckermann S, Hirst NG, et al. Healthcare resource use and medical costs for the management of oesophageal cancer. Br J Surg 2011;98:1589-98. [Crossref] [PubMed]
8. Gordon LG, Hirst NG, Mayne GC, et al. Modeling the cost-effectiveness of strategies for treating esophageal adenocarcinoma and high-grade dysplasia. J Gastrointest Surg 2012;16:1451-61. [Crossref] [PubMed]
9. Browning AF, Chong L, Read M, et al. Economic burden of complications and readmission following oesophageal cancer surgery. ANZ J Surg 2022;92:2901-6. [Crossref] [PubMed]
10. Shih YC, Halpern MT. Economic evaluations of medical care interventions for cancer patients: how, why, and what does it mean? CA Cancer J Clin 2008;58:231-44. [Crossref] [PubMed]
11. Weinstein MC, Torrance G, McGuire A. QALYs: the basics. Value Health 2009;12:S5-9. [Crossref] [PubMed]
12. Whitehead SJ, Ali S. Health outcomes in economic evaluation: the QALY and utilities. Br Med Bull 2010;96:5-21. [Crossref] [PubMed]
13. McNamee P, Glendinning S, Shenfine J, et al. Chained time trade-off and standard gamble methods. Applications in oesophageal cancer. Eur J Health Econ 2004;5:81-6. [Crossref] [PubMed]
14. Brazier J, Roberts J, Deverill M. The estimation of a preference-based measure of health from the SF-36. J Health Econ 2002;21:271-92. [Crossref] [PubMed]
15. Devlin NJ, Brooks R. EQ-5D and the EuroQol Group: Past, Present and Future. Appl Health Econ Health Policy 2017;15:127-37. [Crossref] [PubMed]
16. King MT, Costa DS, Aaronson NK, et al. QLU-C10D: a health state classification system for a multi-attribute utility measure based on the EORTC QLQ-C30. Qual Life Res 2016;25:625-36. [Crossref] [PubMed]
17. Jacobs M, Macefield RC, Blazeby JM, et al. Systematic review reveals limitations of studies evaluating health-related quality of life after potentially curative treatment for esophageal cancer. Qual Life Res 2013;22:1787-803. [Crossref] [PubMed]
18. King MT, Norman R, Mercieca-Bebber R, et al. The Functional Assessment of Cancer Therapy Eight Dimension (FACT-8D), a Multi-Attribute Utility Instrument Derived From the Cancer-Specific FACT-General (FACT-G) Quality of Life Questionnaire: Development and Australian Value Set. Value Health 2021;24:862-73. [Crossref] [PubMed]
19. Carter GC, King DT, Hess LM, et al. Health state utility values associated with advanced gastric, oesophageal, or gastro-oesophageal junction adenocarcinoma: a systematic review. J Med Econ 2015;18:954-66. [Crossref] [PubMed]
20. Zhu Y, Liu K, Ding D, et al. Pembrolizumab Plus Chemotherapy as First-Line Treatment for Advanced Esophageal Cancer: A Cost-Effectiveness Analysis. Adv Ther 2022;39:2614-29. [Crossref] [PubMed]
21. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372: [Crossref] [PubMed]
22. Brazier J, Ara R, Azzabi I, et al. Identification, Review, and Use of Health State Utilities in Cost-Effectiveness Models: An ISPOR Good Practices for Outcomes Research Task Force Report. Value Health 2019;22:267-75. [Crossref] [PubMed]
23. Petrou S, Kwon J, Madan J. A Practical Guide to Conducting a Systematic Review and Meta-analysis of Health State Utility Values. Pharmacoeconomics 2018;36:1043-61. [Crossref] [PubMed]
24. JBI. JBI Manual for Evidence Synthesis. JBIMES-20-01; 2024.
25. Cooper N, Coyle D, Abrams K, et al. Use of evidence in decision models: an appraisal of health technology assessments in the UK since 1997. J Health Serv Res Policy 2005;10:245-50. [Crossref] [PubMed]
26. Papaioannou D, Brazier J, Paisley S. Systematic searching and selection of health state utility values from the literature. Value Health 2013;16:686-95. [Crossref] [PubMed]
27. Brazier J, Deverill M. A checklist for judging preference-based measures of health related quality of life: learning from psychometrics. Health Econ 1999;8:41-51. [Crossref] [PubMed]
28. Bulamu NB, Vissapragada R, Chen G, et al. Responsiveness and convergent validity of QLU-C10D and EQ-5D-3L in assessing short-term quality of life following esophagectomy. Health Qual Life Outcomes 2021;19:233. [Crossref] [PubMed]
29. Malthaner RA, Yu E, Sanatani M, et al. The quality of life in neoadjuvant versus adjuvant therapy of esophageal cancer treatment trial (QUINTETT): Randomized parallel clinical superiority trial. Thorac Cancer 2022;13:1898-915. [Crossref] [PubMed]
30. de Boer AG, Stalmeier PF, Sprangers MA, et al. Transhiatal vs extended transthoracic resection in oesophageal carcinoma: patients' utilities and treatment preferences. Br J Cancer 2002;86:851-7. [Crossref] [PubMed]
31. Qu T, Zhang S, Zhong Y, et al. Cost Effectiveness of Adding Pembrolizumab to Platinum and Fluoropyrimidine-Based Chemotherapy as First-Line Treatment for Advanced Esophageal Cancer: A US Healthcare Payer's Perspective. Pharmacoeconomics 2022;40:1247-59. [Crossref] [PubMed]
