Patient preferences for stages II–IV melanoma treatments in the UK: results from a cross-sectional study

Abstract

Aim

To understand patient preferences for treatment attributes in adjuvant and metastatic melanoma in the UK.

Materials and Methods

Patients with stages II–IV melanoma completed an online survey from November 2022 to February 2023 comprising two discrete choice experiments (DCE) exploring efficacy, safety, and dosing regimen.

Results

Out of 104 respondents, 100 and 102 patients completed DCEs for adjuvant (Adj-DCE) and metastatic (Met-DCE) settings, respectively. 4-year relapse-free survival (RFS) for Adj-DCE (33%), and 4-year overall survival (OS) for Met-DCE (40%) had the highest relative attribute importance (RI), followed by risk of grade ≥3 adverse events (29% versus 26%, respectively). Patients with advanced melanoma or BRAF mutation had significantly higher RI for RFS and OS.

Conclusion

Patients, especially those with advanced melanoma or BRAF mutation, perceive efficacy as the most important factor in determining choice of treatment.

ARTICLE HIGHLIGHTS

• In the UK, melanoma is the 5^th most common cancer with around 19,921 newly diagnosed cases in 2022. Fortunately, treatment options for patients with melanoma have expanded significantly over the past few years. Each of these new therapies carries a different benefit-to-risk profile and varies with regard to different aspects of patient convenience.

• The aim of this study was to understand patient preferences for adjuvant and metastatic treatment attributes in stage II/III and stage IV melanoma in the UK and to determine predictive factors for treatment preferences in different subpopulations.

• A discrete-choice experiment (DCE) was conducted via an online quantitative treatment preference survey to elicit patients’ relative preferences for treatment attributes, including safety, dosing regimen, and efficacy for adjuvant and metastatic settings, respectively.

• Analysis of the DCE data using Hierarchical Bayesian (HB) model revealed distinct preferences among the 104 survey respondents.

• Efficacy, as assessed by 4-year RFS in the adjuvant setting and 4-year OS in the metastatic setting, was the most important attribute, followed by severe adverse events (AEs) and dosing regimen.

• Immune checkpoint inhibitors (ICI) were consistently ranked as the preferred options across all patient segments. Based on the hypothetical drug profile presented in the DCE, the fixed dose combination of nivolumab and relatlimab was the preferred regimen in the metastatic disease setting while either pembrolizumab or nivolumab were preferred in the adjuvant setting.

• Our study highlights the importance of considering patients’ prioritization of treatment efficacy as the primary factor when making decisions about melanoma care, both in the adjuvant and metastatic settings. Although efficacy and safety were attributed to have relatively similar importance for treatment preference in the adjuvant setting, patients with more advanced disease or those carrying the BRAF mutation placed heightened emphasis on treatment efficacy.

1. Introduction

Malignant cutaneous melanoma, hereafter referred to as melanoma, is a neoplasm originating from the malignant transformation of melanocytes located in the basal layer of the epidermis [1]. Melanoma is among the 10 most common cancers reported by the American Cancer Society in the US, and its incidence, which increased by 2–3% annually from 2015 to 2019 [2], is expected to reach 99,700 new cases in 2024 in the US [2] and to increase to 2.56 million individuals in the EU population [3]. In the United Kingdom (UK), melanoma is the 5^th most common cancer [4,5] with around 16,200 new diagnoses and 2,300 deaths in 2017 [6]. BRAF mutations are present in the tumors of roughly 50% of patients with melanoma [2,7].

Fortunately, over the last 10 years several new treatment options have been approved [8]. Treatment management depends on the disease stage, genetic profile, patient characteristics, overall health, and treatment goal. The introduction of immune checkpoint inhibitors (ICI) and BRAF/MEK inhibitors has provided new options for patients and improved outcomes. Each of these new therapies carries a different benefit-to-risk profile and varies with regards to dosing regimen and potential adverse events (AEs) [8–10].

In the UK, BRAF/MEK inhibitors, such as dabrafenib/trametinib [11] or binimetinib/encorafenib [12], are reimbursed by the National Institute for Health and Care Excellence (NICE) for patients with BRAF mutations. Both combinations are approved as first line treatments, as alternatives to ICI, in BRAF mutation-positive patients with unresectable stage III and metastatic melanoma [13–17]. Dabrafenib with trametinib is also approved for resected stage III BRAF mutation-positive melanoma [11]. These drugs are administered orally twice or once daily (BID/QD) and are generally well tolerated, although approximately 50% of patients are likely to experience grade 3/4 AEs [13–16].

