Skip to main content
Download PDF

Cost-effectiveness of novel diagnostic tools for idiopathic pulmonary fibrosis in the United States

BMC Health Services Research volume 25, Article number: 385 (2025) Cite this article

Abstract

Objectives

Novel non-invasive machine learning algorithms may improve accuracy and reduce the need for biopsy when diagnosing idiopathic pulmonary fibrosis (IPF). We conducted a cost-effectiveness analysis of diagnostic strategies for IPF.

Methods

We developed a decision analytic model to evaluate diagnostic strategies for IPF in the United States. To assess the full spectrum of costs and benefits, we compared four interventions: a machine learning diagnostic algorithm, a genomic classifier, a biopsy-all strategy, and a treat-all strategy. The analysis was conducted from the health sector perspective with a lifetime horizon. The primary outcome measures were costs, Quality-Adjusted Life-Years (QALYs) gained, and Incremental Cost-Effectiveness Ratios (ICERs) based on the average of 10,000 probabilistic runs of the model.

Results

Compared to a biopsy-all strategy the machine learning algorithm and genomic classifer reduced diagnostic-related costs by $14,876 and $3,884, respectively. Use of the machine learning algorithm consistently reduced diagnostic costs. When including downstream treatment costs and benefits of anti-fibrotic treatment, the machine learning algorithm had an ICER of $331,069 per QALY gained compared to the biopsy-all strategy. The genomic classifier had a higher ICER of $390,043 per QALY gained, while the treat-all strategy had the highest ICER of $3,245,403 per QALY gained. Results were sensitive to changes in various input parameters including IPF treatment costs, sensitivity and specificity of novel screening tools, and the rate of additional diagnostics following inconclusive results. High treatment costs were found to drive overall cost regardless of the diagnostic method. As treatment costs lowered, the supplemental diagnostic tools became increasingly cost-effective.

Conclusions

Novel tools for diagnosing IPF reduced diagnostic costs, while overall incremental cost-effectiveness ratios were high due to treatment costs. New IPF diagnosis approaches may become more favourable with lower-cost treatments for IPF.

Peer Review reports

Introduction

Idiopathic Pulmonary Fibrosis (IPF) is the most common Interstitial Lung Disease (ILD), a category of relatively uncommon, chronic lung diseases which lead to substantial morbidity and mortality due to progressive lung function impairment [1, 2]. IPF occurs primarily in older adults, particularly over 75 [3], and is more common among men and people who smoke cigarettes [2]. IPF is caused by the interaction of multiple environmental and genetic factors, and repetitive local micro-injuries to aging lung tissue with no singular cause identified [2, 4]. While the disease trajectory varies for patients, the prognosis is generally poor, with high morbidity and a median survival of 3.8 years [5]. Therapeutics have come to market in the past decade, which slow the decline in lung function and reduce respiratory events, such as acute exacerbations [6,7,8,9]. Pooled analyses suggest there may be improvements to survival with long-term treatment [10,11,12,13]. However, these medications come with high costs relative to their benefits [14, 15]. Given their high cost and the still developing evidence of the benefits of treatment on survival, the accurate and early diagnosis of IPF is essential to ensuring the maintenance of quality of life.

Current guidelines for diagnosing IPF recommend the collection of detailed clinical history to rule out causes of other ILD (e.g. chronic exposure to birds associated with hypersensitivity pneumonitis), pulmonary function testing, a variety of blood-based inflammatory biomarkers, and chest imaging with high-resolution computed tomography [16, 17]. Multidisciplinary discussion (MDD) is used to determine whether or not a supplemental assessment is necessary. The overlap between ILD subtypes is significant and subjective interpretation of imaging leads to diagnostic disagreement among clinicians and even subspecialized pulmonologists [18]. Such disagreement can exacerbate existing delays in referral to specialist care [19]. Recommended supplemental diagnostics includes either surgical lung biopsy or bronchoalveolar lavage, with surgical lung biopsy representing a key element of the current gold standard for diagnosis. Surgical lung biopsies are costly [20], come with a severe risk of death and complications [21] and have some limitations to the accuracy of histological confirmation [22].

Recently, novel screening tools have emerged for the diagnosis of IPF that avoid high cost and potentially life threatening surgical lung biopsy. A pathology-based genomics classification system was developed [23, 24]. However, it still requires an invasive transbronchial biopsy with collection of at least 4–5 tissue samples and can result in mixed effects [23]. Machine learning tools which analyze medical images also present another opportunity for non-invasive diagnostics [25, 26]. In pulmonary care, machine learning has been beneficial in interpreting chronic obstructive pulmonary disease and pulmonary function testing [27]. Initial research suggests that such algorithms can aid in the assessment of ILD cases, although their impact on survival is unknown [28].

The adoption of non-invasive diagnostic algorithms may provide multiple potential benefits in the diagnosis of IPF. Among those cases that an MDD would refer to a biopsy, a diagnostic algorithm may divert patients away from a biopsy with an accurate, non-invasive diagnosis. A supplemental diagnostic algorithm could potentially reduce the number of costly biopsies and may prevent biopsy-caused morbidity and mortality. Much remains unknown about the efficacy and long-term impacts of using machine learning algorithms to diagnose IPF.

We conducted a cost-effectiveness analysis of a novel machine learning algorithm, Fibresolve™, and a genomic classifier, Envisia™, as supplements to an MDD where a high degree of confidence in a diagnosis is lacking to determine the costs and benefits of alternative diagnostics for IPF. The machine learning algorithm has been shown to improve the classification of IPF cases compared to MDD alone without the use of invasive diagnostic measures [29]. The classifier algorithm uses CT imaging and is targeted to serve as an adjunct to MDD in confirming cases of IPF non-invasively. Details of the development of the algorithm are available elsewhere [29]. This model results in a binary classification related to IPF diagnosis, and does not evaluate simply usual interstitial pneumonia (UIP) patterns. The genomic classifier has also been shown to serve as a viable alternative to surgical lung biopsy, and details are available elsewhere [23, 24]. This cost-effectiveness analysis provides clinicians and health systems with information on the potential costs and effects of novel diagnostics for IPF, a high cost chronic illnesses.

