The impact of increased post-progression survival on the cost-effectiveness of interventions in oncology

Figures & data

Figure 1 Model schematic.

Figure 2 Progression-free survival and overall survival curves.

Table 1 Details of TAs and marginal cost-effectiveness ratios

TA	Broad indication	Drug name	Annual PPS background costs	PPS utility	Marginal cost-effectiveness ratio	ICER	NICE Decision
458	Breast (stage III or IV)	Trastuzumab emtansine	£2,072	0.53	£3,910	£98,525 to £131,473	R
423	Breast (stage III or IV)	Eribulin	£2,799	0.679	£4,122	£35,624 to £36,244	CR
263	Breast (stage IV)	Bevacizumab (with capecitabine)	£9,648	0.496	£19,452	£77,318	NR
463	RCC (advanced)	Cabozantinib	£2,017	0.777	£2,596	–	R
417	RCC (advanced)	Nivolumab	£3,676	0.663*	£5,549	£42,417 to £83,829	R
333	RCC (advanced)	Axitinib	£4,147	0.61	£6,798	–	R
405	Bowel (metastatic)	Trifluridine (with tipiracil hydrochloride)	£2,319	0.64	£3,623	£44,032	R
377	Prostate (metastatic)	Enzalutamide	£37,024*	0.612*	£60,497	£27,076 to £95,685	R
428	Lung (stage III or IV)	Pembrolizumab	£1,853	0.763	£2,429	£23,424 to £49,048	CR
447	Lung (stage IV)	Pembrolizumab	£6,545	0.668	£9,798	£44,896	CR
416	Lung (stage III or IV)	Osimertinib	£7,255	0.678	£10,701	–	CR
406	Lung (stage III)	Crizotinib	£2,327*	0.565*	£4,119	–	R
411	Lung (stage III or IV)	Necitumumab	£4,016	0.55	£7,302	£60,133 to £119,912	NR
403	Lung (stage III or IV)	Ramucirumab (with docetaxel)	£11,007	0.599	£18,375	£84,985 to £1,106,497	NR
395	Lung (stage III)	Ceritinib	£3,764	0.46	£8,183	£62,456	R
374	Lung (stage III)	Erlotinib and Gefitinib	£9,949	0.4734	£21,016	£15,359 to £61,132	CR Erlotinib; Gefitinib NR
347	Lung (recurrent or metastatic)	Nintedanib	£4,380*	0.638	£6,866	£27,008 to £50,776	R
310	Lung (recurrent or metastatic)	Afatinib	£18,744*	0.52*	£36,046	–	CR
421	Breast (stage III)	Everolimus	–	–	Incalculable	£61,046	R
432	RCC (metastatic)	Everolimus	–	–	Incalculable	£52,261 to £58,316	R
378	Bowel (advanced)	Ramucirumab (with paclitaxel)	–	0.587	Incalculable	£53,830 to £188,640	NR
307	Bowel (metastatic)	Aflibercept (with irinotecan and fluorouracil)	–	–	Incalculable	£30,474 to £36,294	NR
473	Head & neck (recurrent or metastatic)	Cetuximab (with platinum-based therapy)	–	0.52	Incalculable	–	CR
422	Lung (stage III)	Crizotinib	–	0.61	Incalculable	–	R
402	Lung (stage III or IV)	Pemetrexed	–	–	Incalculable	£70,538	CR
412	Prostate (metastatic)	Radium-223 dichloride	–	0.47 and 0.56	Incalculable	£25,963	CR
391	Prostate (metastatic)	Cabazitaxel (with prednisone)	–	0.6266	Incalculable	£49,327	CR
259	Prostate (metastatic)	Abiraterone acetate (with prednisolone)	–	0.5	Incalculable	£52,851 to £170,550	CR
316	Prostate (metastatic)	Enzalutamide	£1,941	–	Incalculable	£14,795 to £102,751	R

Notes: *indicates instances where multiple inputs were used in models. The marginal cost-effectiveness ratio was calculated by averaging all relevant utility inputs and applying only ongoing background PPS costs (not one-off costs).

Abbreviations: TA, technology appraisal; PPS, post-progression survival; ICER, incremental cost-effectiveness ratio; R, recommended; CR, conditional recommendation; NR, not recommended.