In the UK, NICE recommended ICI for the treatment of melanoma include the PD-1 class (nivolumab; pembrolizumab), the CTLA-4 class (ipilimumab), and the LAG-3 class (relatlimab). Ipilimumab [17,18], and combination of nivolumab with either ipilimumab [17–20] or relatlimab [21,22], are approved for the treatment of advanced (unresectable or metastatic) melanoma. Nivolumab as a monotherapy is approved for the treatment of completely resected melanoma with lymph node involvement and advanced melanoma [19,23]. Pembrolizumab has a larger indication and is approved for the treatment of resected stage IIb/IIc/III and advanced melanoma [24,25]. These drugs are administered intravenously in an outpatient hospital setting, with a dosing schedule ranging from every 2 to 6 weeks (Q2W to Q6W), depending on the treatment.

PD-1 class drugs, given as monotherapy or, as per standard-of-care, in combination with a CTLA-4 inhibitor, have shown efficacy in both BRAF mutant and wild-type patients [26–28].

Interestingly, the randomized controlled trial RELATIVITY-047 [29] showed that a fixed-dose combination of nivolumab and relatlimab demonstrated superior progression-free survival (PFS) benefit compared to nivolumab monotherapy without any new safety signals compared to CTLA-4/PD-1 [26].

In BRAF mutation-positive patients, ICI and BRAF/MEK inhibitors are predominantly used as systemic therapies in clinical practice, both in the adjuvant and metastatic settings and have demonstrated significant improvements in survival outcomes. They are well established as standards of care for patients [9,10,26].

A key challenge in clinical practice, however, is understanding the right treatment sequence and combination strategy, which can be a complex decision that depends on multiple factors, including the stage of the disease, genetic mutations, patient characteristics and preference, potential AEs, route of administration (RoA) [30], and prior therapy received.

There is growing evidence from recent clinical studies, such as SECOMBIT [31] and DREAMseq [32], suggesting that patients with a BRAF mutation may have a better benefit from receiving the PD-1/CTLA-4 combination followed by BRAF/MEK inhibitors in the metastatic setting. In the DREAMseq trial, the sequence of PD-1/CTLA-4 as first line followed by BRAF/MEK inhibitors for patients with BRAF positive melanoma was associated with better 2-year OS than the inverse sequence [32–34].

Understanding the preferences of patients with melanoma is critical for fostering joint decision making and supporting physicians in developing personalized treatment plans. Previous research suggests that patients want to be more involved in the discussion around their treatment plan as well as ensuring that their preferences agree with their treating physicians [35]. Therefore, as patients and physicians face decisions regarding melanoma treatment options, it is important to understand patient preferences regarding medication attributes and to incorporate these into the treatment decision. This can help improve patient experiences, outcomes, and overall therapeutic effectiveness in real-world clinical practice. Additionally, it is important to identify areas of discrepancy between patients and physicians to provide better treatment and increase treatment satisfaction for patients [8].

To our knowledge, no previous study in the UK has assessed patient preferences in melanoma treatment. Therefore, by using a discrete-choice experiment (DCE), we aimed to understand patient preferences for attributes of clinical outcomes (such as efficacy, toxicity, dosing regimen) associated with treatments in the adjuvant (stages II–III) and metastatic (stage IV) melanoma settings, by describing differences in patient choices for therapy for both groups in the UK. We also aimed to explore how treatment preferences vary in different subpopulations. These results may provide valuable information to health authorities and physicians to help increase the emphasis on patient-centric care, facilitating shared decision-making and guiding optimal treatment choices for patients with melanoma, based on their preferences.

2. Materials and methods

2.1. Study design

A 30-minute online cross-sectional quantitative treatment preference survey was completed by patients between November 2022 and February 2023. The survey began with a few questions to confirm the patient’s eligibility. All information collected, including eligibility criteria, was self-reported by the patients. At the beginning of the survey, patients were asked to review the information letter and provide their opt-in consent.

Prior to initiating the quantitative survey, cognitive interview pretests were conducted with five patients, who were not included in the study dataset, to obtain feedback, and assess the interpretation, relevance, and comprehensiveness of the survey content. Based on the feedback and the learnings from those interviews, the survey was revised and updated as appropriate to become the final global quantitative preference survey for data collection.

This study and all relevant participant-facing documents received approval from the Oracle institutional review board on July 11, 2022. Participant anonymity was maintained throughout the study.

2.2. Population

This study included 104 adult patients (≥18 years) from the UK, diagnosed with stages II–IV melanoma (most advanced or current stage) as their primary cancer, either with active disease or in remission, as of September 2020. Patients were excluded if: they could not report their most advanced and current stage of melanoma; they were unable or unwilling to provide informed consent; or they were unable to have access to a computer. Patient respondents were classified as in the adjuvant setting or metastatic setting based on the self-reported answers to the stage of melanoma and the disease description statement.

