Quality-adjusted life years for HER2-positive, early-stage breast cancer using trastuzumab-containing regimens in the context of cost-effectiveness studies: a systematic review

ABSTRACT

Introduction

This study aims to provide a comprehensive assessment of economic and health-related quality of life (HRQoL) outcomes for human epidermal growth factor receptor 2 (HER2)-positive, early-stage breast cancer patients treated with trastuzumab-containing regimens, by focusing on both Incremental Cost-Effectiveness Ratios (ICERs) and quality-adjusted life years (QALYs).

Methods

A systematic search was conducted across PubMed, Embase, and Scopus databases without language or publication year restrictions. Two independent reviewers screened eligible studies, extracted data, and assessed methodology and reporting quality using the Drummond checklist and Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022), respectively. Costs were converted to US dollars (US$) for 2023 for cross-study comparison.

Results

Twenty-two articles, primarily from high-income countries (HICs), were included, with ICERs ranging from US$13,176/QALY to US$254,510/QALY, falling within country-specific cost-effectiveness thresholds. A notable association was observed between higher QALYs and lower ICERs, indicating a favorable cost-effectiveness and health outcome relationship. EQ-5D was the most utilized instrument for assessing health state utility values, with diverse targeted populations.

Conclusions

Studies reporting higher QALYs tend to have lower ICERs, indicating a positive relationship between cost-effectiveness and health outcomes. However, challenges such as methodological heterogeneity and transparency in utility valuation persist, underscoring the need for standardized guidelines and collaborative efforts among stakeholders.

Registration

PROSPERO ID: CRD42021259826.

KEYWORDS:

• Early-stage breast cancer
• economic evaluation
• HER2-positive
• Qalys
• systematic review
• trastuzumab

1. Introduction

Breast cancer is the most widely diagnosed cancer globally, with human epidermal growth factor receptor 2 (HER2) amplification observed in about 20% of patients in the early stages of the disease [1–3]. The presence of HER2 is associated with aggressive breast cancer, with patients typically having a higher risk of disease recurrence and a lower overall survival (OS) rate [4,5]. The progression of the disease and its long-term treatment significantly impacts the health-related quality of life (HRQoL) [6–9]. The goals of the treatment in the early stage are to prevent the recurrence of the disease and to improve the long-term clinical outcome of the patients [10]. Trastuzumab, combined with chemotherapy for HER2-positive early-stage breast cancer, has been proven to significantly reduce the risk of recurrence, improve survival, and have a favorable effect on HRQoL [11–13].

Given the importance of cost-effectiveness analyses [14], economic evaluations of breast cancer treatments and of shifts in treatment paradigms are essential. As trastuzumab often entails higher acquisition costs and substantial expense, it is imperative to evaluate its cost-effectiveness to inform financing decision for policymakers [15]. Furthermore, the use of adjuvant trastuzumab therapy adds to the overall cost burden by necessitating additional cycles of chemotherapy administration and monitoring for potential cardiotoxic effects [15].

The quality-adjusted life year (QALY) is an important measure in economic evaluations, serving as an indicator of health outcomes [16]. It offers a comprehensive perspective by integrating both the quantity and quality of life into a single metric [16]. Incorporating both the length of life and the health-related quality of life (HRQoL), it provides a more holistic understanding of the impact of treatments or interventions, influencing policy formulation and decision making [16,17]. Hence, the methodological robustness and quality of the input parameters used in economics evaluations are important considerations [18]. The sources and quality of utilities used are particularly relevant in valuations of oncology management since patients’ quality of life can potentially impact significantly more on measurements of QALYs than on potential moderate benefits in relation to life expectancy [19]. The growing number of studies that have assessed the importance of economic evaluation is not, however, supported by the attention on the methods and sources for quantifying health state utilities and QALYs [20]. Therefore, it is important to foreground and evaluate approaches for determining, identifying, and analyzing health state utility to ensure that methods are performed with minimal bias and to reduce the misallocation of resources [21]. Given the importance of cost-effectiveness analyses and the pivotal role of QALYs as a measure of health outcomes, it is imperative to also consider the Incremental Cost-Effectiveness Ratios (ICERs) in evaluating the economic impact of trastuzumab-containing regimens for HER2-positive, early-stage breast cancer. By focusing on both ICERs and QALYs, this study aims to provide a comprehensive assessment of economic and health-related quality of life outcomes in this patient population.

2. Methods

The protocol for this systematic review has been registered in Prospero under the ID, CRD42021259826. This study was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline and checklist (Appendix, Table A).

2.1. Eligibility criteria

Published articles in any language and without limitations on the year of publication were considered in this systematic review. In summary, this systematic review encompasses health economic evaluations reporting both QALYs and ICERs associated with trastuzumab treatment. All selected articles focused on female participants with HER2-positive, early-stage breast cancer, and the intervention groups consisted of trastuzumab-containing regimens compared to regimens without trastuzumab. Articles reporting on the utility and measurement of QALYs, including sources and methods of measuring health utility, were also included. Additionally, articles reporting ICERs were included to ensure a comprehensive evaluation of economic outcomes. Review articles, unpublished articles that had not undergone peer review, and posters and abstracts for which full-text articles were unavailable were excluded from the review.

2.2. Search strategy and study selection

We conducted a comprehensive search of the literature published up to February 2023 in three electronic databases: PubMed, Embase, and Scopus, and applied the following search terms: breast cancer, trastuzumab, and quality-adjusted life years. The list of references was manually searched to identify any potential and relevant articles to be included in the review. The details of search strategy are presented in Appendix, Table B.

After duplicate articles had been removed, two reviewers (SMK and FDAS) screened the titles and abstracts of articles according to the study’s inclusion criteria. Subsequently, the full texts of the articles were reviewed until the final analysis of the included studies was assessed. Two reviewers reviewed the literature and extracted data independently, and any disagreements regarding their choices were resolved through discussion. The extracted data consists of the authors, publication year, country, time horizon, discounting, health states values, method, population, cited sources, overall survival, costs, ICER and QALYs. All costs and ICERs were adjusted into US dollars (US$) for the year 2023 using the Gross Domestic Product (GDP) deflator by taking into account the purchasing power parity (PPP) conversion rates from the International Monetary Fund [22]. The PRISMA flowchart was generated to record the systematic search process.

2.3. Quality assessment

The methodological quality assessment employed the 10-item Drummond checklist [17] which comprehensively covered several key aspects including the formulation of research questions, description of alternatives, evaluation of intervention effectiveness, identification and measurement of costs and outcomes, valuation of costs and outcomes, discounting procedures, incremental analysis, uncertainty analysis, and interpretation of results.

Additionally, as this systematic review incorporates health economic evaluation studies, it is essential to provide a comprehensive overview, adhering to established reporting standards. To assess the adequacy and transparency of reporting of all included studies, we employed the Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) [23]. The utilization of the CHEERS 2022 checklist ensures comprehensive and transparent reporting of health economic evaluations, thereby enhancing the quality and reliability of the findings presented in this study [23].

