On 2017 May 03, Andrea Messori commented:

Please refer to the Letters' section of the Journal.

Presenting the results of a pharmacoeconomic study: incremental cost-effectiveness ratio vs net monetary benefit

Andrea Messori and Sabrina Trippoli

HTA Unit, ESTAR Regional Health Service, 50135 Firenze (Italy)

The article by Wouters and colleagues (1) presents an exhaustive overview on how QALYs can be used in cost-effectiveness analysis. In this framework, the authors also mention the incremental cost-effectiveness ratio (ICER), which is the parameter typically employed to express the results of a cost-effectiveness study. The article, however, does not discuss the net monetary benefit (NMB), which is another parameter employed to express the results of a cost-effectiveness study.

The incremental cost (deltaC) and the incremental effectiveness (deltaE) are the two main parameters of pharmacoeconomics and cost-effectiveness analysis, along with the willingness-to-pay threshold (lambda). The decision rule (e.g. in the case of a favourable pharmacoeconomic result) is (deltaC/deltaE)<lambda (Equation 1), if based on the ICER, or (deltaE x lambda - deltaC) > 0 (Equation 2), if based on the NMB. Likewise, an unfavourable pharmacoeconomic result is when (deltaC/deltaE)>lambda or when (deltaE x lambda - deltaC) < 0; NMB is defined as deltaE x lambda - deltaC, while ICER is defined as deltaC/deltaE. Despite its apparent complexity, most part of pharmacoeconomic methodology is described by the two simple equations reported above (i.e. Equations 1 and 2), but whether the ICER or the NMB is the best parameter for the purposes of pharmacoeconomic decision-making remains on open question.

The study by Cowper et al evaluating new versus old oral anticoagulants in patients with atrial fibrillation (2) is a typical ICER-based cost-effectiveness analysis in which the ICER of apixaban versus warfarin is compared against a willingness-to-pay threshold. This analysis can be taken as an example for comparing ICER vs NMB.

In one of the base-case analyses of the study by Cowper et al, QALYs per patient were 7.94 for apixaban and 7.54 for warfarin, while pharmacological costs per patient were $22,934 and $4,392, respectively. These data yielded, for apixaban versus warfarin, an ICER of $46,355 per QALY gained, a value that remains within the willingness-to-pay threshold of $50,000 per QALY gained and is therefore considered favourable (or “high value care”). As pointed our by Hlatky (3), in interpreting a specific ICER value, more than a single willingness-to-pay threshold is frequently considered (e.g. the threshold between $50,000 and $150,000 or the threshold above $150,000), and this allows us to better understand a pharmacoeconomic result expressed on the basis of an ICER.

In the methodology of pharmacoeconomics, the net monetary benefit (NMB) plays a role similar to that of ICER, but some differences are important.

Firstly, the ICER –by definition- has always an incremental nature and consequently the absolute cost-effectiveness ratio (calculated for a single treatment in the absence of any comparison) makes little sense and, for this reason, is rarely employed. In contrast, the NMB can be calculated for a single treatment in the absence of any comparison (absolute NMB) or can conversely be calculated as an incremental parameter [according to the equation: (incremental NMB) = (incremental QALYs per patient) x (willingness-to-pay threshold) – (incremental cost per patient)]. Another feature of NMB is that the incremental NMB for the comparison of A vs B can be estimated as the absolute NMB calculated for A minus the absolute NMB calculated for B. In this sequence of calculations, calculating the absolute NMB makes sense because the absolute NMB (separately calculated for the experimental treatment and for the control treatment) represents an intermediate step in the calculation of the incremental NMB (Table 1)

The values of absolute NMB for apixaban and warfarin (Table 1) are, respectively, 374,066 and $372,608 per patient (calculated according to Equation 2). Hence, the incremental NMB for apixaban vs warfarin is simply the difference of the above two values, i.e. $1,458 per patient.