Recruitment was done via a multi-channel direct-to-patient approach based on a mix of recruiters’ proprietary patient databases, referrals from physicians, patient association group recruitment (Melanoma UK and Melanoma Focus), and outreach via social media and online forums. This direct-to-patient approach was cost- and time-effective, allowing for a large sample size.

2.3. Data collection

After providing informed consent, participants completed two DCEs to explore patient preferences with different setting attributes; one in the adjuvant setting (stages II-III) and one in the metastatic setting (stage IV) melanoma. Each DCE included a series of choice tasks between two hypothetical treatment profiles comprising seven different attribute levels for clinical outcomes of efficacy (relapse-free survival (RFS) and overall survival (OS)), safety (AEs such as: nausea, diarrhea, pyrexia…), and dosing regimen (mode of administration and dosing schedule), from randomized clinical trials, as well as a no treatment/observation option (just for the adjuvant setting) (Supplementary Tables 1 and 2). Respondents were asked to choose which hypothetical treatment they preferred across a series of choice tasks, 14 for the adjuvant DCE and 13 for the metastatic DCE, and responses to these questions were used to estimate preference weights for each attribute level. Attributes were selected based on previous qualitative research, and levels were selected based on clinical trials of treatments included in a competitive set for both the adjuvant and metastatic settings to ensure having the right profiles of these treatments (Supplementary Tables 3 and 4). The competitive set in the adjuvant setting comprised nivolumab, pembrolizumab, the combination dabrafenib/trametinib. The competitive set in the metastatic setting comprised nivolumab, pembrolizumab, dabrafenib/trametinib, encorafenib/binimetinib, and the combinations nivolumab/ipilimumab and nivolumab/relatlimab. The hypothetical treatments had their attribute levels shuffled according to an experimental design. No actual treatments were named in the survey. The levels for the attributes in the DCEs were first rated on five-point Likert scales to familiarize the patients with the levels. An example choice set is presented in Supplementary Table 5. Additional information such as sociodemographic characteristics, clinical characteristics, and treatment management were collected. Patients were considered as BRAF positive if they self-reported a positive test for BRAF mutation or if they reported that they received BRAF/MEK inhibitors.

2.4. Sample size

Overall, 104 patients completed the treatment preference survey. However, quality checks revealed that a few patients provided low-quality data in either the adjuvant or metastatic DCE (or both), leading to their exclusion from the respective DCE analyses. As a result, 100 patients were included in the analysis of the adjuvant DCE, and 102 in the metastatic DCE analysis.

2.5. Statistical analysis

All data were collected using a programmed web-based survey. Lighthouse studio from Sawtooth Software version 9.14.2 was used for DCE design and analysis, while Daisy Software version 2.4.95 was used for other variables. A descriptive analysis was done for all study variables using counts and percentages for categorical variables and means, standard deviations, median and range for continuous variables. Hierarchical Bayesian (HB) logistic regression was used to estimate preference weights for each attribute level and relative attribute importance (RI). The parameter estimates (preference weights) from the HB analysis enabled the calculation of the conditional RI of each attribute. The underlying choice-probability model in the HB model was conditional logit, using effects coding for the attribute levels. The model assumed that the preferences were normally distributed. Subgroup analyses were conducted to assess factors potentially influencing RI, using Student’s t-test for comparing continuous variables between two groups, and ANOVA for comparisons across three or more groups. Preferences of patients for both settings were compared. A p-value <0.05 was considered statistically significant.

2.6. Treatment preference score calculation

Relative treatment preference scores for the different available treatments were estimated according to the preference weights and the clinical outcomes collected from the literature. The total preference scores for nivolumab, pembrolizumab, and dabrafenib/trametinib in the adjuvant, as well as nivolumab, pembrolizumab, nivolumab/relatlimab, nivolumab/ipilimumab, dabrafenib/trametinib, and encorafenib/binimetinib in the metastatic setting, were calculated. Initially, each therapy was characterized based on attribute levels derived from their respective clinical trials (Supplementary Tables 3 and 4) [36–40]. In the case of the RELATIVITY-047 trial [29], data were not mature enough to provide an estimate of the 4-year OS for nivolumab/relatlimab, therefore we employed the survival curve trends from CheckMate-067 [26] to extrapolate the 4-year OS. Then, the preference weights associated with the attribute levels for each therapy was summed. For those cases where the treatment outcome had not been explicitly included as a level, its preference weight was measured through a linear interpolation, given that the levels agreed included both a minimum and a maximum value.

