The growing interest in genomic medicine interventions reflects an increasing need to incorporate economic considerations into real-world implementation strategies and evidence-based policy decision (1). In this evolving landscape, next-generation sequencing (NGS) technologies have garnered substantial attention (2). NGS encompasses various techniques, such as genome sequencing (GS), which decodes the organism’s complete genomic sequence, and exome sequencing (ES), which specifically targets the protein-coding regions (3). These advanced techniques are not only challenging traditional genetic testing paradigms (4) but also raise concerns related to the cost-effectiveness (2). GS and ES have shown significant potential in improving the early diagnosis of rare and inherited diseases, particularly in cases with atypical or non-specific symptoms (5–7). They are also accelerating the discovery of novel genes implicated in various genetic disorders (8, 9). As for oncology, GS and ES facilitate the detection of somatic variants that inform the development of personalized therapies (10, 11). However, despite these advancements, their implementation poses challenges from both clinical and public health perspectives, including complexities in data interpretation, limited workforce capacity, risks of overdiagnosis, and broader ethical and societal concerns (1).

Given this context, the demand for robust economic evidence to support the integration of GS and ES into clinical practice has grown substantially (12–17), but their complex and far-reaching implications drive an ongoing debate on which outcome metrics best capture their true economic value (18–21). Already in 2018, a broad systematic review of economic evaluations (EEs) underscored the limited availability of high-quality evidence in clinical settings and noted that diagnostic yield (DY) - while useful - remains a narrow and insufficient outcome measure to be used for these technologies (12). Moreover, Alam et al. critically examined EEs of GS and ES in pediatric populations, highlighting the need to enhance methodological rigor (13). Recently, Ferket et al. proposed conceptual frameworks to address longstanding methodological challenges in genomic EEs, advocating for the inclusion and modeling of broader health outcomes such as quality-adjusted life years (QALYs) (22). Despite existing limitations in this area (18–21), the incorporation of measures like survival gains and quality of life improvements is still widely recognized as essential for informing resource allocation decisions and ensuring comparability with other health interventions (23). To date, no comprehensive synthesis exists that investigates how health outcomes are defined, measured, and incorporated into full EEs of these technologies. Therefore, the primary aim of this review was to examine the methodological approaches used to quantify and integrate health outcomes in EEs of GS and ES. A secondary objective was to report descriptively findings on cost-effectiveness, however the review was not designed to determine whether GS/ES are cost-effective but rather to assess how health outcomes are measured and incorporated within current evaluation frameworks.

This review was conducted in accordance with the Cochrane Handbook for Systematic Reviews (24), the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement (25) and the Center for Reviews and Dissemination guidance on undertaking systematic reviews of economic evaluations (26). The review protocol was registered in PROSPERO under the identifier CRD42023430992. Since primary data collection was not part of this study, ethical review board approval and informed consent were not necessary.

The database search identified 45,007 articles (Figure 1). After removing duplicates, 24,079 titles and abstracts were screened, yielding 378 full-text articles. Ultimately, 12 studies (31–42) met the inclusion criteria. Included studies were published between 2019 and 2025, four were from the United States of America (38–41), four from the Netherlands (32–35), two from Australia (36, 37), one from the United Kingdom (31) and one from Israel (42). All were CUA, mainly using health economic models, except for two cohort studies (37, 41) reporting real-world data. Four studies (32–35) focused on sequencing cancer treatment decisions and one to guide treatment choices for patients with major depressive disorder (MDD) (31); the remaining seven studies (36–42) examined genetic diagnosis of rare diseases in pediatric populations or during pregnancy.

This systematic review investigated traditional full EEs of GS- and ES-based interventions that measured and integrated health outcomes, identifying CUAs conducted across five countries in the recent literature. Although cost-effectiveness results were reported, they were not the primary focus of the analysis; instead, we concentrated on how existing studies measure and incorporate health outcomes and where key methodological advances are still needed. Our findings reflect a substantial degree of methodological heterogeneity across the included studies: differences in time horizons, discount rates, analytical perspectives, and utility sources were frequent. Such inconsistencies limit the interpretability and generalizability of cost-effectiveness outcomes and illustrate the absence of standardized methodological expectations for economic evaluations of GS and ES.

We provided an overview of EE methods in two main clinical applications of GS and ES: treatment guidance and the diagnosis of genetic disorders.