32. Marguet S, Adenis A, Delaine-Clisant S, et al. Cost-Utility Analysis of Continuation Versus Discontinuation of First-Line Chemotherapy in Patients With Metastatic Squamous-Cell Esophageal Cancer: Economic Evaluation Alongside the E-DIS Trial. Value Health 2021;24:676-82. [Crossref] [PubMed]
33. Shenfine J, McNamee P, Steen N, et al. A pragmatic randomised controlled trial of the cost-effectiveness of palliative therapies for patients with inoperable oesophageal cancer. Health Technol Assess 2005;9:1-121. [Crossref] [PubMed]
34. Wang Y, Shi J, Du L, et al. Health-related quality of life in patients with esophageal cancer or precancerous lesions assessed by EQ-5D: A multicenter cross-sectional study. Thorac Cancer 2020;11:1076-89. [Crossref] [PubMed]
35. Bulamu NB, Chen G, Ratcliffe J, et al. Health-Related Quality of Life Associated with Barrett's Esophagus and Cancer. World J Surg 2019;43:1554-62. [Crossref] [PubMed]
36. Doherty MK, Leung Y, Su J, et al. Health utility scores from EQ-5D and health-related quality of life in patients with esophageal cancer: a real-world cross-sectional study. Dis Esophagus 2018;
37. Wildi SM, Cox MH, Clark LL, et al. Assessment of health state utilities and quality of life in patients with malignant esophageal Dysphagia. Am J Gastroenterol 2004;99:1044-9. [Crossref] [PubMed]
38. Hermus M, van der Wilk BJ, Chang RTH, et al. Patient preferences for active surveillance vs standard surgery after neoadjuvant chemoradiotherapy in oesophageal cancer treatment: The NOSANO-study. Int J Cancer 2023;152:1183-90. [Crossref] [PubMed]
39. Ding H, Mao A, Lin J, et al. Using a Chinese time trade-off approach to explore the health utility level and quality of life of cancer patients in urban China: a multicentre cross-sectional study. Support Care Cancer 2021;29:2215-23. [Crossref] [PubMed]
40. Janssen MF, Bonsel GJ, Luo N. Is EQ-5D-5L Better Than EQ-5D-3L? A Head-to-Head Comparison of Descriptive Systems and Value Sets from Seven Countries. Pharmacoeconomics 2018;36:675-97. [Crossref] [PubMed]
41. Dawoud D, Lamb A, Moore A, et al. Capturing what matters: updating NICE methods guidance on measuring and valuing health. Qual Life Res 2022;31:2167-73. [Crossref] [PubMed]
42. Australian Government. Guidelines for preparing assessments for the Medical Services Advisory Committee. 2021.
43. Janssen MF, Birnie E, Bonsel GJ. Quantification of the level descriptors for the standard EQ-5D three-level system and a five-level version according to two methods. Qual Life Res 2008;17:463-73. [Crossref] [PubMed]
44. Pilz MJ, Seyringer S, Hallsson LR, et al. The EORTC QLU-C10D is a valid cancer-specific preference-based measure for cost-utility and health technology assessment in the Netherlands. Eur J Health Econ 2024;25:1539-55. [Crossref] [PubMed]
45. Pan CW, He JY, Zhu YB, et al. Comparison of EQ-5D-5L and EORTC QLU-C10D utilities in gastric cancer patients. Eur J Health Econ 2023;24:885-93. [Crossref] [PubMed]
46. Döhmen A, Obbarius A, Kock M, et al. The EORTC QLU-C10D distinguished better between cancer patients and the general population than PROPr and EQ-5D-5L in a cross-sectional study. J Clin Epidemiol 2025;177:111592. [Crossref] [PubMed]
47. Chung KC, Muthutantri A, Goldsmith GG, et al. Symptom impact and health-related quality of life (HRQoL) assessment by cancer stage: a narrative literature review. BMC Cancer 2024;24:884. [Crossref] [PubMed]
48. Pourrahmat MM, Kim A, Kansal AR, et al. Health state utility values by cancer stage: a systematic literature review. Eur J Health Econ 2021;22:1275-88. [Crossref] [PubMed]
49. Courrech Staal EF, van Sandick JW, van Tinteren H, et al. Health-related quality of life in long-term esophageal cancer survivors after potentially curative treatment. J Thorac Cardiovasc Surg 2010;140:777-83. [Crossref] [PubMed]
50. Rutegård M, Lagergren J, Rouvelas I, et al. Population-based study of surgical factors in relation to health-related quality of life after oesophageal cancer resection. Br J Surg 2008;95:592-601. [Crossref] [PubMed]
51. Kidane B, Ali A, Sulman J, et al. Health-related quality of life measure distinguishes between low and high clinical T stages in esophageal cancer. Ann Transl Med 2018;6:270. [Crossref] [PubMed]
52. Housman B, Flores R, Lee DS. Narrative review of anxiety and depression in patients with esophageal cancer: underappreciated and undertreated. J Thorac Dis 2021;13:3160-70. [Crossref] [PubMed]
53. Howard GS, Dailey PR. Response-shift bias: A source of contamination of self-report measures. J Appl Psychol 1979;64:144-50.
54. Ilie G, Bradfield J, Moodie L, et al. The Role of Response-Shift in Studies Assessing Quality of Life Outcomes Among Cancer Patients: A Systematic Review. Front Oncol 2019;9:783. [Crossref] [PubMed]
55. Sprangers MA, Schwartz CE. Integrating response shift into health-related quality of life research: a theoretical model. Soc Sci Med 1999;48:1507-15. [Crossref] [PubMed]