Methods

Model overview

We developed a decision analytic model to evaluate the effects of a machine learning algorithm and genomic classifier for the diagnosis of IPF in patients with chronic lung disease with probable to indeterminate UIP patterns based on chest computed tomography (Fig. 1). We adopted a health sector perspective and focused the analysis on patients undergoing dedicated work-up for ILD for whom an MDD had residual uncertainty about the diagnostic subtype (e.g. IPF, hypersensitivity pneumonitis, etc.), before the use of invasive procedures. The cost-effectiveness impact inventory is presented in Appendix 1 [30]. We compared the machine learning algorithm and genomic classifier to two hypothetical strategies that represent opposite ends of the theoretical spectrum with respect to IPF management – a biopsy-all strategy and a treat-all strategy – providing the full range of potential costs and benefits. The scenarios reflect the use of the machine learning algorithm and genomic classifier as a supplement to MDD, where the algorithm and classifier provide additional information on cases the MDD is unable to classify.

Fig. 1
figure 1

Model schematic. Legend: Blue square represents a decision node which indicates the point at which a decision is made; green circle represents a probability node at which point there are multiple uncertain outcomes; red triangle represents a terminal node at which point there is a final outcome which includes both costs and health outcomes. Green A indicates that the area of the tree not shown uses the same structure presented in the green box with parameters updated for the genomic classifier. Orange B indicates that the area of the tree not shown uses the same structure presented in the orange box with parameters upated for the biopsy-all strategy. Yellow C indicates that the area of the tree not shown uses the same structure presented in the yellow box with parameters upated for the biopsy-all strategy

Full size image

The model is simulated using Tree Age Pro Health Care 2022. Individuals in the algorithm and genomic classifier arms who were classified as having IPF were allocated to true positive or false positive states. Those with inconclusive diagnoses from these first-line diagnostics may receive a biopsy. If patients in this group did not undergo a biopsy they were not treated and received costs and benefits associated with a true negative or false negative state. Individuals who underwent a biopsy, in either the algorithm, genomic classifier or biopsy arms, faced a risk of death from the biopsy. Patients who underwent biopsy and did not die were allocated to either a true positive, false positive, true negative, or false negative state and accrued treatment benefits, harms, and costs accordingly. True positives received treatment for IPF using nintedanib, a standard-of-care antifibrotic therapy. Negative cases were provided symptom management, assuming that both true and false negatives had a similar prognosis. In the treat-all arm, all patients were in either a true positive or false positive state and accrued respective treatment costs, benefits, and harms. We assumed that all states were final and that no additional follow-up care would later identify false negative or false positive cases.

Model inputs

Model input parameters are presented in Table 1.

Table 1 Model inputs
Full size table

Probabilities

We set the incidence of IPF at 40.4%, as this was the baseline incidence of IPF in the pivotal clinical study of the machine learning algorithm [29]. We varied this in sensitivity analysis.

The sensitivity and specificity of the machine learning screening algorithm came from a retrospective analysis of patients with ILD collected from US clinical sites [29]. The sensitivity and specificity of the genomic classifier came from a recent meta-analysis [32]. The sensitivity and specificity of biopsy for the identification of IPF are based on retrospective studies and assumptions consistent with previous modelling studies [33, 37].

Given the uncertainty in how the algorithm and classifier’s inconclusive results would be interpreted by clinicians, the probability of biopsy following an inconclusive screen by the algorithm or the genomic classifier was set initially at 0.75. We assumed the rate did not vary between the algorithm and genomic classifier arms.

The probability of dying from a surgical biopsy can range considerably from 3–16.7% [38, 39]. As a base case estimate, we used 6.7% [21].

Costs

Costs were set at $5,000 for the algorithm (gross price parity with genomic classifiers in similar use cases) and $16,963 for the genomic classifier (estimated based on Medicare coverage [31] and the cost of bronchoscopy [20]) in base case analyses. The total lifetime expected costs of treatment for IPF treatment with nintedanib (including all follow-up costs such as pulmonary office visits and home oxygen) and the total lifetime costs of symptom management were based on a study of the cost-effectiveness of IPF treatments in the US [15]. Additional costs included the cost of a surgical lung biopsy [20]. All costs were inflated to 2022 dollars using the Bureau of Labor Statistics Consumer Price Index for Medical Care Services.

Quality of life

Limited quality-of-life data exists for IPF. Rather than directly model quality of life linked to the decay of forced vital capacity, as done in CEAs of IPF treatment [15, 40,41,42], we used lifetime benefits from IPF treatment or symptom management as determined by a previous cost-effectiveness analysis of nintedanib and pirfenidone in the US [15]. The quality of life with treatment was used for individuals who were correctly allocated to the true positive state and false positive with treatment benefits. The quality of life with symptom management was assigned to patients allocated to negative states. The quality of life with symptom management is also used for individuals who are false positive but do not have any treatment benefit with an additional negative quality of life adjustments for potential treatment complications. The most common adverse events from nintedanib were gastrointestinal, and serious adverse events were experienced by roughly 30% of trial participants [9, 13]. To account for this, we assume 30% of patients receive a 0.1 disutility [36]. Finally, individuals who undergo a biopsy have a small quality of life decrement of 0.013 due to the biopsy procedure [34, 35].