3. Results

3.1. Study participant characteristics

The characteristics of the 104 study participants who completed the treatment preference survey between November 2022 and February 2023 are shown in Table 1. The mean age was 51.9 ± 11.2 years, 49% were female and more than 70% of them completed high school and had a partner. Interestingly, 86.5% of the population was fully or part-time employed before their melanoma diagnosis compared to 57.7% who were employed at the time of completing the survey (data not shown). Patients in the study were distributed amongst living in urban areas (46.2%), followed by suburban areas (30.7%). At the time of the study, 37.5% of patients were cancer free, on average within the 11 months prior to completing the survey (maximum of 28 months). The proportion of patients with stage III and metastatic melanoma at the time of the survey was 26.9% for each stage respectively. Almost half of participants reached stage III melanoma, while around one third of patients reached stage IV melanoma, as their most advanced cancer stage. BRAF mutation was estimated in 51% of patients, which was the most reported biomarker (Table 1). Out of the one third of patients who, at the time of survey, had melanoma that reached a metastatic stage (n = 34), 29 were first diagnosed with a localized cancer that then spread, and the remaining five were first diagnosed at a metastatic stage. Three quarters of the sample were diagnosed with melanoma within the past 5 years.

Table 1.

Sociodemographic and disease-related characteristics of the study participants (N = 104).

Demographic characteristics
Age at diagnosis, years
Mean (SD)	51.9 (11.2)
Median (Q1-Q3)	52.0 (44–61.3)
Min-Max	26–77
Sex, N (%)
Male	53 (51.0)
Female	51 (49.0)
Marital status, N (%)
Single/never married	13 (12.5)
Committed relationship/married	75 (72.1)
Separated/divorced	13 (12.5)
Widowed	2 (1.9)
Education, N (%)
Less than high school	10 (9.6)
High School	36 (34.6)
College/University	42 (40.4)
Postgraduate	16 (15.4)
Current employment status, N (%)
Full time	44 (42.3)
Part time	16 (15.4)
Temporary leave of absence	8 (7.7)
Not employed	7 (6.7)
Retired	24 (23.1)
Other	5 (4.8)
Living area, N (%)
Major metropolitan area1	26 (25.0)
Urban area2	22 (21.2)
Suburb of a large city3	20 (19.2)
Small city4	12 (11.5)
Rural or small town5	24 (23.1)
Clinical characteristics
Current stages, N (%)
Stage II	4 (3.8)
Stage III	28 (26.9)
Stage IV	28 (26.9)
Cancer-free	39 (37.5)
Unknown	5 (4.8)
Most advanced stage, N (%)
Adjuvant
Stage II	19 (18.3)
Stage III	51 (49.0)
Metastatic
Stage IV	34 (32.7)
Time since diagnosis, months
Mean (SD)	43.5 (28.6)
Median (Q1-Q3)	38.5 (21.8–58.5)
Min-Max	2–138
BRAF status, N (%)
Mutant	53 (51.0)
Wild-type	10 (9.6)
Unknown	41 (39.4)
Treatment management
Surgery, N (%)
Yes	100 (96.2)
No	4 (3.8)
Number of surgeries to remove melanoma
Mean (SD)	1.8 (0.96)
Median (Q1-Q3)	2 (1.0–2.0)
Min-Max	1.0–6.0
Treatment ever received, N (%)
Yes	97 (93.3)
No	7 (6.7)
Type
Radiotherapy	29 (29.9)
Chemotherapy	42 (43.3)
ICI	42 (43.3)
BRAF/MEK inhibitors	35 (36.1)
Interferon therapy	6 (6.2)
Others	3 (3.0)

ICI = Immune checkpoint inhibitors; SD = Standard Deviation; min = minimum; max = maximum.

¹population >500,000; ²population between 100,000 and 500,000; ³population >100,000; ⁴population between 30,000 and 100,000; ⁵population <30,000.

In terms of treatment, 96.2% of the patients had resection surgery for their melanoma, with a mean of 1.8 surgeries per patient, and 75% had their last surgery in the past 2 years and 7 months. Medical treatment was received by 93.3% of patients; more than half received adjuvant treatment, and 43% peri-adjuvant treatment (data not shown). Patients reported ICI and chemotherapy as the most received treatments (43.3% for both), followed by BRAF/MEK inhibitors (36.1%) and then radiotherapy (29.9%) (Table 1). For patients with metastases diagnosed with a localized melanoma, anticancer treatments were mostly administered after becoming metastatic.

Travel-time to a cancer facility was reported to be between 30 and 60 minutes for most of the population (56.7%). Around 70% of the population did not perceive traveling to the hospital as a financial burden; 22.1% were neutral toward it.

3.2. Adverse events

When asked about AEs, the most reported AE was nausea (69.2%), followed by diarrhea (50.0%). Less than 10% reported never experiencing any AE (Figure 1(A)). Around 56% of the population reported having an AE that was difficult to manage, with nausea and diarrhea being the most common ones (26.9% and 21.2% respectively) (Figure 1(B)). Moreover, 41% of participants reported to have been hospitalized due to AEs (data not shown).

