Budget impact from the incorporation of positron emission tomography – computed tomography for staging lung cancers

OBJECTIVE To estimate the budget impact from the incorporation of positron emission tomography (PET) in mediastinal and distant staging of non-small cell lung cancer. METHODS The estimates were calculated by the epidemiological method for years 2014 to 2018. Nation-wide data were used about the incidence; data on distribution of the disease´s prevalence and on the technologies’ accuracy were from the literature; data regarding involved costs were taken from a micro-costing study and from Brazilian Unified Health System (SUS) database. Two strategies for using PET were analyzed: the offer to all newly-diagnosed patients, and the restricted offer to the ones who had negative results in previous computed tomography (CT) exams. Univariate and extreme scenarios sensitivity analyses were conducted to evaluate the influence from sources of uncertainties in the parameters used. RESULTS The incorporation of PET-CT in SUS would imply the need for additional resources of 158.1 BRL (98.2 USD) million for the restricted offer and 202.7 BRL (125.9 USD) million for the inclusive offer in five years, with a difference of 44.6 BRL (27.7 USD) million between the two offer strategies within that period. In absolute terms, the total budget impact from its incorporation in SUS, in five years, would be 555 BRL (345 USD) and 600 BRL (372.8 USD) million, respectively. The costs from the PET-CT procedure were the most influential parameter in the results. In the most optimistic scenario, the additional budget impact would be reduced to 86.9 BRL (54 USD) and 103.8 BRL (64.5 USD) million, considering PET-CT for negative CT and PET-CT for all, respectively. CONCLUSIONS The incorporation of PET in the clinical staging of non-small cell lung cancer seems to be financially feasible considering the high budget of the Brazilian Ministry of Health. The potential reduction in the number of unnecessary surgeries may cause the available resources to be more efficiently allocated.

Economic evaluation of diagnostic and therapeutic interventions is gaining importance to support decisions concerning the incorporation and dissemination of new health care technologies. 22 Those analyses, however, do not provide all necessary information for decision-making, as they do not assess the feasibility for the introduction of the best alternative considering available budgets. 11 The further conduction of budget impact analyses to evaluate short and medium-term financial consequences regarding the incorporation, changed use, or withdrawal of a technology from the set of available interventions in the health care system is required. 2,8 Brazil reports a high number of lung cancer cases: 27,330 new cases are estimated for 2014. a Non-small cell lung carcinomas (NSCLC) account for 75.0%-85.0% of cases, which can be potentially cured with surgical resection in the localized disease. 5,b Often, the diagnosis is achieved in advanced stages. Thus, due to the disease spread to mediastinal lymph nodes or distant metastases at the time of diagnosis, only 20.0% of patients are considered operable. c Evaluating the disease extension at the diagnosis is essential for defining therapies. That avoids improper procedures which can influence patients' survival and quality of life. The clinical staging is mainly conducted by means of computed tomography of the thorax and upper abdomen (CT of thorax), according to the clinical guidelines for the diagnosis and treatment of lung cancer, as disclosed by the Brazilian Ministry of Health (MH) in 2012. d That exam is mainly based in morphological changes.
Positron emission tomography (PET) which is either combined to computed tomography (PET-CT) or not, is based on metabolic activity, rather than only on anatomical aspects. Both are more accurate than conventional imaging techniques in the evaluation of mediastinal and in distant areas involvement. 5 Its inclusion in the traditional diagnostic strategies may result in better management of cases, with reduced numbers of unnecessary surgeries 21,23 and decreased morbidity and mortality. Another advantage would be staging the lung disease and distant metastases with a single exam. 14

INTRODUCTION
PET is starting to be disseminated in Brazil, and it was included in the Brazilian Unified Health System (SUS) payrolls for procedures in April 2014. e The economic evaluation for the use of PET-CT in the staging of NSCLC, conducted for the MH in 2013, found that PET-CT is more cost-effective when compared to the currently offered management strategy, which is CT-based. b The results confirm international findings, 4 which show benefits in its inclusion for the staging of NSCLC patients, mainly for preventing unnecessary surgeries, that pay off for the additional costs for using of the new technology.
The study from 2013 did not evaluate the financial impacts from offering the procedure in Brazil's public health care service network. Budget impact analyses are scarce in Brazil, especially concerning diagnostic imaging. In a health care system which is set to offer universal and comprehensive care, the concern with using resources is shown to be important considering the dichotomic relationship among budget availability, extension of care, and continuous advancements in technology.
This study aimed to estimate the budget impact of the inclusion of PET-CT in the mediastinal and distant staging of non-small cell lung cancer.