3. Results

3.1. Search structure

Our search of the selected databases yielded a total of 6,315 articles, of which 3,607 were eliminated as they were duplicates (Figure 1). The titles and abstracts of the remaining papers were independently screened by two reviewers, leaving 318 articles that were then screened for full-text availability. Following this screening, 94 studies remained, for which the full text of the articles was available. Of these articles, 22 health economic studies were ultimately deemed eligible for the review according to the study’s inclusion criteria. Most of the studies reported that the QALY values in the trastuzumab arms were higher than those in the control groups, with the range of values being 0.69 to 16.17. Two comparative studies in which a trastuzumab regimen was compared respectively with trastuzumab emtansine (T-DM1) [24] and neratinib [25] regimens reported lower scores for the quantification of QALYs in the trastuzumab groups. Table 1 presents summary details of the targeted studies.

Figure 1. Flowchart depicting the process of selecting studies for inclusion in the review.

Table 1. Summary details of the studies included in the review.

No	Author, year	Country	Time horizon	Discounting (%)	Age (years)	Trastuzumab group	Control group	QALYs value (trastuzumab)	QALYs value (control)
1	Garrison et al., 2007 [26]	U.S.A.	Lifetime	3	50	Doxorubicin 60 mg/m2 and cyclophosphamide 600mg/m2 every three weeks for four cycles. This is followed by paclitaxel, either at a dose of 175 mg/m2 every three weeks or 80 mg/m2 weekly for 12 weeks, alongside trastuzumab for the same duration. After the initial 12 weeks, trastuzumab is continued alone on a weekly basis for an additional 40 weeks	Anthracycline/taxane-based chemotherapy	11.78	10.08
2	Aboutorabi et al., 2014 [27]	Iran	20 years	3	50	The AC-T arm involves intravenous administration of 100 mg/m2 docetaxel, 60 mg/m2 doxorubicin, and 600 mg/m2 cyclophosphamide in each session, repeated six times every 3 weeks, extending over approximately 4 months. Trastuzumab is also included in the treatment regimen, but its duration is not specified	Anthracycline/taxane-based chemotherapy	11.98	11.11
3	Hall et al., 2011 [28]	UK	Lifetime	3.5	50	Trastuzumab	Anthracycline/taxane-based chemotherapy	7.71	7.22
4	Pichon-Riviere et al., 2015 [29]	Argentina	Lifetime	5	55	Trastuzumab (12 months)	No trastuzumab (chemotherapy regimen was not explained)	8.70	8.12
5	Gupta et al., 2020 [30]	India	Lifetime	3	>50	Trastuzumab is administered with adjuvant chemotherapy, with an initial infusion at 8 mg/kg for the first cycle, followed by infusions at 6 mg/kg for the remaining 16 cycles	Anthracycline/taxane-based chemotherapy	6.60	5.50
6	Younis et al., 2020 [24]	Canada	Lifetime	1.5	49	Trastuzumab 6 mg/kg every 3 weeks intravenously	T-DM1 3.6 mg/kg every 3 weeks intravenously	16.16	18.32
7	Hajjar et al., 2019 [31]	U.S.A.	Lifetime	3	40 and above 80	Adjuvant paclitaxel 80 mg/m^2 weekly for 3 months and trastuzumab 4 mg/kg on day 1 and then 2 mg intravenously per kg weekly for one year	Anthracycline/taxane-based chemotherapy	16.17	9.57
8	Schwartz et al., 2019 [25]	U.S.A.	Lifetime	3	18 and older	Adjuvant Trastuzumab	Neratinib	15.28	15.67
9	Purmonen et al., 2011 [32]	Finland	Lifetime	3	50	Trastuzumab for 1 year	No trastuzumab (chemotherapy regimen was not explained)	8.37	7.71
10	Gershon et al., 2019 [33]	Congo	Lifetime	3	≥45	Trastuzumab 8mg/kg followed by 16 doses of 6 mg/kg	Anthracycline/taxane-based chemotherapy	9.89	8.88
11	Genuino et al., 2019 [34]	Philippines	Lifetime	3.5	50	Doxorubicin 60 mg/m2 and cyclophosphamide 600 mg/m2 are administered on day 1 every 3 weeks for 4 cycles. This is followed by docetaxel 100 mg/m2 on day 1, along with an initial dose of trastuzumab at 8 mg/kg, then 6 mg/kg on day 1 every 3 weeks for 4 cycles. Finally, trastuzumab is continued at a dose of 6 mg/kg on day 1 every 3 weeks for an additional 14 cycles	Anthracycline/taxane-based chemotherapy	8.99	7.99
12	Kongsakon et al., 2019 [35]	Thailand	Lifetime	3	50	An initial dose of trastuzumab at 8 mg/kg intravenously, followed by subsequent doses of 4 mg/kg every 3 weeks for a duration of 1 year. Additionally, paclitaxel is administered intravenously at a dose of 175 mg/m2 every 3 weeks for a total of 6 cycles	Paclitaxel 175/m2 intravenously IV every 3 weeks for 6 cycles	5.01	4.59
13	Buendía et al., 2013 [36]	Colombia	Lifetime	5	50	Trastuzumab	Anthracycline/taxane-based chemotherapy	7.62	6.83
14	Kurian et al., 2007 [37]	U.S.A.	2 years	3	49	Doxorubicin 60 mg/m^2 plus cyclophosphamide 600 mg/m^2 are administered intravenously for four cycles every 3 weeks, followed by paclitaxel administered at 80 mg/m^2, with trastuzumab administered weekly; the initial dose of trastuzumab is 4 mg/kg, with subsequent doses of 2 mg/kg. Trastuzumab treatment continues weekly for a total duration of 1 year	Anthracycline/taxane-based chemotherapy	9.79	9.35
15	Seferina et al., 2017 [38]	The Netherlands	Lifetime	1.5	Not reported	Trastuzumab treatment	No trastuzumab (chemotherapy regimen was not explained)	13.93	13.10
16	Liberato et al., 2007 [39]	Italy	15 years	3	50	During the first quarter of adjuvant treatment, patients receive paclitaxel and trastuzumab at a dosage of 2 mg/kg/week	Anthracycline/taxane-based chemotherapy	11.09	9.55
17	Lang et al., 2016 [40]	Taiwan	20 years	3	50	Trastuzumab was used as an adjuvant treatment and reimbursed for only 1 year pre- or post-operation or after standard chemotherapy	Anthracycline/taxane-based chemotherapy	4.98	3.35
18	Hassett et al., 2020 [41]	U.S.A.	5 years	Not reported	55	Trastuzumab with pertuzumab	Trastuzumab emtansine with pertuzumab	4.67	4.62
19	Skedgel et al., 2009 [42]	Canada	25 years	3	49	12-month adjuvant trastuzumab after adjuvant chemotherapy	No trastuzumab (chemotherapy regimen was not explained)	0.59	0.39
20	Macedo et al., 2010 [43]	Portugal	45 years	3	Not reported	Adjuvant trastuzumab for 1 year	No trastuzumab (chemotherapy regimen was not explained)	11.69	9.68
21	Konishi et al., 2022 [44]	Japan	20 years	2	70	Adjuvant trastuzumab for 1 year	No trastuzumab (chemotherapy regimen was not explained)	13.27	12.91
22	Botelho et al., 2022 [45]	Brazil	50 years	5	49	Adjuvant trastuzumab for 1 year	No trastuzumab (chemotherapy regimen was not explained)	7.31	6.48

Abbreviations: A = doxorubicin; C = cyclophosphamide; QALYs = quality-adjusted life years; T = paclitaxel; T-DM1 = trastuzumab emtansine.