Figure 1.

Side effects associated with Melanoma treatments of the study participants (N = 104). (A) Side effects ever experienced. (B) Side effects difficult to manage.

3.3. DCE findings in adjuvant and metastatic settings

Of 104 patients recruited, 100 and 102 patients completed a DCE for treatment in the adjuvant and metastatic setting, respectively.

Responses to the DCE led to the preference weights (Supplementary Figure 1(A,B)). A consistent pattern was observed across both adjuvant and metastatic setting preference weights. There was a proportional relationship between the best outcomes and the highest preference weights within each attribute for the ordinal variables (like relapse-free survival (RFS)/OS, risk of nausea, etc…). Regarding the categorical attribute, dosing regimen, “pills taken twice a day indefinitely” was the most preferred attribute compared to the rest of the options. Interestingly, an “intravenous (IV) every 4 weeks” had a higher preference than “every 6 weeks” and the latter was more preferred than “every 3 weeks for 4 doses and then every 12 weeks” (Supplementary Figure 1(A,B)).

The adjuvant DCE setting revealed that the 4-year RFS was the most important attribute (32.7%), followed closely by the risk of grade ≥3 AEs (28.6%), then by dosing regimen (15.9%) and the risk of each specific AE, with hypothyroidism being the least important (2.4%) (Figure 2(A)). Patients aged 50 years or older attributed more importance to the risk of grade ≥3 AEs than those who were younger than 50 years old (30.8% vs. 25.6%) (Figure 3(A)). When comparing patients based on their most advanced melanoma setting, those who had metastatic melanoma attributed much importance to 4-year RFS than those treated in the adjuvant setting (41.8% vs. 28.5%). Contrarily, grade ≥3 AEs, dosing regimen and pyrexia were considered more important for patients in the adjuvant DCE setting (Figure 3(B)). Patients with a BRAF positive melanoma considered 4-year RFS as more important than patients whose status was negative or unknown (35.5% vs. 29.7%) (Figure 3(C)). Patients who have never been hospitalized due to side effects attributed greater importance on risk of grade ≥3 AEs (32.5% vs. 23.3%), while hospitalized patients prioritized 4-year RFS (26.5% vs. 41.3%).

Figure 2.

Patient relative importance of attributes across the study sample. (A) For adjuvant treatment. (B) For metastatic treatment. Bars represent the 95% CI for each attribute.

Figure 3.

Patient relative importance of attribute for adjuvant treatment and influence of sociodemographic and disease-related characteristics. (A) Relative importance of treatment attributes across the study sample and by age. (B) Relative importance of treatment attributes across the study sample and by most advanced setting. (C) Relative importance of treatment attributes across the study sample and by patient BRAF status. * Indicates differences that are significantly superior (p < 0.05) vs the compared group. AE = Adverse Event; RFS = Relapse Free Survival.

The metastatic DCE setting had a similar pattern to that in the adjuvant setting, revealing that 4-year OS was the most important attribute (40.3%), followed by the risk of grade ≥3 AEs (25.9%) and dosing regimen (15.3%) (Figure 2(B)). Patients younger than 50 years old attributed more importance to the dosing regimen than their older counterparts (18.3% vs. 13.0%). While those who were ≥50 years considered the risk of grade ≥3 AEs as more important than their younger counterparts (29.1% vs. 21.7%) (Figure 4(A)). Patients with metastases put more weight on 4-year OS than patients treated in the adjuvant setting (50.4% vs. 35.5%), while the latter group attributed more importance to the risk of grade ≥3 AEs and dosing regimen (Figure 4(B)). The 4-year OS attribute was more important for BRAF positive patients than those with a negative or unknown status (44.2% vs. 36.2%). Patients with BRAF negative/unknown status attributed more importance to the risk of grade ≥3 AEs than those who were BRAF positive (28.5% vs. 23.3%) (Figure 4(C)). Patients who had never been hospitalized due to side effects attributed greater importance on risk of grade ≥3 AEs (29.7% vs. 20.3%), while hospitalized patients prioritized 4-year RFS (35.7% vs. 47.2%).

Figure 4.

Patient relative importance of attribute for metastatic treatment and influence of sociodemographic and disease-related characteristics. (A) Relative importance of treatment attributes across the study sample and by age. (B) Relative importance of treatment attributes across the study sample and by most advanced setting. (C) Relative importance of treatment attributes across the study sample and by patient BRAF status. * Indicates differences that are significantly superior (p < 0.05) vs the compared group. AE = Adverse Event; RFS = Relapse Free Survival.