In the first study area, research has mainly examined GS in oncology—especially advanced lung cancer (32–34), largely driven by a Dutch initiative (48)—with more recent applications in prostate cancer (35). These analyses generally showed strong methodological quality, especially in scenario definition and sensitivity analysis, which are essential for addressing uncertainty in early technology assessment (49). Across studies, sequencing costs, targeted therapy prices, and the probability of detecting actionable variants emerged as the main drivers of cost-effectiveness (32–35). The only study on psychiatric disorder (31) yielded similar results, suggesting that ES-guided treatment could become more cost-effective with lower treatment costs and improved diagnostic accuracy, while also highlighting pharmacogenomics as a promising area for future investigation. However, most treatment-guidance studies relied on simulated cohorts; only one used real-world data (34), limiting generalizability (50). In addition, GS costs were sometimes taken from commercial providers (32, 33), potentially over- or underestimating true costs. It should also be noted that GS/ES-guided treatment for cancer patients appeared potentially cost-effective when assessed against the Dutch willingness-to-pay threshold of €80,000 per QALY; however, this finding has limited generalizability, as differences in cost structures, treatment pathways, and WTP thresholds across healthcare systems may hinder its direct transferability to other settings., Additionally, due to data gaps, studies relied on assumptions about DY, uptake, treatment effectiveness, and patient response. For instance, Simons et al. estimated treatment effectiveness for rare variants due to the lack of randomized controlled trial data (33), a common challenge in precision medicine (51). This limitation could be addressed by emerging study designs, such as umbrella and basket trials, which enable a more efficient evaluation of targeted therapies in genetically defined subpopulations (52). Finally, uptake rates were often unreported, and some authors (34, 35) stressed the need to account for diagnostic waiting times, as these can influence treatment initiation and survival and should be incorporated to improve the realism of future evaluations (53). GS and ES are particularly valuable for the timely diagnosis of rare diseases, and our review suggests they may be cost-effective in this context; however evidence remains preliminary and context-dependent (36–42). Diagnostic strategies varied widely (e.g., rapid vs. standard sequencing, proband-only vs. trio sequencing), as did populations, clinical settings, and delivery models. Cost inclusion also differed markedly. While some studies included long-term care (38, 39), disability management (42), and reproductive services (36, 37), most lacked detailed downstream cost data and often excluded indirect costs, such as informal caregiving, productivity losses, and special education needs. Given that rare diseases frequently impose lifelong disability and family burden (54), failing to incorporate societal and informal costs may lead to systematic underestimation of true economic impact. Moreover, due to data scarcity, several studies relied on proxy conditions to estimate long-term outcomes and costs (38, 39). Methodological inconsistencies were common across diagnostic EEs, including differences in perspective, discount rates, time horizons, cohort size, and reporting transparency (37, 41, 42). To improve the comparability and reliability of future evaluations, the adoption of standardized, transparent reporting frameworks has been strongly recommended (55).

In estimating health benefits, most studies relied on common utility sources (43, 44) and validated multi-attribute utility instruments, including EQ-5D, HUI, and SF-6D (45–47). These instruments are also the most frequently recommended by HTA guidelines internationally (56, 57). In oncology and psychiatry, health states and disease progression were relatively well defined. In contrast, modeling outcomes for rare diseases proved more complex due to highly heterogeneous conditions and limited data on natural history. Here, health states often reflected levels of childhood disability and relied on caregiver-reported utilities, which are appropriate when patients cannot self-report their HRQOL (57). However, traditional utility measures remain limited in capturing the full value of genomic technologies. In particular, they struggle to incorporate non-health benefits, such as personal utility and the “value of knowing” (18, 19, 21, 58). There is growing methodological interest in expanding current frameworks by integrating patient preferences through discrete choice experiments (22), or connecting diagnostic yield with long-term survival and quality-of-life outcomes (20). Nonetheless, QALYs remain central to current HTA settings and are not fully replaceable (23, 59).

To strengthen future CUAs of genomic technologies, researchers should leverage updated utility data based on validated instruments, consider broader population-representative proxy conditions, and employ structured “impact inventories” to systematically map both health and non-health consequences (27).

Several genomics-specific challenges remain insufficiently addressed in current EEs, including test heterogeneity, patient stratification, incidental findings, data reanalysis, and spillover effects (19, 22, 60). While large genomic initiatives, such as population-based WGS studies (61), may help improve data availability for modeling, most evaluated studies did not yet incorporate these developments.

Notably, no study fully modeled the costs and benefits of incidental findings, despite their potential clinical, psychological, ethical and legal implications (22, 62–64). Similarly, data reanalysis—an increasingly important practice enabling reclassification of variants—was considered in only two studies (36, 37), even though it may substantially increase diagnostic yield over time (65, 66). Cascade testing and reproductive planning, which are particularly relevant in pediatric and familial genetic disorders, were also rarely included, despite their recognized importance (18).

Organizational and implementation aspects were almost entirely neglected. Only one study accounted for genetic counseling costs (36), and none fully evaluated workforce training, service delivery pathways, or infrastructure requirements. These omissions may lead to underestimation of real-world costs and limit the applicability of results for healthcare system planning. Even in the single study exploring sequencing as a population-level screening strategy (42), test delivery and organizational feasibility were not addressed.

To our knowledge, this is the first systematic review focused specifically on full economic evaluations of GS and ES that model health outcomes, with particular attention to how these outcomes are defined, quantified, and incorporated into economic models. This approach provides a structured methodological overview of current practices in a highly debated area of genomic medicine (67–69). Nevertheless, several limitations should be acknowledged. First, our primary objective was methodological rather than comparative; therefore, although we reported key parameters such as ICERs, willingness-to-pay thresholds, and iNMBs, we did not analyze these outcomes further. Second, by restricting inclusion to full EEs measuring health outcomes (i.e., QALYs and LYGs), and excluding other types of economic evaluations such as cost–benefit and cost–consequence analyses, we may have overlooked broader dimensions of value. Cost–benefit analyses can capture wider societal impacts, while cost–consequence analyses allow outcomes to be presented in a disaggregated manner. Third, limiting inclusion to English- and Italian-language publications may have introduced language bias and excluded relevant studies published in other languages; however, this was necessary to ensure accurate appraisal of complex methodological content. Finally, we did not conduct a meta-analysis due to substantial heterogeneity in study designs, cohort characteristics, and comparative strategies across the included studies.

In conclusion, although the value of GS and ES extends beyond clinical outcomes, current economic evaluations—particularly CUAs—struggle to capture this broader impact. Evidence on their cost-effectiveness in terms of QALYs remains limited and varies between treatment guidance and rare disease diagnosis. Future evaluations should incorporate long-term outcomes and real-world data, improve uncertainty handling, and better capture broader elements of value, including personal utility and organizational implications, to support informed and equitable decision-making.