3.2. Setting and discounting

Studies were mainly conducted in high-income countries (HICs) [24,25,28,31,32,37–44,46], some upper-middle- income countries (UMICs) [29,35,36,45], and low- or middle-income countries (LMICs) [27,30,33,34]. All of the studies applied Markov modeling, with most of the studies [27–30,46] adopting a long-term time horizon, ranging from 20 years to lifetime. Thirteen studies [25,27,30–33,35,37,39,40,43,46,47] discounted for both costs and QALYs by 3% while the others utilized discount rates by 1,5% [24,38], 2% [44], 3,5% [34], and 5% [29,45].

3.3. Intervention used

Trastuzumab (8 mg/kg i.v.) was administered as an adjuvant treatment to the intervention group for a one-year period in all the included studies. An anthracycline/taxane-based chemotherapy regimen was used for the control group in most of the studies [27,28,30,31,33,34,36,37,39,40,46]. One study [35] compared a trastuzumab-containing regimen with paclitaxel monotherapy administered to an intervention group, and one study [25] compared the use of trastuzumab with neratinib. In addition, trastuzumab administered with pertuzumab was compared with T-DM1 plus pertuzumab in one study [41], and T-DM1 monotherapy was used in another study [24]. Finally, five studies [29,32,38,42–45] did not clearly specify the regimen used for the intervention.

3.4. ICER of the included studies and its association with QALYs

The ICERs reported in the majority of included studies fell within country-specific cost-effectiveness thresholds, indicating that the trastuzumab-containing regimen is a cost-effective strategy (Table 2 and Appendix, Table a C). The range of ICERs reported in studies conducted in HICs [28,31,32,37–41,43,44,46] ranged from US$13,176/QALY to US$254,510/QALY. Conversely, studies conducted in UMICs [29,36,45] suggested that trastuzumab is not cost-effective, with ICERs ranging from US$4,468 to US$191,925, except for one study [35] that concluded the favorable results. In LMICs, all studies concluded that trastuzumab is not a cost-effective treatment [27,30,33,34]. The one-way sensitivity analysis identified the most influential factors affecting ICERs to be the cost of trastuzumab, discount rates for both costs and outcomes, and hazard ratios. Majorities studies reporting higher QALYs tended to demonstrate lower ICERs. This association suggests a favorable relationship between cost-effectiveness and health outcomes, where interventions yielding higher QALYs were associated with lower costs per additional unit of health gain.

Table 2. Summary of ICERs of the included studies.

No	Author, year	Currency and year	Costs, US$ 2023 (trastuzumab)	Costs, US$ 2023 (comparator)	ICER, US$ 2023 (Cost per QALY)	Threshold, US$ 2023	Most influential factors reported in one-way sensitivity analysis
1	Garrison et al., 2007 [46]	US$ (2006)	107,728	42,039	38,629	54,864–82,296	Discount rate, cost of trastuzumab, and probability of metastatic
2	Aboutorabi et al., 2014 [27]	US$ (2010)	78,160	16,991	70,366	Not reported	Cost of trastuzumab, discount rate for outcomes, and HR in trastuzumab group
3	Hall et al., 2011 [28]	£ (2008)	Not reported	Not reported	52,800	61,388	Cost of trastuzumab, hazard ratio, discount rate for benefits, percentage of local recurrence versus distant recurrence, and relative percentage of distant recurrences that are brain metastatic
4	Pichon-Riviere et al., 2015 [29]	US$ (2012)	57,448	16,092	101,925	12,034	Cost of trastuzumab, hazard ratio of trastuzumab and discount rate
5	Gupta et al., 2020 [30]	US$ (2019)	5,729	Not reported	2,257	Not reported	Cost of trastuzumab, DFS utility after 1 year, and transition probability from a disease-free to metastatic in the chemotherapy group
6	Younis et al., 2020 [24]	CAD$ (2018)	257,717	175,899	Dominate	19,783	Not specified (one-way sensitivity analysis was not performed)
7	Hajjar et al., 2019 [31]	US$ (2016)	222,851	135,754	13,203	62,371	Discount rate, probability of relapse, and cost of adjuvant therapy
8	Schwartz et al., 2019 [25]	US$ (2018)	182,675	379,689	497,424	59,771	DFS HR, annual discount rate, and probability of recurrence between 9 and 10 years
9	Purmonen et al., 2011 [32]	Euro (2008)	101,845	88,784	19,840	49,600	Discount rate and trastuzumab treatment cost
10	Gershon et al., 2019 [33]	US$ (2018)	26,999	3,094	24,530	23,908	Discount rate, HR, cost of trastuzumab
11	Genuino et al., 2019 [34]	PHP (2017)	275,694	247,761	28,018	7,414	HR for DFS, duration of efficacy of trastuzumab, discounting rate for outcomes, cost of trastuzumab therapy, and transition probabilities from DFS to metastatic
12	Kongsakon et al., 2019 [35]	US$ (2012)	56,939	36,416	4,468	4,522	Transition probabilities of metastatic to DFS, cost of trastuzumab, cost of computed tomography scan, and quality of life for metastatic stage
13	Buendía et al., 2013 [36]	US$ (2010)	181,547	103,303	98,058	20,574	Discount rate
14	Kurian et al., 2007 [37]	US$ (2005)	286,541	184,547	60,133	60,295	Discount rate, cost of trastuzumab, median survival after breast cancer recurrence, and cost of treating metastatic breast cancer
15	Seferina et al., 2017 [38]	Euro (2012)	391,049	385,327	21,067	128,626	Duration of the treatment effect
16	Liberato et al., 2007 [39]	Euro (2007)	138,017	61,825	32,086	33,829	Time horizon of the analysis
17	Lang et al., 2016 [40]	US$ (2015)	190,348	73,141	71,862	84,496	Transition probabilities of DFS for both the trastuzumab and non-trastuzumab group, health utility for non-trastuzumab group in disease free state, utility for trastuzumab in metastatic state
18	Hassett et al., 2020 [41]	US$ (2020)	236,417	302,026	254,510	Not reported	Recurrence risk
19	Skedgel et al., 2009 [42]	CAD$ (2007)	Not reported	Not reported	84,273	58,287	Discount rate and cost of trastuzumab
20	Macedo et al., 2010 [43]	Euro (2007)	138,018	111,550	13,176	Not reported	Discount rate, QALY values, and costs not associated with trastuzumab
21	Konishi et al., 2022 [44]	¥ (2020)	36,505	29,058	20,398	56,449	Transition probabilities from DFS to metastatic
22	Botelho et al., 2022 [45]	US$ (2018)	42,224	25,522	23,429	10,102	HR, the cost of trastuzumab, and hospital cost in the DFS, local, regional, and contralateral, metastatic and death states

Abbreviations: DFS = Disease-free survival; HR = Hazard ratio; ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life year.