Table 2 Replicated TA403 model results with corrected discount rate

	Treatment	Comparator	Incremental
Costs
Tx cost - pre-progression	£24,412	£1,134	£23,278
Background: pre-progression	£1,689	£1,195	£494
Background: post-progression	£11,059	£9,706	£1,353
Cost of death	£0	£0	£0
AEs	£807	£656	£152
Total cost	£37,968	£12,691	£25,276
QALYs
Pre-progression QALYs	0.423	0.299	0.124
Post-progression QALYs	0.602	0.528	0.074
AEs	−0.003	−0.003	0.000
Total QALYs	1.022	0.825	0.197
ICER			£128,233

Abbreviations: Tx, treatment; AE, adverse event; QALY, quality-adjusted life year; ICER, incremental cost-effectiveness ratio.

Table 3 Results of Scenario 1 (separation of post-progression background costs)

Scenario	Cost timings	PPS monthly cost	Total treatment costs	Total comparator costs	Incremental costs	ICER	Marginal cost-effectiveness ratio
Base case	Original	£635	£37,968	£12,691	£25,276	£128,233	£12,721
	80% at death	£162	£37,417	£13,267	£24,150	£122,521	£3,245
	80% at progression	£162	£37,851	£13,632	£24,219	£122,868
	40% at death and 40% at progression	£162	£37,634	£13,450	£24,185	£122,695
Disease Scenario 1: Shorter OS	Original	£635	£29,059	£4,006	£25,053	£144,878	£12,721
	80% at death	£160	£28,044	£3,810	£24,234	£140,143	£3,205
	80% at progression	£160	£28,700	£4,548	£24,152	£139,668
	40% at death and 40% at progression	£160	£28,371	£4,179	£24,193	£139,906
Disease Scenario 2: Longer OS	Original	£635	£39,050	£13,900	£25,150	£133,296	£12,721
	80% at death	£157	£38,392	£14,272	£24,120	£127,839	£3,145
	80% at progression	£157	£38,674	£14,498	£24,176	£128,137
	40% at death and 40% at progression	£157	£38,533	£14,385	£24,148	£127,988

Abbreviations: OS, overall survival; PPS, post-progression survival; ICER, incremental cost-effectiveness ratio.

Table 4 Results of Scenario 2 (adopting end-of-life criteria using arbitrary QALY weightings)

	QALY weighting
	1.00	1.10	1.30	1.50	1.70	1.90	2.10	2.30	2.50
Marginal cost-effectiveness ratio for all disease scenarios	£12,721	£11,565	£9,786	£8,481	£7,483	£6,695	£6,058	£5,531	£5,088

Notes: As survival gain in pre- and post-progression was equal across all disease scenarios and monthly costs were the same, the impact of QALY weighting on the marginal cost-effectiveness ratio was the same across all scenarios.

Abbreviations: QALY, quality-adjusted life year; ICER, incremental cost-effectiveness ratio.

Table 5 Results of Scenario 3 (adopting end-of-life criteria using population norm QALY weightings)

Scenario	Change	PPS utility	“Population norm” utility	Mean survival gain	Additional discounted QALYs	Marginal cost-effectiveness ratio	ICER
Base case	Original	0.599	–	4.4 months	–	£12,721	£128,233
	Adjusted	0.599	0.800	4.4 months	0.067	£9,525	£95,848
Disease Scenario 1	Original	0.599	–	3.74 months	–	£12,721	£144,878
	Adjusted	0.599	0.800	3.74 months	0.061	£9,525	£106,978
Disease Scenario 2	Original	0.599	–	4.03 months	–	£12,721	£133,178
	Adjusted	0.599	0.800	4.03 months	0.062	£9,525	£100,190

Abbreviations: PPS, post-progression survival; QALY, quality-adjusted life year; ICER, incremental cost-effectiveness ratio.

Table 6 Marginal cost-effectiveness ratios by QALY weighting with applied transition costs

	Base case	Applied transition costs (ie, reduced monthly cost of survival) and QALY weighting
		1.00	1.10	1.30	1.50	1.70	1.90	2.10	2.30	2.50
Base case	£12,721	£3,245	£2,950	£2,496	£2,164	£1,909	£1,708	£1,545	£1,411	£1,298
Disease Scenario 1	£12,721	£3,205	£2,914	£2,466	£2,137	£1,885	£1,687	£1,526	£1,394	£1,282
Disease Scenario 2	£12,721	£3,145	£2,859	£2,419	£2,097	£1,850	£1,655	£1,498	£1,367	£1,258

Abbreviation: QALY, quality-adjusted life year.