In both DCEs, exploratory subgroup analyses (data not shown) indicated that the relative importance of efficacy was lower among patients currently in remission compared to those who were not. Moreover, patients who reported having received chemotherapy assigned greater importance to grade ≥3 AEs and dosing regimen, and less importance to efficacy, compared to those who had not received chemotherapy. In contrast, patients who had received immunotherapy placed greater emphasis on efficacy and lower importance on grade ≥3 AEs and dosing regimen than those who had not received immunotherapy.

3.4. Relative treatment preference scores

In the adjuvant DCE setting, ICI monotherapies, either nivolumab or pembrolizumab, were the most preferred therapies (with mean preference of 181.6 ± 3.6 and 176.9 ± 6.7 respectively), versus BRAF/MEK inhibitors available in this setting such as dabrafenib/trametinib, regardless of the subgroups (BRAF mutation, patient setting, and melanoma stages). In the metastatic DCE setting, dual immune checkpoint fixed-dose combination of nivolumab and relatlimab was the most preferred therapy (with a mean preference of 132.6 ± 7.3), while BRAF/MEK inhibitors being the least preferred, regardless of the subgroups (Table 2).

Table 2.

Relative treatment preference scores

Adjuvant DCE (N = 100)
Treatment’s type		Overall (N = 100)	Subgroup analysis
			BRAF mutation		Treatment setting		Disease stage
			BRAF mutant (N = 52)	BRAF wild type or unknown (N = 48)	Adjuvant (N = 68)	Metastatic (N = 32)	Stage II (N = 18)	Stage III (N = 50)
ICI	Nivolumab	181.6 (3.6)	176.9 (5.1)	186.6 (4.9)	182.1 (3.9)	180.5 (7.5)	181.7 (17.2)	182.3 (4.7)
	Pembrolizumab	176.9 (6.7)	181.9 (9.3)	171.5 (9.8)	165.0 (7.6)	202.3 (12.6)	137.5 (13.9)	174.9 (8.7)
BRAF/MEK inhibitor	Dabrafenib/ Trametinib	124.4 (5.8)	118.4 (7.9)	130.9 (8.4)	124.2 (6.9)	124.9 (10.7)	148.1 (11.9)	115.6 (8.1)
No treatment	No treatment	−93.3 (18.9)	−126.3 (25.3)	−57.5 (27.7)	−43.5 (22.5)	−199.1 (26.7)	–	–

Metastatic DCE (N = 102)
Treatment’s type		Overall (N = 102)	Subgroup analysis
			BRAF mutation		Treatment setting
			BRAF mutant (N = 52)	BRAF wild type or unknown (N = 50)	Adjuvant (N = 69)	Metastatic (N = 33)
ICI	Nivolumab/Relatlimab	132.6 (7.3)	142.3 (10.3)	122.5 (10.2)	122.8 (8.4)	153.1 (13.8)
	Nivolumab	76.2 (5.7)	84.3 (7.9)	67.7 (8.0)	72.0 (6.5)	84.8 (11.1)
	Pembrolizumab	33.2 (2.5)	31.0 (2.8)	35.5 (4.2)	30.9 (3.2)	38.1 (4.1)
	Nivolumab/Ipilimumab	20.9 (13.7)	42.3 (20.8)	−1.3 (17.5)	−20.0 (15.0)	106.5 (22.5)
BRAF/MEK inhibitors	Dabrafenib/Trametinib	−57.6 (9.1)	−66.4 (13.5)	−48.6 (12.1)	−33.3 (11.0)	−108.6 (12.5)
	Encorafenib/Binimetinib	−94.0 (7.5)	−88.7 (9.7)	−99.6 (11.5)	−86.6 (10.1)	−109.5 (9.2)

Data are given as Mean (SE). DCE = Discrete Choice Experiment; ICI = Immune checkpoint inhibitors.

Stages II and III for adjuvant patients only, BRAF mutation includes both adjuvant and metastatic patients.

4. Discussion

To our knowledge, this is the first DCE study in the UK examining patient preferences regarding systemic therapies for melanoma, which involved attributes covering three treatment aspects: clinical outcomes, risk of toxicities and dosing regimen. By considering patient preferences and incorporating them into treatment decision-making, physicians can enhance patient satisfaction, treatment adherence and, therefore, overall treatment outcomes.

Findings from this study indicate that patients prioritize efficacy as the most crucial factor when making decisions about their melanoma care, whether in an adjuvant or a metastatic setting. In the adjuvant setting, however, relative efficacy consideration held similar importance as safety. These results are in line with other similar studies that have shown that efficacy was the most important consideration to patients [41–44], followed by AEs [41]. Beusterien et al. reported that in the adjuvant melanoma setting, patients and physicians tend to agree on their preferences. They favor treatment over no treatment and consider potential efficacy to be more significant than the risks of toxicities [44]. In the metastatic melanoma setting, Liu et al. reported that there was a slight switch in importance weight between nurses and patients regarding objective response rate and AEs (i.e., metastatic melanoma patients gave more importance to AEs than objective response rate). Nonetheless, the difference was not significant, and the highest preference rate was for overall survival for both nurses and patients, showing that the perceptions of healthcare providers and patients were well aligned [41]. Melanoma remains a potentially lethal disease, and clinical decision-making in such contexts often places a strong emphasis on efficacy and survival outcomes, such as OS and RFS. However, while these outcomes are crucial, there is suggestive evidence that treatment decisions reflect a shared decision-making model that respects patient preferences, values, and quality of life considerations [45].