3.5. Health state utility values

The utilities associated with the use of trastuzumab for treating HER2-positive, early-stage breast cancer were difficult to compare because assumptions relating to the values used to estimate the utility, the duration of the study, and the health states utilized differed across the studies (Table 3). Most of the included studies identified the health states utility value as the source of QALYs associated with trastuzumab, namely disease-free survival, local recurrence, distant recurrence, and cardiotoxicity. Trastuzumab-associated utility was mentioned in 6 targeted studies [27,28,31,37,42,46], with values ranging from 0.71 to 0.94. In addition, three articles [27,31,46] discussed utilities that were not associated with the use of trastuzumab. Health state utility values used for quantifying QALYs were also associated with local recurrence [24,27–30,33,36,39,42,43] and ranged between 0.61 and 0.82. One study included local recurrence both during and after the first year [38], five studies [27–29,33,42] considered distant recurrence as a QALY source, and five studies [34,37,41,44,46] considered recurrence. Disease-free survival was presented in ten studies [27,28,30,34,36,38,40,43–45]. Two studies [30,43] considered disease-free survival. Cardiotoxic events related to trastuzumab administration were assumed to occur during treatment and were quantified in some studies [27–29,31,34,36–38,42]. One study [31] provided a detailed explanation as to why cardiotoxicity was considered a health state utility given high and low toxicity levels induced by trastuzumab. Moreover, one study [24] considered invasive disease-free survival and eleven studies [30,32,34–36,38–40,43–45] reported metastatic state utility values, which ranged from 0.484 to 0.762.

Table 3. Methods used and sources of the health state utilities in the included studies.