Analysis

Given the uncertainty in numerous model input parameter including how the results of the algorithm and classifier may be used by clinicians, changing costs for IPF treatments, poor data on quality of life, and impact of treatment on survival, we present average results from 10,000 probablistic simulations in lieu of expected value from the model and 95% confidence intervals (95% CIs) for costs and QALYs. Distributions for model input parameters used in the Monte Carlo simulation are specified in Table 1. We compared costs and health improvements from each intervention arm. We consider the impact of interventions on total costs as well as diagnostic-only costs. Incremental cost-effectiveness ratios (ICERs) were calculated and cost-effectiveness acceptability curves, which present the probability that an intervention is deemed cost-effective at a specific threshold, were plotted. Strategies that had greater costs and fewer QALYs were dominated.

Scenario and sensitivity analyses

We conducted a series of scenario analyses to explore the impact that changes in model assumptions had on outcomes. Our scenario analyses (1) examine changes in base case results with 25% and 50% reductions in lifetime treatment costs of IPF given that cheaper generic treatment options are becoming available; (2) determine the potential effects of receiving treatment following an inconclusive result from the algorithm or genomic classifier without being biopsied; and (3) potential for patients in the false positive state to receive treatment benefits. Recommendations on the use of antifibrotic treatments is a rapidly evolving area, and evidence on the potential for clinical benefit from the treatment of other progressive types of ILD is limited [43].

To further explore the uncertainty around different model input parameters, we conducted multiple one-way sensitivity analyses where a single parameter is varied while all others are held constant. We further evaluated the impact of changes in the costs of IPF treatments, as well as examining the effect of the sensitiviy and specificity of the supplemental screening tools, the prevalence of IPF in the sample undergoing diagnostics, and the proportion of inconclusive algorithim or genomic classifier results that receive subsequent biopsy.

In addition to ICERs, sensitivity analyses are reported in terms of net monetary benefit (NMB), which provides a single numeric value of the economic benefit of an intervention, measured in dollars, where an intervention with the highest NMB is seen as most favourable. NMB is calculated by converting the health benefit of an intervention into dollars based on a willingness-to-pay per QALY gained and then subtracting the total costs of an intervention [44]. NMB provides a simple way to compare strategies directly across disparate sensitivity analyses. A negative NMB indicates that an intervention’s opportunity costs outweigh benefits, while a positive NMB indicates the intervention provides health benefits beyond its opportunity costs [44]. Given the lack of an established willingness-to-pay threshold in the United States [45], we evaluate sensitivity analyses at multiple thresholds ($50,000, $100,000, $150,000, $200,000, and $250,000 per QALY).

Results

Effectiveness

Our base case probabilistic analysis found biopsy-all resulted in the lowest number of QALYs (3.63, 95% CI 3.62–3.64), followed by the machine learning algorithm with 3.75 (95% CI 3.74–3.76) QALYs, the genomic classifier with 3.77 (95% CI 3.76–3.78) QALYs, and treat-all with 3.87 (95% CI 3.85–3.88) QALYs (Table 2).

Table 2 Cost-effectiveness of idiopathic pulmonary fibrosis diagnosis strategies with 95% confidence intervals from probabilistic analyses
Full size table

The use of the machine learning algorithm and genomic classifiers were associated with small increases in true positive (1.95 and 4.62% points, respectively) and increased false positives (6.36 and 3.04% points, respectively) compared to the biopsy-all strategy (Appendix 2). The algorithm and classifier reduced biopsy-related mortality by 3.03 and 3.15% points, respectively (Appendix 3). Treat-all increased the number of true positives and led to a ten-fold increase in false positive cases compared to biopsy-all.

Costs

Regarding diagnostic costs, biopsy-all was the most costly strategy when considering diagnostic costs only (Table 3). When compared to the biopsy-all strategy, the machine learning algorithm and genomic classifier both reduced diagnostic-only costs by $14,876 and $3,884, respectively. The treat-all strategy has no additional diagnostic costs, and as such, reduced costs by $42,088 compared to biopsy-all.

Table 3 Cost breakdowns of idiopathic pulmonary fibrosis diagnosis strategies with 95% confidence intervals from probabilistic analyses
Full size table

Conversely, when incorporating both long-term treatment and diagnostic costs and benefits, biopsy-all was the least costly strategy with average lifetime costs of $345,680 (95% CI $341,807-$349,554) per patient (Tables 2 and 3, Appendix 2). The total lifetime costs with the machine learning algorithm was the next least costly intervention, with average lifetime costs of $385,585 (95% CI $380,881-$390,290) per patient followed by the genomic classifier at $393,020 (95% CI $388,379-$397,660) per patient and treat-all with average lifetime costs of $715,949 (95% CI $705,517-$726,381) per patient. In each intervention pathway, total lifetime costs increase as more individuals are allocated to treatment states in the model.

Cost-effectiveness of long-term treatment plus diagnostic options

ICERs for each of the intervention pathways from our probabilistic analysis with a health care sector perspective are reported in Table 2. Compared to a biopsy-all diagnostic strategy, the machine learning diagnostic algorithm pathway had an ICER of $331,069 per QALY when including the value of long-term treatment and diagnostic costs and benefits. The genomic classifier diagnostic pathway had an ICER of $390,043 per QALY compared to the algorithm pathway. Finally, the treat-all pathway had an ICER of $3,245,403 per QALY compared to the genomic classifier diagnostic pathway.

The cost-effectiveness acceptability curve illustrates the uncertainty regarding the cost-effectiveness of these intervention pathways (Fig. 2). At low willingness-to-pay levels, biopsy-all is the most attractive diagnostic and treatment selection strategy. As willingness-to-pay increases, biopsy-all becomes increasingly less attractive. Yet, even at a willingness-to-pay threshold as high as $250,000/QALY, the machine learning algorithm and genomic classifier approaches were unlikely to be deemed cost-effective with similar probabilities (23.1% and 21.6%, respectively), while the biopsy-all strategy had a 42.1% chance of being cost-effective.