METHODS
The budget impact estimation has adopted SUS's perspective as a financing agent of health care services, as indicated by the Brazilian guideline. f The chosen horizon was a five-year one (2014 to 2018), considering the possible morosity in the reallocation of government budgets and restrictions in the availability and access to PET-CT. were estimated by admitting a 75.0% coverage for the SUS-supported patient population. g Three analysis scenarios were defined: reference (strategies of management that are widely used, based on CT of thorax for all patients); alternative 1 (use of PET-CT restricted to patients with previous negative CT results, allowing for coverage of situations with more limited access to PET-CT); alternative 2 (use of CT and PET-CT for all cases, with further clinical management being defined by the combined results of the two examsonly patients with both negative images would directly proceed to pulmonary resection). This last strategy yielded a higher reduction in the number of unnecessary surgeries in the cost-effective study used as basis, b with small differences in the incremental cost-effectiveness ratio between the two usage methods for PET-CT in the conducted sensitivity analyses.
Only direct costs of procedures involved in the staging and therapies of patients were considered ( Table 2). As the PET-CT procedure was not included in SUS payrolls when the analyses were conducted, we used values as estimated by micro-costing. 3 The values were calculated again to have a 30.0% reduction in the F18-fluoro-2-deoxy-D-glucose costs ( 18 FDG), h to consider the recent increase in the number of private input producers which took place when the Federal Government lost its monopoly for radiopharmaceuticals in 2006. For all procedures figuring in SUS payroll charts, values regarding November 2013 were used, which were listed in SUS Management System for the Chart of Procedures, Medications, and Orthoses, Prosthetics, and Special Materials. i For the budget impact estimates, the same decision trees and parameters that were used in the cost-effectiveness study, conducted for MH in 2013, were used again here. b The new cases projected for each year and the costs of procedures fed the trees related to each analysis scenario, which generated estimates for quantities of conducted procedures and total costs associated to that target population. The yearly budget impacts and the budgets for the period between 2014 and 2018 were calculated for each scenario. No discounts rates or values regarding adjust for inflation were introduced, in compliance to international 12,13,16 and national f guidelines for this type of study.
The incremental budget impact for each examined year was calculated by means of the difference between the total budget impacts for the alternative and reference scenarios. The incremental difference among the alternative strategies was evaluated, which enabled the analysis of a wider and more restricted offer of technology.
Univariate and extreme scenarios sensitivity analyses were conducted to consider the uncertainties related to parameter values and premises used. 15 The evaluated parameters in the first ones were: the annual variation rate of lung cancer cases; costs of PET-CT procedure; prevalence of mediastinal and distant lesions; probability of conducting confirmatory mediastinoscopy; and CT and PET-CT sensitivity. The same ranges of values that were obtained in the literature and used in the study for the MH were used here. b The parameters were simultaneously modified in the extreme scenarios sensitivity analysis. The "best-case scenario" corresponded to minimizing the budget impact from PET incorporation for any alternative scenario adopted. The minimum values in the range that figures in Table 2 for the following parameter were employed: costs of PET-CT, annual variation rate for the number of new cases and CT sensitivity. Simultaneously, the following were employed considering their maximum values: biopsy sensitivity, share of patients having undergone mediastinoscopy procedure; and prevalence of metastases in mediastinal lymph nodes (N2/3) and distant metastases (M1). The "worst-case scenario" corresponded to the same parameters varying in the opposite direction to the one mentioned above.
Moreover, the influence from the rate by which the technology is disseminated at SUS was analyzed. It is possible that, even with it being included in SUS payrolls, delays may take place until it is fully offered, due to the current geographical availability of equipment and qualified staff for its operation. Sixty percent of patients were considered eligible for using PET-CT in 2014, with 10.0% increases with each year, until full access was achieved in 2018.
Written authorization was obtained from the Project (CNPq 564797/2010-3) coordinator, concerning the usage of data and model of the cost-effectiveness study.