No	Author, year	Health states	Reference method	Population	Cited sources	Health state utility value
1	Garrison et al., 2007 [46]	Year 1: AC-T including treatment	SG and TTO	General population	Oestreicher et al., 2005 [48]	0.77
		Year 1: AC-TH treatment	SG and TTO	General population	Oestreicher et al., 2005 [48]	0.71
		Year2+ stable disease utility	SG and TTO	General population	Oestreicher et al., 2005 [48]	0.9
		Year2+ recurrence utility	SG and TTO	General population	Hornberger et al., 2002 [49]	0.6
2	Aboutorabi et al., 2014 [27]	Treatment with AC-T	VAS	Patients	Earle et al. 2000 [19]	0.94
		Treatment with AC-TH	VAS	Patients	Earle et al., 2000 [19]	0.94
		Disease free survival	VAS	Patients	Earle et al., 2000 [19]	0.98
		Congestive heart failure	SG and TTO	Patients	Lewis et al., 2001 [50]	0.64
		Local recurrence	TTO, EQ5D	Patients	Lidgren et al., 2007 [51]	0.615
		Distant recurrence	SG	Exploratory interview with expert physicians	Lloyd et al., 2006 [52]	0.615
3	Hall et al., 2011 [28]	Disease free survival	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.77
		Trastuzumab treatment	Assumption	Not reported	Authors’ assumption	0.77
		Asymptomatic cardiac toxicity	Assumption	Not reported	Authors’ assumption	0.77
		Symptomatic cardiac toxicity	EQ-5D	Patients	Calvert et al., 2005 [53]	0.60
		Local recurrence	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.78
		Second primary cancer (first y)	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.70
		Distant recurrence	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.69
		Distant recurrence	EORTC QLQ-C30	Patients	Dinnes et al., 2001 [54]	0.69
4	Pichon-Riviere et al., 2015 [29]	Remission	EQ-5D	General population and patients	Lidgren et al., 2008 [55]	0.97
		Local recurrence	EQ-5D	General population and patients	Lidgren et al., 2008 [55]	0.82
		Distance recurrence	EQ-5D	General population and patients	Lidgren et al., 2008 [55]	0.58
		Cardiac toxicity	EQ-5D	Patients	Calvert et al., 2005 [53]	0.50
5	Gupta et al., 2020 [30]	Disease free survival in first year	Review	Patients	Chen et al., 2009 [56]	0.75
		Disease free survival after first year	Review	Patients	Chen et al., 2009 [56]	0.85
		Local recurrence	Review	Patients	Chen et al., 2009 [56]	0.81
		Metastatic	Review	Patients	Chen et al., 2009 [56]	0.48
6	Younis et al., 2020 [24]	DFS on treatment	KATHERINE EQ-5Db with Canadian tariffs	Patients	Cohort	0.81
		DFS off treatment	KATHERINE EQ-5Db with Canadian tariffs	Patients	Cohort	0.83
		Local recurrence	Assume equal to DFS on treatment	Not reported	Authors’ assumption	0.81
		Remission	Assume equal to DFS on treatment	Not reported	Authors’ assumption	0.83
		First-line (metastatic breast cancer)	SG	General population and patients	Lloyd et al., 2006 [52]	0.76
		Subsequent line (metastatic breast cancer)	SG	General population and patients	Lloyd et al., 2006 [52]	0.51
7	Hajjar et al., 2019 [31]	Therapy (no adjuvant trastuzumab)	SG and TTO	General population	Kurian et al., 2007 [37]; Garrison et al., 2007 [46]	0.81
		Utility due to high toxicity (all trastuzumab arms)	VAS, CEA registry, EQ-5D	General population	Kurian et al., 2007 [37]; Skedgel et al., 2013 [47]	0.56
		Utility due to low toxicity (all trastuzumab arms)	VAS, CEA registry, EQ-5D	General population	Kurian et al., 2007 [37]; Skedgel et al., 2013 [47]	0.85
		Adjuvant therapy (TH)	VAS	Patients	Slamon, 2011 [57]; Kurian et al., 2007 [37]	0.79
		Adjuvant therapy (ACTH and TCH)	Not reported	General population	Kurian et al., 2007 [37]; Garrison et al., 2007 [46]	0.78
		40–49	EQ-5D	General population	Skedgel et al., 2013 [47]	0.84
		50–59	EQ-5D	General population	Skedgel et al., 2013 [47]	0.83
		60–69	EQ-5D	General population	Skedgel et al., 2013 [47]	0.83
		70–79	EQ-5D	General population	Skedgel et al., 2013 [47]	0.81
		>80	EQ-5D	General population	Skedgel et al., 2013 [47]	0.78
		Utility of relapse	SG and VAS	General population	Kurian et al., 2007 [37]	0.55
		Utility after therapy (relapse-free)	SG	General population	Garrison et al., 2007 [46]	0.9
8	Schwartz et al., 2019 [25]	Baseline	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.86
		Month 1–2	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 3–5	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 6–8	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 9–11	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 12	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.85
		Remission, placebo	EQ-5D	Patients	Delaloge et al., 2019 [58]
		Baseline	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.86
		Month 1–2	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.86
		Month 3–5	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 6–8	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.85
		Month 9–11	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.84
		Month 12	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.85
		Remission, no drug (after 12 months)	EQ-5D	Patients	Delaloge et al., 2019 [58]	0.85
		Disease progression	Not reported	Not reported	Hornberger et al., 2002 [49]	0.60
9	Purmonen et al., 2011 [32]	Primary breast cancer, 1st year	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.69
		Primary breast cancer, subsequent years	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.78
		Metastatic breast cancer	EQ-5D	Patients	Lidgren et al., 2007 [51]	0,68
		Dead	EQ-5D	Patients	Lidgren et al., 2007 [51]	0.00
10	Gershon et al., 2019 [33]	Remission	Not reported	General population and patients	Norum et al., 2007 [59]; Garrison et al., 2007 [46]	0.94
		Local recurrence	Not reported	Patients	Piccart-Gebhart et al., 2005 [11]	0.82
		Distant recurrence	Not reported	Patients	Piccart-Gebhart et al., 2005 [11]	0.58
11	Genuino et al., 2019 [34]	Disease free survival	EQ-5D-3L	Patients	Not reported	0.83
		CHF	EQ-5D-3L	Patients	Not reported	0.67
		Recurrence	EQ-5D-3L	Patients	Not reported	0.83
		Metastatic	EQ-5D-3L	Patients	Not reported	0.76
12	Kongsakon et al., 2019 [35]	Non metastatic	Review	General population	Oestreicher et al., 2005 [48]; Hornberger et al., 2002 [49]	0.90
		Metastatic	Review	General population	Oestreicher et al., 2005 [48]; Hornberger et al., 2002 [49]	0.60
13	Buendía et al., 2013 [36]	Disease free survival	Not reported	Not reported	Published literature unknown	0.85
		Local recurrence	Not reported	Not reported	Not reported	0.81
		Metastatic	Not reported	Not reported	Not reported	0.48
		Cardiac events	Not reported	Not reported	Not reported	0.70
14	Kurian et al., 2007 [41]	Symptomatic heart failure	VAS	Patients	Lewis EF et al., 2001 [50]	0.64
		Cardiac toxicity with symptoms	Assumption	Not reported	Assumption	1.00
		Breast cancer recurrence	SG and VAS	Nurses who were acting as proxy patients	Lewis EF et al., 2001 [50]	0.55
		Weighted average, cardiac toxicity utility	VAS	Patients	Lewis EF et al., 2001 [50]	0.91
		Well	TTO	General population	Fryback DG et al., 1993 [60]	1.00
		Well, age 45–54 years	TTO	General population	Fryback DG et al., 1993 [60]	0.90
		Well, age 55–64 years	TTO	General population	Fryback DG et al., 1993 [60]	0.87
		Well, age 65–74 years	TTO	General population	Fryback DG et al., 1993 [60]	0.83
		Well, age >75 years	TTO	General population	Fryback DG et al., 1993 [60]	0.79
15	Seferina et al., 2017 [38]	Disease free survival	EQ-5D with UK tariffs	Patients	Cohort	0.73
		Disease free survival after first year	EQ-5D with UK tariffs	Patients	Cohort	0.81
		Local recurrence first year	EQ-5D with UK tariffs	Patients	Cohort	0.72
		Local recurrence after first year	EQ-5D with UK tariffs	Patients	Cohort	0.71
		Distant metastases in first year	EQ-5D with UK tariffs	Patients	Cohort	0.58
		Distant metastases after first year	EQ-5D with UK tariffs	Patients	Cohort	0.60
		Cardiac toxicity	EQ-5D	Patients	Calvert MJ et al., 2005 [53]	0.6
		Disutility	EQ-5D with UK tariffs	Patients	Cohort	0.13
16	Liberato et al., 2007 [39]	Disease free (in patients without symptomatic cardiac dysfunction)	TTO	Patients	Kirsch J, 2000 [61]	0.97
		Disease free (in patients with symptomatic cardiac dysfunction)	EQ-5D	Patients	Kirsch J, 2000 [61]	0.51
		Disease free after local relapse	Not reported	Patients	Kirsch J, 2000 [61]	0.92
		Local recurrence	Observational study	Patients	Not reported	0.82
		Metastatic disease	Observational study	Patients	Not reported	0.58
17	Lang et al., 2016 [40]	Disease free survival	Not reported	Patients	Not reported	0.83
		Local recurrence to metastatic	Not reported	Patients	Not reported	0.81
		Metastatic to death	Not reported	Patients	Murray CJ et al., 2000 [62]; Hedden L et al., 2012 [63]; Lidgren M et al., 2007 [51]	0.65
18	Hasset et al., 2020 [41]	Recurrence	SG	General population and patients	Lloyd et al., 2006 [52]	0.60
		Post-surgery (off treatment)	VAS	Patients	de Koning et al., 1991 [64]	0.87
19	Skedgel et al., 2009 [42]	Well, off treatment	CEA registry	Not reported	CEA registry	0.90
		Well, on adjuvant trastuzumab	VAS	Patients	Kurian et al., 2007 [37]	0.92
		Cardiotoxicity	CEA registry	Not reported	Gold, 1996 [65]	0.83
		1st local cancer recurrence	CEA registry	Not reported	CEA registry	0.70
		2nd local cancer recurrence	CEA registry	Not reported	CEA registry	0.50
		Well after LCR	CEA registry	Not reported	CEA registry	0.90
		Distant recurrence	CEA registry	Not reported	CEA registry	0.60
20	Macedo et al., 2010 [43]	Disease free survival in first year	TTO	Patients	Tengs and Wallace, 2000 [66]	0.75
		Disease free survival after first year	TTO	Patients	Tengs and Wallace, 2000 [66]	0.84
		Metastatic disease in first year	TTO	Patients	Hayman, 1998 [67]	0.48
		Metastatic disease after first year	Basic reference gamble	Patients	Carter, 1998 [68]	0.48
		Local recurrence	SG	Patients	Tengs and Wallace, 2000 [66]	0.81
21	Konishi et al., 2022 [44]	Disease free survival in first year	FACT-G	Patients	Taira et al., 2021 [69]	0.81
		Disease free survival after first year	FACT-G	Patients	Taira et al., 2021 [69]	0.83
		Recurrence	Observational study	Patients	Liberato et al., 2007 [39]	0.82
		Metastatic	Observational study	Patients	Liberato et al., 2007 [39]	0.58
		Death	Observational study	Patients	Liberato et al., 2007 [39]	0
22	Botelho et al., 2022 [45]	Disease free survival in the first year	TTO	Patients	Macedo et al., 2010 [43]	0.749
		Local, regional, and contralateral	SG	Patients	Macedo et al., 2010 [43]	0.81
		Metastatic	TTO	Patients	Macedo et al., 2010 [43]	0.484

Abbreviations: A = Doxorubicin; C = Cyclophosphamide; CEA = Cost-effectiveness analysis; EORTC QLQ-C30 = European Organization for Research and Treatment of Cancer Quality-of-life Questionnaire Core 30; EQ-5D = EuroQoL-5 Dimensions; EQ-5D-3L = EuroQoL-5 Dimensions-3Levels; T = Paclitaxel; H = Trastuzumab; TTO = Time trade-off; SG = Standard gamble; VAS = Visual Analog Scale.