Fig. 2
figure 2

Cost-effectiveness acceptability curve. Legend: The figure presents probability that a given intervention is deemed cost-effectiveness based at a given willingness-to-pay threshold. The probability of being deemed cost-effective is based on 10,000 Monte Carlo simulations of the model

Full size image

Scenario analyses

Our scenario analyses that examine the effect of reduced lifetime IPF treatment costs found increased probability of cost-effectiveness for the machine learning algorithm, genomic classifier, and treat-all strategies. With a 25% reduction in lifetime treatment costs, the low probability that the machine learning algorithm and the genomic classifier are cost-effective is relatively unchanged (Appendix 4). Reductions of treatment costs by 50% result in more favourable ICERs for the machine learning algorithm and an increased probability of cost-effectiveness for the treat-all strategy (Appendix 4). In our one-way sensitivity analysis of lifetime treatment costs for IPF, we found that with lifetime treatment costs between ~$420,000 and ~$180,000 the machine learning algorithm diagnostic pathway may be cost-effective (Appendix 5). At lifetime treatment costs below roughly $177,000, treat-all becomes the most cost-effective strategy.

Our analyses which examined potential treatment use among those with inconclusive diagnostic results and treatment benefits for individuals with false positive results did not find meaningful improvements in cost-effectiveness (results not shown).

Sensitivity analyses

Inconclusive algorithm results referred to biopsy

We examined the change in net monetary benefit as the probability that an inconclusive algorithm result is referred to biopsy ranges from 0 to 1 (Appendix 6). At high willingness-to-pay thresholds, the machine learning biopsy and genomic classifier become more cost-effective than biopsy all when the biopsy referral rate is less than 0.66. With referral rates reduced further, the machine learning algorithm and genomic classifier become increasingly attractive even at lower willingness-to-pay thresholds. At lower levels, these strategies reduce biopsy costs, biopsy-related deaths, and the number of individuals who are allocated to high-cost treatment arms.

When examining diagnostic costs alone, the use of the machine learning algorithm reduced diagnostic costs compared to the biopsy-all strategy by $37,373 to $7,609, depending on the biopsy referral rate. Even when all inconclusive results were referred for biopsy, the machine learning algorithm reduced overall diagnostic costs by $7,609. Similar results were seen for the genomic classifier. When no inconclusive results were referred for biopsy, the genomic classifier pathway reduced diagnostic costs by $25,410. However, when all inconclusive results were referred for biopsy, the genomic classifier pathway increased diagnostic costs by $3,302.

Sensitivity and specificity of supplemental diagnostics

Given the similarities in the test characteristics of the machine learning algorithm and genomic classifier, the results of our analyses are dependent on the sensitivity and specificity of these tools. In multiple instances when ranging the sensitivity and specificity of the supplemental screening modalities across their 95% confidence intervals, the machine learning algorithm is dominated (either absolutely, where costs are higher and outcomes lower, or through extended dominance, where the next most effective intervention has a lower ICER) by the genomic classifier as sensitivity and specificity change (Appendix 7).

Prevalence of IPF

Finally, we examined the effect of changes in the prevalence of IPF in the sample undergoing supplemental diagnostics (Appendix 8). As the prevalence of IPF increases, the ICER for the machine learning algorithm diagnostic pathway becomes increasingly favourable. However, it is not cost-effective even with a high willingness-to-pay thresholds and 60% prevalence of IPF, the upper end of our sensitivity analysis. At the lower end (30% prevalence), the machine learning algorithm pathway is dominated by the genomic classifier pathway given the algorithm’s higher false positive rate.

Discussion

Our analysis highlights the complicated role that diagnostic technologies play in identifying diseases with limited treatment options. IPF is challenging to diagnose and expensive to treat [14, 15, 17]. Use of the machine learning algorithm reduced diagnostic costs by $14,876 in the base case, ranging from $7,609 to $37,373 in sensitivity analyses. The genomic classifier also reduced diagnostic costs, but by a smaller margin. These results are based on the effects of the supplemental diagnostic tools in the target population and reducition in surgical lung biopies following the use of supplemental diagnostics. Both the machine learning algorithm and genomic classifier reduced biopsies and biopsy-related deaths compared to a biopsy-all strategy.

When incorporating treatment effects into a comprehensive assessment of long-term costs and benefits of novel diagnostic tools, we find that the machine learning algorithm and genomic classifier have high ICERs relative to the biopsy all strategy. High costs were driven in large part due to the high costs and limited impact of treatment, as has been documented in other cost-effectiveness analyses [15]. As a corollary, the biopsy-all strategy at the current costs for IPF treatments was overall the most cost-effective strategy given greatest selectivity in treatment. The treat-all strategy was least cost-effective.

Sensitivity analyses illustrate that the main driver of cost-effectiveness was the high cost and poor cost-effectiveness of IPF treatment. Across multiple sensitivity analyses, we found reductions in ICERs when more patients were diverted from the high-cost treatments, such as lower sensitivity of the supplemental diagnostic tools and lower rates of biopsy referral. As treatment costs lower, ICERs become more favourable and other parameters, such as quality of life for false positives and true negatives, become increasingly important. Given the novelty of machine learning diagnostics there is little evidence for how such tools might be reimbursed as such our base case reimbursement cost was based on an assumption. However, in sensitivity analysis which explore lower costs in line with other analyses of machine learning tools [46,47,48] we found limited impact on the ICER unless treatment costs were lower.