RESULTS
The current diagnostic and therapeutic management model for NSCLC patients in Brazilian health care services, which is focused on CT use, would result in 397.5 BRL (246.9 USD) million in expenditures j in five years for SUS.  The introduction of PET-CT in NSCLC staging would imply an increase in total expenditures for SUS (Table 3) due to its complementary, non-replaceable nature, regardless of the strategy for its use. Its restricted use in patients with negative CT of thorax results would determine a total impact of 555.5 BRL (345.0 USD) million over the period (+39.8% as compared to the current management). Its use for all patients would cause an impact of 600.1 BRL (372.8 USD) million (+51.0%).
The financial impact from the more restricted PET-CT offer would imply an additional allocation of 158 BRL (98.2 USD) million in five years (Table 3). Extending the offer to all potential candidates would involve 202.7 BRL (125.9 USD) million in additional resources, with 44.6 BRL (27.7 USD) million being the difference between the strategies at the end of the period.
The cost of PET-CT procedure was the parameter with the biggest impact in the univariate sensitivity analyses ( Figure) using the values from the range in Table 2 The variation in the share of patients submitted to mediastinoscopy to confirm imaging exam results, between 0% and 100%, was shown to be important, given their costs to SUS. Non-performance of mediastinoscopy corresponded to a reduction in the total budget impact of 24.6 BRL (15.3 USD) million in the "PET-CT for CT-" scenario, and 20.3 BRL (12.6 USD) million in the use of "PET for all". Its conduction in all patients, on the other hand, would lead to increases in both scenarios of the same amounts mentioned above.
The use of the lower value of the range of the growth of staging-eligible NSCLC cases produced decreases in the budget impact regardless of the analyzed scenario: from 14 BRL (8.7 USD) million, in the "PET-CT for CT-" scenario, and 15.1 BRL (9.4 USD) million, with the offer of "PET-CT for all". Using the upper limit of that parameter resulted in increases of 3.7 BRL (2.3 USD) million and 4 BRL (2.5 USD) million, respectively.
The extreme scenarios sensitivity analyses (Table 4) showed significant reduction in total budget impact in the "best-case scenario": 90.

DISCUSSION
The incorporation of PET-CT in the staging of NSCLC, a highly relevant neoplasia in Brazil's nosological scenario, would imply total expenditures of 555.5 BRL (345.0 USD) million to SUS, in case its use is restricted to patients with   One of SUS's challenges lies in its compliance to the principle that health services should follow the principle of universality. Offering PET-CT to all candidates may not be feasible due to financial, infrastructural, or human resources limitations, among others. That acknowledgment, plus the fact that the literature and the study to the Ministry of Health point towards higher health benefits for the group with previous negative CT exams 6 In the alternative scenarios, the total budget impact estimated would correspond to 0.105% of the MH budget (restricted offer) or to 0.113% (inclusive offer). The dissemination of PET-CT into the clinical practice took place in a context in which concern with expenditures and impacts for health care systems was building up. Thus, the technology was the subject of several cost-effectiveness studies in several countries. Budget impact evaluations for its implementation are less frequent in the literature, and that is maybe so because they are conducted internally in the governmental environment which is involved with offering the technology. Nonetheless, directly comparing the results of those budget impact analyses with the ones herein is inappropriate. That is so because the management and organization of health care systems, structures of their models, epidemiological data, and especially the underlying cost structures greatly differ among studies. 17 Comparing budget impact estimates that are conducted in our reality would be ideal. Even though the MH has internally simulated the budget impact from PET at SUS, p its estimation methods and likelihood of bearing Although employing PET does not show a significant increase in the survival of patients, 4 its use allows for better (financial, material, and human) resource distribution in the system, as it more accurately identifies the extension of disease and allows planning the therapeutic strategy that is the most adequate to each case. Such smoother method would prevent unnecessary surgical procedures, which is more relevant when there are famous problems with access to health care services in the country, especially regarding oncology, and significant regional discrepancies in its offer. 10,q Budget impact studies are scarce, and only more recently they have gained guidelines on good practices more established. This study followed the main available guidelines on budget impact analyses from Task Force on Good Research Practices from the International Society for Pharmacoeconomics and Outcomes Research 13 and the ones from the Ministry of Health, which were recently published. f Required adaptations were made, as they mainly focus on therapeutic procedures.
Despite our using a nine-year time series (2006 to 2014) to estimate future lung cancer new cases, it was not possible to predict possible alterations arising from population changes or in the prevalence of some of its risk factors. Besides that, this study used parameter values from the cost-effectiveness study. Thus, the same limitations from before remain, as a gap in the national data regarding some epidemiological parameters, accuracy measurements for diagnostic technologies from international studies, and from the missing information about the share of patients who are submitted to mediastinoscopy within the country. The multiple sensitivity analyses conducted aimed at shedding some light on those uncertainties, and potentializing the knowledge regarding the extent of the impact they generate to SUS.
Trueman et al 20 discuss the incompatibility between the effort to maximize efficiency, which is the core target of economists, and the limits for the current budgets, which is commonly the main need from managers. Budget impact analyses do not show the best way to distribute available resources in the economy, whose most proper evidence come from comprehensive economic evaluation studies, such as the cost-effectiveness ones. Furthermore, the decisions to incorporate technologies in health care systems must take into account other factors, such as the availability of human and budget resources, political factors, and aspects regarding equal access to health care.
Data that is similar to the ones in this study, along with the evidence the technology is cost-effective in Brazil, may allow decisions taken to be properly backed up. Thus, the incorporation of PET in the clinical staging of potentially resectable NSCLC seems to be financially feasible considering the high total budget from Brazil's Ministry of Health and the potential reduction in the number of unnecessary surgeries better staged patients are submitted to. This may cause the available resources to be more efficiently distributed.