3.6. Methods used to quantify health state utilities

The methods used to quantify health state utilities in the selected studies entailed direct and indirect evaluation or expert opinion [70]. We found multiple reference method being used to measure utilities, with different measurements reported in the cited studies. In 7 studies [25,28,29,31,32,38,39], QALYs were based on utility estimates obtained indirectly through EuroQoL-5 Dimensions (EQ-5D) scores reported in the studies [47,51,53,55,58,61], which were obtained from participants’ self-reporting of their conditions or hypothetical health states using a generic scale, and the application of country-specific tariffs for the general population for measuring the utility values. The Visual Analogue Scale (VAS) was also applied in 5 studies [27,31,37,41,47]. Other methods used to measure utility included standard gamble (SG) scores in six studies [24,31,41,43,45], in which health states were compared against a gamble of death and perfect health. The time trade-off (TTO) method was applied in four studies [27,39,43,45] by referring to some previous studies [60,66,67]. One study derived utilities from EuroQoL-5 Dimensions-3Levels (EQ-5D-3 L) [34], FACT-G [44], and another study from European Organization for Research and Treatment of Cancer Quality-of-life Questionnaire Core 30 (EORTC-QLQ-C30) [28]. A combination of techniques was applied in some studies [27,31,37,46]. In addition, one study [24] combined direct and indirect methods, applying the utility derived from EQ-5D-3 L measurements with Canadian tariffs relating to breast cancer in women at screening, during treatment, and at 6-month intervals for 1 year after the study’s completion. Furthermore, in one study [38], which estimated the utilities from real-world data, a cross-sectional survey was performed using EQ-5D instruments with UK tariffs. Finally, some studies [24,28,37] included assumptions made by the authors.

In this review, health state utility values associated with early-stage breast cancer patients were obtained from measurements applied across different population targets, namely communities, patients, and experts and physicians. Most of the primary articles [25,27,29–34,37–41,43–45] obtained the utility values based on breast cancer patients’ preferences. Four studies [31,35,37,46] targeted general populations, specifically valuing the utilities used according to age group. Four studies [24,29,33,41] derived the utility values from the general population as well as patients and explained differences in utility values measured in disease-free patients and those with breast cancer which the utility values for HER2-positive breast cancer patients were generally found to be lower than in general population. Only one study [27] derived the weights of QALYs from clinical experts during an exploratory interview. Table 2 provides summary details of the health state utilities for each included study.

3.7. Sources of the health state utilities in the included studies

Of the literature cited in the studies as references for quantifying the utilities, one specific study [51] was the most frequently cited source for measuring utilities to estimate QALYs in the included studies [27,28,32]. Another paper [52] cited by three studies [27,28,32] explained metastatic recurrence using SG scores to measure the utilities. A third study [60], cited by Kurian et al. [37] estimated the utility associated with a disease-free state according to age groups (45–54 years, 55–64 years, 65–74 years, and >75 years), measured using the TTO method. Two studies [37,46] were cited in a study [31] included in this review, which estimated the utilities corresponding to regimens with and without trastuzumab. In addition, utility relating to cardiac toxicity was calculated in the above study using EQ-5D with reference to two studies [37,47]. Figure 2 visualizes the connections between the studies included in the review and those cited as the sources of utility.

Figure 2. Bibliometric network of the connections between the selected studies included in this review and those cited as the sources of utilities.

3.8. Overall survival values in the included studies

Only a few of selected studies [24,25,27–30,32,35,37–39,42,44–46] explicitly provided survival data (Table 4). Most of those studies [28–30,36,38,42,45] obtained OS data from the HERA trial [72,80], whereas one study [30] used a combination of the HERA trial [72] and the CONCORD study [73] and calibrated the model for the scenario to enable the overall survival value to be predicted. Three studies reported OS data derived from a combined analysis of data from the NCCTG N9831 [12] and NSABP B-31 [71] trials, which were projected and extrapolated to a 20-year horizon and a lifetime horizon in the model. The survival data from the BCIRG 006 trial [57] was utilized in one study [27], and a trial [77] in which a monoclonal antibody plus chemotherapy regimen for metastatic disease was evaluated was used to define the survival value in two studies [37,39]. Survival outcomes were derived from the ExteNET [74] trial in one study [25], whereas the FinHer trial [75] was comprised in another study [32]. Finally, a study [38] used the survival value reported in a real-world cohort study [78].

Table 4. Overall survival values in the included studies.

No	Author, year	Overall survival value (%)		Sources and methods	Reference(s)
		Trastuzumab with chemotherapy	Chemotherapy without trastuzumab
1	Garrison et al., 2007 [46]	91	87	The data were projected from the combined analysis of data from the 4-year NSABP B-31 and NCCTG N9831 trials and extrapolated to 20-year and lifetime horizons using a Markov model	[12,71]
2	Aboutorabi et al., 2014 [27]	85	93	Based on data from the BCIRG 006 trial	[57]
3	Hall et al., 2011 [28]	92.40	89.70	Data from the HERA trial; based on the assumption of DFS that predicts the OS	[72]
4	Pichon-Riviere et al., 2015 [29]	97.40	94.20	Data from the HERA trial	[72]
5	Gupta et al., 2020 [30]	Not reported	66.10	Data from the CONCORD study, HERA trial, and calibration of the model for the counterfactual scenario to predict survival based on breast cancer survival from 2 Indian cancer registries	[73]
6	Younis et al., 2020 [24]	Not reported	Not reported	Not reported
7	Hajjar et al., 2019 [31]	Not reported	Not reported	Not reported
8	Schwartz et al., 2019 [25]	Not reported	Not reported	Based on data obtained from ExteNET	[74]
9	Purmonen et al., 2011 [32]	Not reported	Not reported	Data from the FinHer trial	[75]
10	Gershon et al., 2019 [33]	Not reported	Not reported	Not reported
11	Genuino et al., 2019 [34]	Not reported	Not reported	Not reported
12	Kongsakon et al., 2019 [35]	85.70	73.50	Data from a combined analysis extrapolated to a lifetime horizon using a cohort simulation model	[76]
13	Buendía et al., 2013 [36]	Not reported	Not reported	Data from the NCCTG N9831 and NSABP B-31 and HERA trials	[12,72]
14	Kurian et al., 2007 [37]	Not reported	Not reported	Data from a pivotal randomized trial of trastuzumab for metastatic disease	[77]
15	Seferina et al., 2017 [38]	Not reported	Not reported	Data from all patients in cohort studies (in real-world settings) and from the HERA trial	[72,78]
16	Liberato et al., 2007 [39]	Not reported	Not reported	Data from a trial	[77]
17	Lang et al., 2016 [40]	Not reported	Not reported	Not reported
18	Hassett et al., 2020 [41]	Not reported	Not reported	Not reported
19	Skedgel et al., 2009 [42]	Not reported	Not reported	Data from the HERA trial of adjuvant trastuzumab	[72]
20	Macedo et al., 2010 [43]	Not reported	Not reported	Not reported
21	Konishi et al., 2022 [44]	95.3	92.4	Data from a trial	[79]
22	Botelho et al., 2022 [45]	79	73	Data from the HERA trial	[80]

Abbreviations: BCIRG 006 = Breast Cancer International Research Group; DFS = Disease Free Survival; FinHer = The Finland Herceptin; HERA = HERceptin Adjuvant; NSABP B-31 = National Surgical Adjuvant Breast and Bowel Project B-31; NCCTG N9831 = North Central Cancer Treatment Group N9831; OS = Overall survival.