Given that IPF treatments are currently accepted in clinical practice and recent guidelines suggest a greater emphasis on treatment over surgical lung biopsy [49], perhaps high ICERs are not a barrier to the adoption of new interventions in this clinical space. Alternatively, a higher ICER threshold could be warranted in this setting of rare conditions and diseases with limited treatment options in order to spur innovation [50]. Previous work has found that lower-cost treatments with similar efficacy may be cost-effective [15]. Generic IPF treatments with lower costs are already available. The cost of generic Pirfenidone is considerably less than the brand name. Additionally, Medicare announced in January of 2025 that nintedanib, one of the primary treatments for IPF and the treatment included in our analysis, is part of the next round of price negotiations. Reductions in the cost of treatment would, in turn, improve the balance of costs and benefits of new IPF screening modalities.

Our overall findings show that the genomic classifier and machine learning algorithm options reduce diagnostic costs. However, they have a low likelihood of being cost-effective due to high-cost treatments, which are themselves unlikely to be cost-effective at conventional thresholds without lower costs [15]. In terms of diagnostic considerations, results of this analysis show the potential promise of alternative diagnostic strategies for IPF in several respects. Both the machine learning algorithm and genomic classifier reduced the number of biopsies and biopsy-related deaths compared to the biopsy-all strategy. Both strategies also reduced diagnostic-related costs compared to the biopsy-all strategy. For the machine learning algorithm, diagnostic costs were reduced even when all inconclusive findings were referred for biopsy. In our base case analysis, the machine learning algorithm strategy resulted in roughly the same number of true positives as the biopsy-all strategy, while the genomic classifier strategy improved the identification of true positives. When the risk of death from surgical lung biopsy is high, the application of a diagnostic algorithm can provide additional useful information with low risk.

It is also important to consider potential additional benefits that were not explicitly modelled here but that may increase the value of machine learning diagnostic algorithms. Machine learning algorithms may improve in the future. The value of scientific spillover would likely improve cost-effectiveness [51]. Future improvements in the specificity of the screening algorithm, which ensures that people are not allocated to a high-cost treatment, may result in more favourable cost-effectiveness. In addition, these algorithms may provide other benefits to patients by providing the opportunity to avoid surgery. The machine learning algorithm may also reduce delays in diagnosis for patients unable to access a medical center able to effectively diagnose IPF or have the surgical expertise for surgical lung biopsies. Delays in IPF diagnosis have been found to decrease survival [19]. Future modelling efforts should include these broader value considerations [52]. However, these may require alternative model structures which could consider the potential impact that non-invasive diagnostic tools have on reducing the time to diagnosis.

Our results are limited by numerous uncertainties and inconsistencies around how IPF is diagnosed and treated in practice. We make assumptions regarding how clinicians may continue to use biopsies following the results from the machine learning and genomic classifier diagnostic tools. Additionally, we assumed all individuals undergoing biopsy would receive a surgical lung biopsy and not transbronchial lung biopsy or cryobiopsy. While the adoption of these procedures is increasing, due to lower mortality rates, uptake is varied across the country and sensitivity is low [53, 54]. As such, our analysis used the current recommended gold standard, surgical lung biopsy [16, 17]. For treatment, we assumed that patients received treatment following the use of diagnostics and that those with negative screening results for IPF did not receive antifibrotic treatment which may not reflect clinical practice if clinicians are prescribing antifibrotics for all fibrosising interstitial lung diseases [55]. Current evidence from Medicare claims suggests that only 10.4% of IPF patients receive antifibrotics, and one quarter of those are treated prior to a coded IPF diagnosis [56]. While other studies using IPF registries have found rates of antifibrotic treatment closer to 60–70% [57, 58]. The uncertainty and inconsistency points to a need for new diagnostic approaches and additional real world evidence evaluating how IPF and ILD are being diagnosed and treated in clinical practice. Cost-effectiveness analyses should be updated once additional evidence is available.

This analysis is also limted by how we model treatment outcomes. We did not model treatment outcomes directly but instead relied on the results of a previous cost-effectiveness analysis of IPF treatment. As such, this analysis relies on how the study measured and calculated long-term quality-of-life and cost outcomes. Future analyses could try to collect cost and quality-of-life data prospectively directly from patients in a trial. However, those types of trials may not be economical for this patient population.

IPF is challenging to diagnose and treat. Novel diagnostic tools for IPF were found to reduce diagnostic costs, but are unlikely to be cost-effective in current practice given high therapeutic costs. These tools may allow for better diagnosis of IPF through reduced biopsies and diagnostic-related costs. The potential for these tools should be monitored as new IPF treatments become available and as lower-cost generic treatments emerge, which may change the cost-effectiveness result for overall disease management.

Data availability

All data used to develop the model are described and available in the paper. The model is available upon reasonable request from the authors.

References

  1. King TE Jr, Pardo A, Selman M. Idiopathic pulmonary fibrosis. Lancet. 2011;378(9807):1949–61. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s0140-6736(11)60052-4.

    Article  PubMed  Google Scholar 

  2. Richeldi L, Collard HR, Jones MG. Idiopathic pulmonary fibrosis. Lancet. 2017;389(10082):1941–52. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S0140-6736(17)30866-8.

    Article  PubMed  Google Scholar 

  3. Sgalla G, Biffi A, Richeldi L. Idiopathic pulmonary fibrosis: diagnosis, epidemiology and natural history. Respirology. 2016;21(3):427–37. https://doiorg.publicaciones.saludcastillayleon.es/10.1111/resp.12683.

    Article  PubMed  Google Scholar 

  4. Lederer DJ, Martinez FJ. Idiopathic pulmonary fibrosis. N Engl J Med. 2018;378(19):1811–23. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMra1705751.

    Article  PubMed  CAS  Google Scholar 

  5. Raghu G, Chen S-Y, Yeh W-S, et al. Idiopathic pulmonary fibrosis in US medicare beneficiaries aged 65 years and older: incidence, prevalence, and survival, 2001–11. Lancet Respir Med. 2014;2(7):566–72.