In this study, safety was consistently considered the second most valued attribute across all patient segments and thus plays a critical role in patient preference. This was also consistent with a systematic review by Livingstone et al. wherein it was reported that impaired quality of life due to AEs appeared to be a second-order consideration [43]. The risk of grade ≥3 AEs consistently predominated among all safety attributes and across all subgroups, followed by diarrhea. There may be an innate fear component associated with grade ≥3 AEs that seems to be minimized once patients have experienced them by gradually adapting and developing coping mechanisms. Patient-physician communication here is crucial, as it allows healthcare providers to discuss potential AEs and the signs and symptoms to be aware of, highlighting that, while each patient will respond differently to treatment, generally AEs typically resolve over time as patients continue their treatment.

Dosing regimen was consistently the third most important attribute amongst the adjuvant and metastatic melanoma settings. From a patient point of view, it seems that dosing regimen preference was predominantly driven by RoA rather than dosing schedule/frequency of administration. Oral formulation was the most preferred RoA but was unlikely to drive preference alone for BRAF/MEK inhibitors over intravenous ICI regimens. However, although not statistically significant, patients seemed to prefer their injectable treatment to be received at least every 4 weeks (Q4W) rather than longer (Q6W) for ICI, particularly in the adjuvant setting. A subcutaneous (SC) formulation that enables a quick administration of ICI at home, if/when available, may influence patient preference over an IV option and would be favored by healthcare systems and enable them to provide faster care to melanoma patients [46].

When considering the subgroup analyses, patients with advanced disease showed a particularly strong desire to prioritize treatment efficacy over safety, compared with those in earlier disease stages. This preference may be attributed to the fact that patients in the metastatic stage have potentially faced more challenging prognoses and are likely seeking treatments with the highest potential to combat their advanced cancer. In contrast, patients in early disease stages may place almost equal importance on safety and efficacy, possibly because they want to ensure a more comfortable treatment experience while still benefiting from effective therapy.

Patients with mutant BRAF tumors also face a poorer prognosis [47] and an increased likelihood of recurrence [48], thus emphasizing the importance of prioritizing efficacy. Additionally, patients with BRAF positive melanoma may be more informed around treatment evidence as they have more choice in available treatment options, and they may have had these discussions with their oncologist. Patients aged ≥50 years attributed more importance to the risk of AE grade ≥3 than those aged <50 years old, who attributed more importance to the dosing regimen in both settings. Younger people may be more fit and willing to accept the risk of AEs in exchange for other attributes such as a more convenient dosing regimen, whereas for older people there may be a greater fear of experiencing an AE that leads to hospitalization.

In both the adjuvant and metastatic DCE, those who have never been hospitalized due to AEs put a greater importance on the risk of AE grade ≥3 than those who have been hospitalized. This could be explained due to a fear of hospitalization compared to those who have gone through the experience, while the latter might be more willing to accept the risk of hospitalization in exchange for a higher efficacy outcome, based on past experiences.

Regarding treatment preference, ICI were consistently ranked as the preferred options across all patient segments in this study. The fixed dose combination of nivolumab and relatlimab was the preferred regimen in the metastatic disease setting, while either pembrolizumab or nivolumab was preferred in the adjuvant setting. This may be due to the higher efficacy along with a well-tolerated, generally manageable safety profile of these treatments [49,50]. This study showed that the combination of nivolumab and ipilimumab was the least preferred option among ICI in the metastatic setting, which is mostly driven by a lower safety profile that is characterized by a higher incidence of grade ≥3 AEs compared to other ICI [26]. However, it should be noted that the combination of nivolumab and ipilimumab was preferred over nivolumab or pembrolizumab monotherapy in the subgroup of 33 patients who had received treatment in the metastatic setting. The gap in preference for ICI vs BRAF/MEK inhibitors narrows in the adjuvant setting. Patient preferences toward immunotherapy may be influenced by strong public and media narratives that portray these treatments as revolutionary or highly effective, often referred to as ‘breakthroughs’ in cancer care. Such perceptions may create bias that is not entirely grounded in clinical evidence. This influence is vital, as it may shape patient expectations and decision-making in ways that diverge from evidence-based recommendations.