3.9. Quality assessment

In terms of methodological quality, most studies comprehensively defined the research questions, described alternatives, assessed intervention effectiveness, adjusted for present value, conducted incremental analysis, performed sensitivity analysis, and interpreted results. However, the valuation of costs and outcomes was not adequately justified in half of the included studies [31,33,35–37,40–43,45,46] (Appendix, Table D).

With reference to the items in the CHEERS checklist, all the studies described the outcomes used to measure benefits and harm according to the inclusion criteria of the studies. However, four studies [31,33,36,40] did not sufficiently describe the methods used to assess the utilities or explain how the outcomes were quantified. Valuation of outcomes was adequately described in half of the targeted studies. In summary, the targeted studies demonstrated strengths in defining the utilities they used. They generally employed appropriate methods for assessing health state utility values, with those adopting a direct approach utilizing validated methods recommended in the literature [81]. Studies applying indirect methods utilized validated questionnaires and explained the tariffs applied. Most primary studies that elicited utilities provided detailed information about the methods used, population targets, instruments used for data collection, and the duration of the study. Detailed results of the quality assessment can be found in the Appendix, Tables E and F.

4. Discussion

To our knowledge, this review represents the first comprehensive examination of economic evaluations pertaining to early-stage breast cancer patients undergoing treatment with trastuzumab-containing regimens. In addition to extracting ICERs, our analysis encompasses a thorough assessment of HRQoL measures and the identification of sources contributing to utilities used in estimating QALYs. Our findings reveal variations in trastuzumab-containing regimens, intervention groups, methods of measuring utilities, and the target populations for defining health state utilities. Most studies included in this review were conducted in HICs and suggest trastuzumab as a cost-effective treatment option. Most studies reported higher QALY values in the trastuzumab arms compared to control groups, with EQ-5D emerging as the most frequently utilized instrument for assessing health state utility values in obtaining QALYs, while the populations targeted to obtain utilities exhibited diversity. Notably, studies reporting lower ICERs tended to demonstrate higher QALYs, indicating a positive correlation between cost-effectiveness and health outcomes. This highlights the importance of considering both factors simultaneously in evaluating the overall value of healthcare interventions.

Cost-effectiveness analysis, as the gold standard of decision-making process in health-related policy making, employs QALY measurement, providing a standardized and widely applicable method for evaluating the comparative value of various interventions. Each cost-effectiveness analysis included in this study has meticulously assessed the cost and effectiveness of competing interventions, offering decision-makers a precise quantitative understanding of their effectiveness. According to our review, trastuzumab-containing regimens were deemed cost-effective in the majority of HICs [28,31,32,37–41,43,44,46]. Furthermore, trastuzumab was found to be cost-effective in Thailand [35], contrary to the findings from other studies conducted in UMICs [29,36,45], which concluded that it was not cost-effective. However, our study revealed that trastuzumab was not a cost-effective strategy in all LMICs included in our analysis. These findings align with a previously published review [82], suggesting that a significant portion of the studies exhibited favorable cost-effectiveness results primarily due to higher willingness-to-pay thresholds in HICs. The trastuzumab-containing regimen in this study demonstrated a cost-effective strategy, yielding more favorable QALY gains despite its higher cost compared to the control group. It is noteworthy that studies reporting higher QALY estimates tended to exhibit lower ICERs, indicating a more efficient allocation of resources. Consequently, the trastuzumab-containing regimen emerged as a more favorable option in healthcare decision-making, offering both economic efficiency and enhanced patient outcomes.

Several studies [31,33,35–37,40,42,43] included in our review did not provide detailed explanations of the methods used to value utilities relating to health states associated with the therapy. Because of the limited number of available primary studies that evaluated the utilities, it was difficult to obtain information on the source and quality of published utilities required for designing economic measures. However, such constraints within cost-effectiveness studies [70,83] must be presented explicitly for decision makers to consider.

The variations in estimates of the utility depend on various factors, notably the measurement method used, the study population, health states, and the study location [84,85]. Regarding the approaches employed for evaluating utilities in breast cancer patients, our findings revealed that when compared to values derived from EQ-5D, utilities assessed via the VAS were lower, while those obtained through the SG and TTO methods were higher, in parallel with another finding [86]. As choice-based instruments, SG and TTO consider the trade-offs made when eliciting preferences for health states with only SG incorporates preferences related to risk [86]. Conversely, the VAS method, while comparatively simpler, carries susceptibility to response spreading bias and end-state aversion bias [87]. Some health technology assessment (HTA) agencies suggest utilizing indirect methods due to their low cost, ease of implementation, and reduced cognitive load in contrast to directly applying elicitation tasks [86]. We additionally discovered that patients expressed greater utility values in contrast to individuals from the general population, supporting findings consistent with existing literature [86]. The difference in utility values among populations might stem from the hypothetical situation proposed to the general population, the adjustment to the health condition by patients, or the likelihood of a response shift [86].

This systematic review demonstrates the broad selection of tools used to measure utility in health economics, reflecting the field’s growing interest. Another review also confirms the diverse range of instruments employed for utility measurement [88]. This offers researchers and policymakers diverse tools for program evaluation and resource allocation. However, comparing programs using different tools might be challenging as they can produce varying scores due to differences in their descriptive systems or valuation methods within the same field. The most widely used approach in this study was EQ-5D, which is able to detect improvements as well as deteriorations in health conditions and to distinguish slight changes in utility [89]. According to one published review [85], the most frequently used method for estimating utility in the early stage of breast cancer is the VAS, followed by the SG, EQ-5D, and lastly the TTO.

The findings of this review notably highlighted the large number of health economic studies that have focused on a trastuzumab-containing regimen used for managing HER2-positive, early-stage breast cancer. The paucity of reviews focusing on a specific drug uncovered by our review strengthens its value when used in the development of qualitative descriptions of the QALYs of early-stage breast cancer patients and the health state utilities used in the included studies. However, only a few of the papers that we reviewed elicited the sources of the weights of QALYs from the population of women with early-stage breast cancer being treated with trastuzumab. We found that the health utility values for patients using trastuzumab were higher compared with those of patients not treated with trastuzumab. This finding is supported by some evidence, which suggests that adding trastuzumab to a chemotherapy regimen for treating early-stage breast cancer leads to a meaningful improvement in HRQoL [90,91].

This study revealed differences in the values of the QALYs derived from the various types of utility approaches used in the reviewed studies. The diversity in utility values acquired using different methods may produce varying results in terms of health states related to particular conditions [92–94]. In addition, when the utilities used in economic evaluations are sourced from the literature, some concerns relating to the patient population, the sources of utilities, and the elicitation procedure should be considered. Furthermore, a systematic review of the literature aimed at identifying utilities constitutes good practice in health economic studies [95]. Our review covers a range of studies within the literature considered as the sources of utilities used to calculate QALYs relating to the use of trastuzumab for managing HER2-positive, early-stage breast cancer.