    Article  PubMed  Google Scholar 

  6. Costabel U, Inoue Y, Richeldi L, et al. Efficacy of nintedanib in idiopathic pulmonary fibrosis across prespecified subgroups in INPULSIS. Am J Respir Crit Care Med. 2016;15(2):178–85. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201503-0562OC.

    Article  CAS  Google Scholar 

  7. King TE Jr, Bradford WZ, Castro-Bernardini S, et al. A phase 3 trial of Pirfenidone in patients with idiopathic pulmonary fibrosis. N Engl J Med. 2014;29(22):2083–92. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMoa1402582.

    Article  CAS  Google Scholar 

  8. Ley B, Swigris J, Day BM, et al. Pirfenidone reduces respiratory-related hospitalizations in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med. 2017;15(6):756–61. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201701-0091OC.

    Article  Google Scholar 

  9. Richeldi L, du Bois RM, Raghu G, et al. Efficacy and safety of nintedanib in idiopathic pulmonary fibrosis. N Engl J Med. 2014;29(22):2071–82. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMoa1402584.

    Article  CAS  Google Scholar 

  10. Lancaster L, Crestani B, Hernandez P, et al. Safety and survival data in patients with idiopathic pulmonary fibrosis treated with Nintedanib: pooled data from six clinical trials. BMJ Open Respir Res. 2019;6(1):e000397. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmjresp-2018-000397.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Loveman E, Copley VR, Scott DA, Colquitt JL, Clegg AJ, O’Reilly KM. Comparing new treatments for idiopathic pulmonary fibrosis–a network meta-analysis. BMC Pulm Med. 2015;18:37. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12890-015-0034-y.

    Article  CAS  Google Scholar 

  12. Nathan SD, Albera C, Bradford WZ, et al. Effect of Pirfenidone on mortality: pooled analyses and meta-analyses of clinical trials in idiopathic pulmonary fibrosis. Lancet Respir Med. 2017;5(1):33–41. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s2213-2600(16)30326-5.

    Article  PubMed  CAS  Google Scholar 

  13. Richeldi L, Cottin V, du Bois RM, et al. Nintedanib in patients with idiopathic pulmonary fibrosis: combined evidence from the TOMORROW and INPULSIS(®) trials. Respir Med. 2016;113:74–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.rmed.2016.02.001.

    Article  PubMed  Google Scholar 

  14. Collard HR, Chen SY, Yeh WS, et al. Health care utilization and costs of idiopathic pulmonary fibrosis in U.S. Medicare beneficiaries aged 65 years and older. Ann Am Thorac Soc. 2015;12(7):981–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1513/AnnalsATS.201412-553OC.

    Article  PubMed  Google Scholar 

  15. Dempsey TM, Thao V, Moriarty JP, Borah BJ, Limper AH. Cost-effectiveness of the anti-fibrotics for the treatment of idiopathic pulmonary fibrosis in the United States. BMC Pulm Med. 2022;22(1):18. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12890-021-01811-0.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med. 2011;15(6):788–824. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.2009-040GL.

    Article  Google Scholar 

  17. Raghu G, Remy-Jardin M, Myers JL, et al. Diagnosis of idiopathic pulmonary fibrosis. An official ATS/ERS/JRS/ALAT clinical practice guideline. Am J Respir Crit Care Med. 2018;1(5):e44-68. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201807-1255ST.

    Article  Google Scholar 

  18. Walsh SLF, Maher TM, Kolb M, et al. Diagnostic accuracy of a clinical diagnosis of idiopathic pulmonary fibrosis: an international case–cohort study. Eur Respir J. 2017;50(2):1700936. https://doiorg.publicaciones.saludcastillayleon.es/10.1183/13993003.00936-2017.

    Article  PubMed  Google Scholar 

  19. Lamas DJ, Kawut SM, Bagiella E, Philip N, Arcasoy SM, Lederer DJ. Delayed access and survival in idiopathic pulmonary fibrosis: a cohort study. Am J Respir Crit Care Med. 2011;1(7):842–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201104-0668OC.

    Article  Google Scholar 

  20. Chiu YW, Kao YH, Simoff MJ, et al. Costs of biopsy and complications in patients with lung Cancer. Clinicoecon Outcomes Res. 2021;13:191–200. https://doiorg.publicaciones.saludcastillayleon.es/10.2147/ceor.S295494.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Hutchinson JP, Fogarty AW, McKeever TM, Hubbard RB. In-Hospital mortality after surgical lung biopsy for interstitial lung disease in the United States. 2000 to 2011. Am J Respir Crit Care Med. 2016;15(10):1161–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201508-1632OC.

    Article  Google Scholar 

  22. Hutchinson J, Hubbard R, Raghu G. Surgical lung biopsy for interstitial lung disease: when considered necessary, should these be done in larger and experienced centres only? Eur Respir J. 2019;53(2):1900023. https://doiorg.publicaciones.saludcastillayleon.es/10.1183/13993003.00023-2019.

    Article  PubMed  Google Scholar 

  23. Raghu G, Flaherty KR, Lederer DJ, et al. Use of a molecular classifier to identify usual interstitial pneumonia in conventional transbronchial lung biopsy samples: a prospective validation study. Lancet Respir Med. 2019;7(6):487–96. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s2213-2600(19)30059-1.

    Article  PubMed  Google Scholar 

  24. Richeldi L, Scholand MB, Lynch DA, et al. Utility of a molecular classifier as a complement to high-resolution computed tomography to identify usual interstitial pneumonia. Am J Respir Crit Care Med. 2021;15(2):211–20. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.202003-0877OC.

    Article  Google Scholar 

  25. Chan HP, Samala RK, Hadjiiski LM, Zhou C. Deep learning in medical image analysis. Adv Exp Med Biol. 2020;1213:3–21. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/978-3-030-33128-3_1.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Litjens G, Kooi T, Bejnordi BE, et al. A survey on deep learning in medical image analysis. Med Image Anal. 2017;42:60–88. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.media.2017.07.005.