4.1. Strengths and limitations

The results of this study should be interpreted within the context of the strengths and limitations. The utilization of the well-established DCE method to explore treatment preferences for patients with melanoma in the UK constitutes a notable strength in this study. The attractiveness of a treatment to individual patients depends on their relative preferences for the attributes expressed and their willingness to accept tradeoffs amongst these attributes. Although there are alternatives to eliciting patient preferences (e.g., revealed preferences, direct ratings/rankings, etc.), DCE is one of the most common approaches for assessing preferences in a healthcare context and has been increasingly applied in healthcare decisions. Recruitment through a multi-channel direct-to-patient approach taken from four different sources helps ameliorate the generalizability of our results. In addition, the design was chosen to estimate the preference weights efficiently and without bias.

This study presents several limitations. Selection bias may be present, as an online patient-reported survey is likely to under-represent people without access to or comfort with online surveys, which may explain why our sample was relatively younger compared to the typical average age of a patient with melanoma as well as including fewer healthy elderly patients, institutionalized patients, and those with the most severe comorbidities and disabilities. Due to the cross-sectional nature of the study, recall bias cannot be excluded. For example, the fact that 43% of patients stated that they had been treated with chemotherapy is unlikely to be true as it is rarely used. It is possible that patients confused chemotherapy with ICI due to its IV mode of administration. Additionally, a general limitation of the DCE methodology involves patients choosing between hypothetical treatment profiles with a limited set of attributes that may not reflect all aspects of a treatment that can influence patient preference. Some attributes had to be neglected to limit the complexity of the choice tasks, which may result in omitted variable bias. However, this risk was mitigated by the fact that the most important attributes were selected in a previous qualitative study and validated through pretests. Choices were intended to simulate possible clinical decisions but obviously do not have the same clinical, financial, or emotional consequences of actual decisions. Thus, differences can arise between stated and actual choices. As such, results from this study may not reflect real-world treatment decisions, which could be influenced by other factors not captured in the survey, such as physician recommendation, patient quality of life, and treatment cost. To limit the potential hypothetical bias, we constructed choice questions that mimicked realistic clinical choices as closely as possible and mapped clearly into clinical evidence. Another limitation is that patients with resectable melanoma had to provide preferences for treatments used in the metastatic setting, and some patients with metastatic melanoma had to provide preferences for treatments used in the adjuvant setting, although they never had experience of being treated in that setting. However, they all had experience (except three) with systemic treatments, and could understand what the treatment attributes mean, which is the most important consideration in a DCE administration. Finally, the relatively limited sample size may restrict the generalizability of the findings and may have reduced the ability to detect differences across subgroups.

Although this study does not directly assess the alignment or divergence between physician and patient preferences, this remains a key element in shared decision-making. A limitation of this study is the lack of data on this, which is central to the process or shared decision-making. Future research should investigate how differences or concordance between physician and patient preferences influence decision-making processes, treatment adherence, and patient satisfaction, particularly in the context of life-threatening conditions such as melanoma.

5. Conclusion

Understanding patient preferences is key to foster joint decision making, enhance patient satisfaction and treatment adherence, and promote a more personalized and patient-centric care approach. Effective and transparent communication with patients, particularly about AEs, is critical to prevent building of beliefs and misconceptions that may discourage patients from receiving effective treatments.

Our study highlights the relative importance of multiple treatment attributes for melanoma patients. Notably, disease stage, prior experiences, and mutational status may influence relative perception of treatment related attributes. This may provide drug manufacturers and healthcare systems with a better understanding of which treatment attributes hold value for patients, aiding in supporting access to optimal treatment options for patients with melanoma in the UK.

Effective communication that is both clear and compassionate is fundamental to helping patients make informed treatment choices. In contexts where fear, hope, or public discourse may distort perceptions of treatment risks and benefits, clinicians play a vital role in helping patients interpret information realistically. This involves not only sharing evidence but also guiding patients through the emotional and cognitive aspects of decision-making, ensuring choices are grounded in understanding rather than assumption.

Funding Statement

This study was sponsored by Bristol Myers Squibb.

Author contributions

Conceptualization: All authors. Data curation: AS, AM and XG. Formal analysis: AM and XG. Funding acquisition: LV and SV. Investigation: AS, AM and XG. Methodology: All authors. Project administration: LV, SV and XG. Resources: All authors. Software: AS and XG. Supervision: LV, SV and XG. Validation: LV, SV, XG and GC. Visualization: All authors. Writing – original draft: AS, AM, AN and XG. Writing – review & editing: All authors.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Ethical disclosure

The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, informed consent has been obtained from the participants involved and recorded as part of the survey. This study and all relevant participant-facing documents received approval from the Oracle institutional review board on July 11, 2022. Participant anonymity was maintained throughout the study.