To date, no firm guidelines have been established on the most appropriate methods for quantifying the utilities of women with breast cancer, including economic measures and the targeted population for evaluating the clinical outcomes of breast cancer treatments. Clarifying the utilities associated with breast cancer treatment is essential in clinical settings, but it also yields important insights for developing treatment guidelines. Establishing clear guidelines for quantifying utilities in women with breast cancer involves a collaborative effort among researchers, clinicians, and policymakers. This process should involve rigorous analysis and validation of different utility measurement methods, considering the diverse aspects of patient experiences and preferences. Our findings offer valuable information on health state utilities that apply to the use of trastuzumab in the management of HER2-positive, early-stage breast cancer as trastuzumab is entering new markets of low- and middle-income countries where previously this treatment might have been unaffordable. The relatively high price of the drug leads to some difficulties in its implementation across health systems and evaluating the cost-effectiveness of trastuzumab in accordance with country contexts is essential for health financing decisions by policymakers [96].

The sources of QALYs used in most of the papers included in this review were generated from the cited papers [11,19,37,47,48,50–54,56–59,97]. However, only two articles [24,38] estimated utilities using cohort studies. The reviewed studies noted the substantial heterogeneity in the valuing of the sources of QALYs. Examples of this heterogeneity included the year of costing, the study perspective, utility measurement, time scale, and discounting, which made it difficult to compare and synthesize the studies. Within economic evaluations, the duration of utilities is a key factor to be considered when estimating QALYs; most of the selected studies considered the utilities to be associated with temporary health states and therefore used an appropriate chaining adaptation.

Adding trastuzumab to chemotherapy within a regimen for treating HER2-positive breast cancer has been proved to lead to a better OS profile [76,98]. Because QALYs estimates are based on improvements in survival, they contribute to the calculation of the weights of the QALYs. In this review, most of the studies exhibited the higher of the QALYs for the regimen combining trastuzumab with chemotherapy, with higher OS values found in the trastuzumab arms. Only a few of the targeted studies reported OS values because of challenges associated with the providing of the evidence. Consequently, available methods that entail assumptions regarding the duration of the treatment effects during the trials are required to calculate the OS value.

This study exhibits several strengths Firstly, it provides a comprehensive examination of economic evaluations concerning early-stage breast cancer patients treated with trastuzumab-containing regimens. Beyond extracting ICERs, it conducts a thorough assessment of HRQoL measures and identifies sources contributing to utilities used in estimating QALYs, offering a comprehensive understanding of the economic and health outcomes associated with trastuzumab treatment. Secondly, its inclusion of studies from diverse income settings enhances the generalizability of findings, providing valuable insights into the cost-effectiveness and health outcomes of trastuzumab across different healthcare contexts. Thirdly, efforts were made to trace utility values back to their derivation methods, population targets, and QALY weighting, facilitating critical insights for oncologists in treatment decision-making. This compilation of QALYs associated with trastuzumab enables holistic evaluations of patient well-being, aiding in the assessment of treatment benefits, potential side effects, and associated costs. Moreover, it facilitates dynamic decision-making by integrating new evidence for improved patient-centered care [99,100]. It is important to note that while QALYs offer a quantitative measure to evaluate treatments, they are not the sole determinant in clinical decision-making. Oncologists consider various factors, including individual patient characteristics, preferences, tumor characteristics, potential side effects, and the overall health of the patient, in addition to QALY assessments, to tailor personalized treatment recommendations.

However, our study has some drawbacks. The health state utilities associated with trastuzumab treatment could not be estimated because of the heterogeneity of the methods used to adjust the utility values. In addition, when evaluating health utility values between patients using trastuzumab and those not treated with trastuzumab, it’s essential to consider of various biases that could influence the observed differences. Patients receiving trastuzumab might differ systematically from those who do not receive this treatment. Factors like disease severity, comorbidities, or patient demographics could influence the decision to administer trastuzumab and could impact health utility values. Moreover, differences in the duration of follow-up between the groups or the timing of health utility assessments could influence results. Short-term assessments might not capture long-term effects, potentially biasing the comparison.

5. Conclusions

In summary, this review provides a comprehensive examination of economic evaluations regarding early-stage breast cancer patients treated with trastuzumab-containing regimens. It highlights variations in study methodologies, utility measurement methods, and health outcomes across different settings. While trastuzumab was generally deemed cost-effective in HICs, its cost-effectiveness in LMICs remains uncertain. Studies reporting higher QALYs tend to have lower ICERs, indicating a positive relationship between cost-effectiveness and health outcomes. The positive correlation between lower ICERs and higher QALYs emphasizes the importance of considering both cost-effectiveness and health outcomes simultaneously in evaluating healthcare interventions. However, challenges such as methodological heterogeneity and transparency in utility valuation persist, underscoring the need for standardized guidelines and collaborative efforts among stakeholders. Despite these limitations, this review offers valuable insights into the economic evaluation of trastuzumab-containing regimens and informs decision-making in breast cancer treatment, particularly in resource-constrained settings.

Article highlights

• Trastuzumab-containing regimens are generally deemed cost-effective in High-Income Countries (HICs), with studies reporting favorable quality-adjusted life year (QALY) gains and lower Incremental Cost-Effectiveness Ratios (ICERs) compared to control groups.

• Variations in utility values depend on measurement methods, study populations, and health states, posing challenges in comparing outcomes and assessing the true value of healthcare interventions.

Declaration of interest

The author has 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.

Reviewer disclosures

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

Abbreviations

BCIRG	=	Breast Cancer International Research Group
CHEERS 2022	=	Consolidated Health Economic Evaluation Reporting Standards
EQ-5D	=	EuroQoL-5 Dimensions
EQ-5D-3 L	=	EuroQoL-5 Dimensions-3Levels
EORTC-QLQ-C30	=	European Organization for Research and Treatment of Cancer Quality-of-life Questionnaire Core 30
FinHer	=	The Finland Herceptin
GDP	=	Gross Domestic Product
HER2	=	Human epidermal growth factor receptor 2
HERA	=	HERceptin Adjuvant
HICs	=	High-income countries
HRQoL	=	Health-related quality of life
HTA	=	Health technology assessment
ICER	=	Incremental cost-effectiveness ratio
LMICs	=	Low- or middle- income countries
NCCTG N9831	=	North Central Cancer Treatment Group N9831
NSABP B-31	=	National Surgical Adjuvant Breast and Bowel Project B-31
OS	=	Overall survival
PPP	=	Purchasing power parity
PRISMA	=	Preferred Reporting Items for Systematic Reviews and Meta-Analyses
QALY	=	Quality-adjusted life years
SG	=	Standard gamble
T-DM1	=	Trastuzumab emtansine
TTO	=	Time trade-off
UMICs	=	Upper- middle- income countries
VAS	=	Visual Analogue Scale

Authors contributions

SM Khoirunnisa designed the study protocol, screened the included studies, analyzed the data, and drafted the manuscript. FDA Suryanegara reviewed the selected studies, analyzed the data, and drafted the manuscript.

D Setiawan and MJ Postma reviewed the data, revised the draft of the manuscript and supervised the project.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14737167.2024.2352006

Additional information

Funding

This study was funded through the Indonesia Endowment Funds for Education (LPDP) under [Grant no. 0005479/PHA/D/BUDI-2019] and by the University of Groningen/University Medical Center Groningen.