References

[1] Wouters OJ, Naci H, Samani NJ. QALYs in cost-effectiveness analysis: an overview for cardiologists. Heart. 2015 Dec;101(23):1868-73.

[2] Cowper PA, Sheng S, Lopes RD, Anstrom KJ, Stafford JA, Davidson-Ray L, Al-Khatib SM, Ansell J, Dorian P, Husted S, McMurray JJ, Steg PG, Alexander JH, Wallentin L, Granger CB, Mark DB. Economic Analysis of Apixaban Therapy for Patients With Atrial Fibrillation From a US Perspective: Results From the ARISTOTLE Randomized Clinical Trial. JAMA Cardiol. 2017 Mar 29. doi:10.1001/jamacardio.2017.0065. [Epub ahead of print]

[3] Hlatky MA. Are Novel Anticoagulants Worth Their Cost? JAMA Cardiol. 2017 Mar 29. doi: 10.1001/jamacardio.2017.0126. [Epub ahead of print]

Table 1. Cost-effectiveness of apixaban vs warfarin in atrial fibrillation: basecase analysis reported by Cowper et al.(2)

a) STARTING VALUES

apixaban: QALYs per patient = 7.94, cost per patient = $22,934

warfarin: QALYs per patient = 7.54, cost per patient = $4,392

willingness-to-pay threshold = $50,000/QALY

b) PHARMACOECONOMIC PARAMETERS

ICER = ($22,934 - $ 4,392) / (7.94 – 7.54) = 18,542 / 0.40 = $46,355/QALY

absolute NMB for apixaban = 7.94 x $50,000 – $22,934 = $374,066

absolute NMB for warfarin = 7.54 x $50,000 – $4,392 = $372,608

incremental NMB = absolute NMB for apixaban - absolute NMB for warfarin = $374,066 - $372,608 = $1,458

c) INTERPRETATION

ICER (equal to $46,355/QALY): favourable because <$50,000/QALY

incremental NMB (equal to $1,458 per patient): favourable because >0

On 2017 May 03, Andrea Messori commented:

Please refer to the Letters' section of the Journal.

Presenting the results of a pharmacoeconomic study: incremental cost-effectiveness ratio vs net monetary benefit

Andrea Messori and Sabrina Trippoli

HTA Unit, ESTAR Regional Health Service, 50135 Firenze (Italy)

The article by Wouters and colleagues (1) presents an exhaustive overview on how QALYs can be used in cost-effectiveness analysis. In this framework, the authors also mention the incremental cost-effectiveness ratio (ICER), which is the parameter typically employed to express the results of a cost-effectiveness study. The article, however, does not discuss the net monetary benefit (NMB), which is another parameter employed to express the results of a cost-effectiveness study.

The incremental cost (deltaC) and the incremental effectiveness (deltaE) are the two main parameters of pharmacoeconomics and cost-effectiveness analysis, along with the willingness-to-pay threshold (lambda). The decision rule (e.g. in the case of a favourable pharmacoeconomic result) is (deltaC/deltaE)<lambda (Equation 1), if based on the ICER, or (deltaE x lambda - deltaC) > 0 (Equation 2), if based on the NMB. Likewise, an unfavourable pharmacoeconomic result is when (deltaC/deltaE)>lambda or when (deltaE x lambda - deltaC) < 0; NMB is defined as deltaE x lambda - deltaC, while ICER is defined as deltaC/deltaE. Despite its apparent complexity, most part of pharmacoeconomic methodology is described by the two simple equations reported above (i.e. Equations 1 and 2), but whether the ICER or the NMB is the best parameter for the purposes of pharmacoeconomic decision-making remains on open question.

The study by Cowper et al evaluating new versus old oral anticoagulants in patients with atrial fibrillation (2) is a typical ICER-based cost-effectiveness analysis in which the ICER of apixaban versus warfarin is compared against a willingness-to-pay threshold. This analysis can be taken as an example for comparing ICER vs NMB.