    Article  PubMed  Google Scholar 

  27. Mlodzinski E, Stone DJ, Celi LA. Machine learning for pulmonary and critical care medicine: a narrative review. Pulm Ther. 2020;6(1):67–77. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s41030-020-00110-z.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Walsh SLF, Calandriello L, Silva M, Sverzellati N. Deep learning for classifying fibrotic lung disease on high-resolution computed tomography: a case-cohort study. Lancet Respir Med. 2018;6(11):837–45. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S2213-2600(18)30286-8.

    Article  PubMed  Google Scholar 

  29. Maddali MV, Kalra A, Muelly M, Reicher JJ. Development and validation of a CT-based deep learning algorithm to augment non-invasive diagnosis of idiopathic pulmonary fibrosis. Respir Med. 2023;219:107428. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.rmed.2023.107428.

    Article  PubMed  Google Scholar 

  30. Sanders GD, Neumann PJ, Basu A, et al. Recommendations for conduct, methodological practices, and reporting of cost-effectiveness analyses: second panel on cost-effectiveness in health and medicine. JAMA. 2016;316(10):1093–103. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jama.2016.12195.

    Article  PubMed  Google Scholar 

  31. Morris T. Veracyte receives Adlt status for envisia genomic classifier from centers for medicare and medicaid services. Veracyte. https://investor.veracyte.com/news-releases/news-release-details/veracyte-receives-adlt-status-envisia-genomic-classifier-centers. Accessed 30 Jul 2023.

  32. Kheir F, Uribe Becerra JP, Bissell B, et al. Use of a genomic classifier in patients with interstitial lung disease: a systematic review and meta-analysis. Ann Am Thorac Soc. 2022;19(5):827–32. https://doiorg.publicaciones.saludcastillayleon.es/10.1513/AnnalsATS.202102-197OC.

    Article  PubMed  Google Scholar 

  33. Shih AR, Nitiwarangkul C, Little BP, et al. Practical application and validation of the 2018 ATS/ERS/JRS/ALAT and Fleischner society guidelines for the diagnosis of idiopathic pulmonary fibrosis. Respir Res. 2021;22(1):124. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12931-021-01670-7.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Bendixen M, Jørgensen OD, Kronborg C, Andersen C, Licht PB. Postoperative pain and quality of life after lobectomy via video-assisted thoracoscopic surgery or anterolateral thoracotomy for early stage lung cancer: a randomised controlled trial. Lancet Oncol. 2016;17(6):836–44. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s1470-2045(16)00173-x.

    Article  PubMed  Google Scholar 

  35. Feller-Kopman D, Liu S, Geisler BP, DeCamp MM, Pietzsch JB. Cost-effectiveness of a bronchial genomic classifier for the diagnostic evaluation of lung Cancer. J Thorac Oncol. 2017;12(8):1223–32. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jtho.2017.04.030.

    Article  PubMed  Google Scholar 

  36. Lloyd A, Nafees B, Narewska J, Dewilde S, Watkins J. Health state utilities for metastatic breast cancer. Br J Cancer. 2006;18(6):683–90. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/sj.bjc.6603326.

    Article  Google Scholar 

  37. Hunninghake GW, Zimmerman MB, Schwartz DA, et al. Utility of a lung biopsy for the diagnosis of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med. 2001;15(2):193–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/ajrccm.164.2.2101090.

    Article  Google Scholar 

  38. Carrillo G, Estrada A, Pedroza J, et al. Preoperative risk factors associated with mortality in lung biopsy patients with interstitial lung disease. J Invest Surg. 2005;18(1):39–45. https://doiorg.publicaciones.saludcastillayleon.es/10.1080/08941930590905206.

    Article  PubMed  CAS  Google Scholar 

  39. Utz JP, Ryu JH, Douglas WW, et al. High short-term mortality following lung biopsy for usual interstitial pneumonia. Eur Respir J. 2001;17(2):175–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1183/09031936.01.17201750.

    Article  PubMed  CAS  Google Scholar 

  40. Clay E, Cristeau O, Chafaie R, Pinta A, Mazaleyrat B, Cottin V. Cost-effectiveness of Pirfenidone compared to all available strategies for the treatment of idiopathic pulmonary fibrosis in France. J Mark Access Health Policy. 2019;7(1):1626171. https://doiorg.publicaciones.saludcastillayleon.es/10.1080/20016689.2019.1626171.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Rinciog C, Diamantopoulos A, Gentilini A, et al. Cost-effectiveness analysis of nintedanib versus Pirfenidone in idiopathic pulmonary fibrosis in Belgium. Pharmacoecon Open. 2020;4(3):449–58. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s41669-019-00191-w.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  42. Rinciog C, Watkins M, Chang S, et al. A cost-effectiveness analysis of nintedanib in idiopathic pulmonary fibrosis in the UK. Pharmacoeconomics. 2017;35(4):479–91. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s40273-016-0480-2.

    Article  PubMed  CAS  Google Scholar 

  43. Finnerty JP, Ponnuswamy A, Dutta P, Abdelaziz A, Kamil H. Efficacy of antifibrotic drugs, nintedanib and pirfenidone, in treatment of progressive pulmonary fibrosis in both idiopathic pulmonary fibrosis (IPF) and non-IPF: a systematic review and meta-analysis. BMC Pulm Med. 2021;21(1):411. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12890-021-01783-1.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  44. Paulden M. Calculating and interpreting icers and net benefit. Pharmacoeconomics. 2020;38(8):785–807. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s40273-020-00914-6.

