Skip to main content

Table 2 Details regarding cost-effectiveness studies of non-small cell lung cancer

From: A systematic review of the cost and cost-effectiveness studies of immune checkpoint inhibitors

Reference, Country, Year

Comparison

Methodologyb

Costs

QALYs

ICER

WTP

Conclusions

Criticisms

Goeree et al., Canada, 2016 [28]

Nivo vs. doc vs. erl for recurrent stage IIIB/IV SCC

PSa + Markov; accounted for toxicity, administration, and end-of-life costs

CAD 139,016 ($107,631) nivo, CAD 38,812 ($30,049) doc, CAD 39,920 ($30,906) erl

1.23 nivo, 0.58 doc, 0.54 erl

Relative to doc, nivo CAD 152,229 ($117,857)/QALY and relative to erl, 141,838 ($109,811)/QALY

No specific amount in Canada

Compared with doc or erl, nivo may or may not be CE depending on WTP threshold

- Overall modeling horizon of 10 years, when low numbers of patients still alive, causing errors in survival extrapolation and thus costs

- Difficulty in interpreting next-line therapies

- Only grade ≥ 3 toxicities accounted for, as a one-time cost

Matter-Walstra et al., Switzerland, 2016 [29]

PD-L1 testing + subsequent decision vs. nivo vs. doc for recurrent non-SCC

Markov; PD-L1 testing cutoffs ≥1% and ≥ 10%; accounted for end-of-life costs and reducing nivo dose and duration

CHF 37,618 ($39,378) doc, CHF 66,208 ($69,306) nivo, CHF 47,410 ($49,628) nivo with dose reduction, CHF 55,394 ($57,992) with duration reduction

0.53 doc, 0.69 nivo

Relative to doc, nivo CHF 177,478 ($185,802)/QALY, reduced dose CHF 60,787 ($63,638)/QALY, reduced duration CHF 110,349 ($115,524)/QALY

CHF 100,000 ($104,690) /QALY

Although not at baseline, nivo is CE by dose reduction and increased PD-L1 threshold

- Overall modeling horizon of complete lifetime, causing errors in survival extrapolation and thus costs

- In addition to lack of a discount rate, did not account for most toxicities, administration, or death costs

- Median PFS favored doc, but 1-year PFS favored nivo; difficult to account for in model

Aguiar et al., USA, 2017 [30]

Nivo, pembro, or atezo ± prior PD-L1 testing vs. doc for recurrent SCC/non-SCC

Decision-analytic model; PD-L1 testing cutoffs ≥1%/≥5%/≥10% for nivo and ≥ 1%/50% for pembro; accounted for administration, monitoring, end-of-life costs

$140,453 nivo for SCC and $100,791 nivo for non-SCC (PD-L1 untested), $82,201 pembro (PD-L1 ≥ 1%), $122,155 atezo (PD-L1 untested); doc $39,516–$48,182

0.82 nivo (SCC), 0.87 nivo (non-SCC), 0.92 pembro, 0.90 atezo, 0.54–0.59 doc

Relative to doc, nivo $155,605/QALY (SCC) and nivo $187,685/QALY (non-SCC), pembro $98,421/QALY, atezo $215,802/QALY

$100,000/QALY

Atezo not CE; pembro is CE; although not at baseline, nivo is CE by increased PD-L1 threshold

- Toxicity types and one-time costs thereof not explained

- Unclear whether patients treated until PD, along with details of next-line therapy (if present)

- Analysis and reflections based on USA Medicare system, limiting broader applicability

Huang et al., USA, 2017 [31]

Pembro vs. doc for recurrent non-SCC

PS; PD-L1 testing cutoff ≥50%; accounted for toxicity, administration, end-of-life costs

$136,921 doc, $297,443 pembro

0.76 doc, 1.71 pembro

Relative to doc, pembro $168,619/QALY

Per capita GDP × 3

Pembro CE at the particular WTP threshold, which is nonstandard and thus questionable

- Overall modeling horizon of 20 years, when exceedingly low numbers of patients still alive, causing errors in survival extrapolation, which was done through retrospective population datasets

- Survival found to most influence costs; thus, extrapolation may markedly influence overall costs

- Only grade ≥ 3 toxicities accounted for, as a one-time cost

Huang et al., USA, 2017 [31]

Pembro vs. several types of chemo for first-line SCC/non-SCC

PS; PD-L1 testing cutoff ≥50%; accounted for toxicity, administration, end-of-life costs

$260,223 chemo, $362,662 pembro

1.55 chemo, 2.60 pembro

Relative to chemo, pembro $97,621/QALY

No specific amount used

Pembro is CE in this setting

- Overall modeling horizon of 20 years, when exceedingly low numbers of patients still alive, causing errors in survival extrapolation (done through retrospective population datasets), which may impact costs secondarily

- Heterogeneity in treating with variety of chemo

- Only grade ≥ 3 toxicities accounted for, as a one-time cost

  1. QALY quality-adjusted life year, ICER incremental cost-effectiveness ratio, WTP willingness to pay (threshold); nivo, nivolumab; doc, docetaxel; erl, erlotinib; SCC squamous cell carcinoma, PS partitioned survival, CAD Canadian dollar, CE cost-effective, PFS progression-free survival, OS overall survival, PD-L1 programmed cell death ligand-1, CHF Swiss francs
  2. aData for partitioned survival model not shown owing to virtual similarity as Markov data
  3. bAll studies consisted of three basic health states (progression-free (stable), progressive disease, and death); all studies performed sensitivity analyses in addition to the base case