In one of the base-case analyses of the study by Cowper et al, QALYs per patient were 7.94 for apixaban and 7.54 for warfarin, while pharmacological costs per patient were $22,934 and $4,392, respectively. These data yielded, for apixaban versus warfarin, an ICER of $46,355 per QALY gained, a value that remains within the willingness-to-pay threshold of $50,000 per QALY gained and is therefore considered favourable (or “high value care”). As pointed our by Hlatky (3), in interpreting a specific ICER value, more than a single willingness-to-pay threshold is frequently considered (e.g. the threshold between $50,000 and $150,000 or the threshold above $150,000), and this allows us to better understand a pharmacoeconomic result expressed on the basis of an ICER.

In the methodology of pharmacoeconomics, the net monetary benefit (NMB) plays a role similar to that of ICER, but some differences are important.

Firstly, the ICER –by definition- has always an incremental nature and consequently the absolute cost-effectiveness ratio (calculated for a single treatment in the absence of any comparison) makes little sense and, for this reason, is rarely employed. In contrast, the NMB can be calculated for a single treatment in the absence of any comparison (absolute NMB) or can conversely be calculated as an incremental parameter [according to the equation: (incremental NMB) = (incremental QALYs per patient) x (willingness-to-pay threshold) – (incremental cost per patient)]. Another feature of NMB is that the incremental NMB for the comparison of A vs B can be estimated as the absolute NMB calculated for A minus the absolute NMB calculated for B. In this sequence of calculations, calculating the absolute NMB makes sense because the absolute NMB (separately calculated for the experimental treatment and for the control treatment) represents an intermediate step in the calculation of the incremental NMB (Table 1)

The values of absolute NMB for apixaban and warfarin (Table 1) are, respectively, 374,066 and $372,608 per patient (calculated according to Equation 2). Hence, the incremental NMB for apixaban vs warfarin is simply the difference of the above two values, i.e. $1,458 per patient.

References

[1] Wouters OJ, Naci H, Samani NJ. QALYs in cost-effectiveness analysis: an overview for cardiologists. Heart. 2015 Dec;101(23):1868-73.

[2] Cowper PA, Sheng S, Lopes RD, Anstrom KJ, Stafford JA, Davidson-Ray L, Al-Khatib SM, Ansell J, Dorian P, Husted S, McMurray JJ, Steg PG, Alexander JH, Wallentin L, Granger CB, Mark DB. Economic Analysis of Apixaban Therapy for Patients With Atrial Fibrillation From a US Perspective: Results From the ARISTOTLE Randomized Clinical Trial. JAMA Cardiol. 2017 Mar 29. doi:10.1001/jamacardio.2017.0065. [Epub ahead of print]

[3] Hlatky MA. Are Novel Anticoagulants Worth Their Cost? JAMA Cardiol. 2017 Mar 29. doi: 10.1001/jamacardio.2017.0126. [Epub ahead of print]

Table 1. Cost-effectiveness of apixaban vs warfarin in atrial fibrillation: basecase analysis reported by Cowper et al.(2)

a) STARTING VALUES

apixaban: QALYs per patient = 7.94, cost per patient = $22,934

warfarin: QALYs per patient = 7.54, cost per patient = $4,392

willingness-to-pay threshold = $50,000/QALY

b) PHARMACOECONOMIC PARAMETERS

ICER = ($22,934 - $ 4,392) / (7.94 – 7.54) = 18,542 / 0.40 = $46,355/QALY

absolute NMB for apixaban = 7.94 x $50,000 – $22,934 = $374,066

absolute NMB for warfarin = 7.54 x $50,000 – $4,392 = $372,608

incremental NMB = absolute NMB for apixaban - absolute NMB for warfarin = $374,066 - $372,608 = $1,458

c) INTERPRETATION

ICER (equal to $46,355/QALY): favourable because <$50,000/QALY

incremental NMB (equal to $1,458 per patient): favourable because >0