    Article  PubMed  Google Scholar 

  45. Neumann PJ, Kim DD. Cost-effectiveness thresholds used by study authors, 1990–2021. JAMA. 2023;329(15):1312–4. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jama.2023.1792.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Hlatky MA, Wilding S, Stuart B, et al. Randomized comparison of chest pain evaluation with FFR(CT) or standard care: factors determining US costs. J Cardiovasc Comput Tomogr. 2023;17(1):52–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jcct.2022.09.005.

    Article  PubMed  Google Scholar 

  47. Mamta B, Mina M, Elizabeth S, et al. LiverMultiScan as an alternative to liver biopsy to monitor autoimmune hepatitis in the National health service in England: an economic evaluation. BMJ Open. 2022;12(9):e058999. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmjopen-2021-058999.

    Article  Google Scholar 

  48. CMS sets National payment rate for the HeartFlow analysis in Doctors’ Offices and imaging centers. HeartFlow. https://www.heartflow.com/newsroom/cms-sets-national-payment-rate-for-the-heartflow-analysis-in-doctors-offices-and-imaging-centers/. Accessed 23 Jul 2023.

  49. Choe J, Kwon BS, Do KH, Hwang HS, Song JW, Chae EJ. Diagnostic and prognostic implications of 2018 guideline for the diagnosis of idiopathic pulmonary fibrosis in clinical practice. Sci Rep. 2021;13(1):16481. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41598-021-95728-7.

    Article  CAS  Google Scholar 

  50. Danzon PM, Drummond MF, Towse A, Pauly MV. Objectives, budgets, thresholds, and opportunity costs—a health economics approach: an ISPOR special task force report [4]. Value Health. 2018;21(2):140–5. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jval.2017.12.008.

    Article  PubMed  Google Scholar 

  51. Garrison L, Mestre-Ferrandiz J, Zamora B. The value of knowing and knowing the value: improving the health technology assessment of complementary diagnostics. London; Office of Health Economics; 2016.

  52. Garrison LP Jr, Kamal-Bahl S, Towse A. Toward a broader concept of value: identifying and defining elements for an expanded cost-effectiveness analysis. Value Health. 2017;20(2):213–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jval.2016.12.005.

    Article  PubMed  Google Scholar 

  53. Romagnoli M, Colby TV, Berthet JP, et al. Poor concordance between sequential transbronchial lung cryobiopsy and surgical lung biopsy in the diagnosis of diffuse interstitial lung diseases. Am J Respir Crit Care Med. 2019;15(10):1249–56. https://doiorg.publicaciones.saludcastillayleon.es/10.1164/rccm.201810-1947OC.

    Article  Google Scholar 

  54. Baba T, Takemura T, Okudela K, et al. Concordance between transbronchial lung cryobiopsy and surgical lung biopsy for interstitial lung disease in the same patients. BMC Pulm Med. 2023;23(1):279. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12890-023-02571-9.

    Article  PubMed  PubMed Central  Google Scholar 

  55. Flaherty KR, Wells AU, Cottin V, et al. Nintedanib in progressive fibrosing interstitial lung diseases. N Engl J Med. 2019;31(18):1718–27. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMoa1908681.

    Article  Google Scholar 

  56. Herberts MB, Teague TT, Thao V, et al. Idiopathic pulmonary fibrosis in the United States: time to diagnosis and treatment. BMC Pulm Med. 2023;23(1):281. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12890-023-02565-7.

    Article  PubMed  PubMed Central  Google Scholar 

  57. Salisbury ML, Conoscenti CS, Culver DA, et al. Antifibrotic drug use in patients with idiopathic pulmonary fibrosis. Data from the IPF-PRO registry. Ann Am Thorac Soc. 2020;17(11):1413–23. https://doiorg.publicaciones.saludcastillayleon.es/10.1513/AnnalsATS.201912-880OC.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Holtze CH, Freiheit EA, Limb SL, et al. Patient and site characteristics associated with pirfenidone and nintedanib use in the United States; an analysis of idiopathic pulmonary fibrosis patients enrolled in the Pulmonary Fibrosis Foundation Patient Registry. Respir Res. 2020;21(1):48. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12931-020-1315-4.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

Download references

Acknowledgements

None.

Funding

Imvaria, Inc.

Author information

Authors and Affiliations

  1. Department of Health Management and Policy, School of Public Health, University of Michigan, 1420 Washington Heights, Ann Arbor, MI, 48109-2013, USA

    Christopher J. Cadham & David W. Hutton

  2. Imvaria, Inc, Berkeley, CA, USA

    Joshua Reicher & Michael Muelly

Authors
  1. Christopher J. Cadham
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Joshua Reicher
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Michael Muelly
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. David W. Hutton
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

The authors contributed to the paper as follows: Study conception and design: CJC, JR, MM, DWH; Acquisition of data: CJC Analysis and interpretation of results: CJC, JR, MM, DWH; Draft manuscript preparation: CJC Critical revision of the manuscript for important intellectual content: CJC, JR, MM, DWH Administrative, technical, or logistic support: CJC Supervision: DWH All authors reviewed the results and approved the final version of the manuscript.

Corresponding author

Correspondence to Christopher J. Cadham.

Ethics declarations

Ethics approval and consent to participate

Not applicable, this study does not involve any human subjects research.

Consent for publication

Not applicable.

Competing interests

CJC received research support funding from Imvaria, Inc for the conduct of this study. JR and MM are a co-founders, employees and shareholders of Imvaria, Inc. DWH’s employer, the Univeristy of Michigan, received payment from Imvaria, Inc. for the creation of this analysis.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cadham, C.J., Reicher, J., Muelly, M. et al. Cost-effectiveness of novel diagnostic tools for idiopathic pulmonary fibrosis in the United States. BMC Health Serv Res 25, 385 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12913-025-12506-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12913-025-12506-1

Keywords

BMC Health Services Research

ISSN: 1472-6963

Contact us