Table S1 Table of exclusions with reasons from the TA review

TA	Broad indication	Drug name	Reason for exclusion
424	Breast	Pertuzumab	Not a standard partitioned survival model.
107	Breast	Trastuzumab	Not a standard partitioned survival model. Submission outside of eligible date range (2006).
109	Breast	Docetaxel	Submission outside of eligible date range (2006).
439	Bowel	Cetuximab and Panitumumab	Not a standard partitioned survival model.
326	Bowel	Imatinib	Not a standard partitioned survival model.
118	Bowel	Bevacizumab and cetuximab	Guidance partially replaced by TA242. Original submission outside of eligible date range (2006).
242	Bowel	Cetuximab, bevacizumab and panitumumab	Submission outside of eligible date range (January 2012).
272	Bladder	Vinflunine	Submission outside of eligible date range (2010).
404	Prostate	Degarelix	Not a standard partitioned survival model.
387	Prostate	Abiterone	Not a standard partitioned survival model.
190	Lung	Pemetrexed	Submission outside of eligible date range (2010).

Abbreviation: TA, technology appraisal.

Table S2 Model inputs used for matched replication

Model input	Value used in replication	Source and comments
Pre-progressed utility	0.706	Reported in Table 56 of the company’s submission, page 194 of committee papers.
Post-progressed utility	0.599	Reported in Table 56 of the company’s submission, page 194 of committee papers.
Pre-progressed background costs	£163 per 3 week cycle	One oncologist visit every 3 week cycle at £136.08; plus one CT scan at £94.90, 0.28 times every 3 week cycle (£136.08 + £26.57 = £162.65). Reported in Table 63 of the company’s submission, page 201 of committee papers.
Post-progressed background costs	£635 per 3 week cycle	One oncologist visit every 3 week cycle at £136.08; plus one CT scan at £94.90, 0.28 times every 3 week cycle; plus one urinalysis at £1.22 every 3 week cycle; plus one blood count at £3.09 every 3 week cycle; plus one renal function test at £12.18 every 3 week cycle; plus one hepatic function test at £8.52 every 3 week cycle; plus one electrolyte test at £4.87 every 3 week cycle; plus anti-cancer treatment for 30% patients at £473.78 every 3 week cycle; plus best supportive care for 70% patients at £429.16 every 3 week cycle (£136.08 + £26.57 +£1.22 + 3.09 + £12.18 + £8.52 + £4.87 + £142.13 + £300.41 = £635.08). Reported in Table 63 of the company’s submission, page 201 of committee papers. The ERG report was used to clarify the proportion of patients on active anti-cancer treatment vs best supportive care, as the submission values did not sum to 100%.
Docetaxel acquisition costs (pre-progression)	£62 per 3 week cycle	Docetaxel at £36.00 every 3 week cycle; plus pre-medication at £32.11 for 98.1% patients every 3 week cycle (£36.00 + £32.11 = £61.50). Reported in Tables 7, 57 and 74 of the company’s submission, pages 72, 196 and 207 of the committee papers.
Docetaxel administration costs (pre-progression)	£165 per 3 week cycle	Reported in Table 62 of the company’s submission, page 200 of committee papers.
Rambucirumab plus docetaxel acquisition costs (pre-progression)	£3,804 per 3 week cycle	Rambucirumab at £3,733 every 3 week cycle; plus docetaxel at £36 every 3 week cycle; plus pre-medication at £35.54 for 97.6% patients every 3 week cycle (£3,733 + £35 + £36 = £3,804) Reported in Tables 7, 57 and 74 of the company’s submission, pages 72, 196 and 207 of the committee papers.
Rambucirumab plus docetaxel administration costs (pre-progression)	£219 per 3 week cycle	Reported in Table 62 of the company’s submission, page 200 of committee papers.
Mean cycles of docetaxel treatment	4.9	Reported in Table 71 of the company’s submission, page 206 of committee papers.
Mean cycles of rambucirumab plus docetaxel treatment	6.1	Reported in Table 71 of the company’s submission, page 206 of committee papers.
Docetaxel adverse event costs	£656	A weighted average assuming: neutropenia affected 39.8% at £356; febrile neutropenia affected 10.0% patients at £2,070; fatigue affected 10.5% patients at £381; nausea/vomiting affected 1.9% patients at £1,975; diarrhea affected 3.1% patients at £1,848; rash affected 0.7% patients at £657; dyspnea affected 8.3% patients at £571; leukopenia affected 12.5% patients at £435; anemia affected 5.7% patients at £1,006 and hypertension affected 2.1% patients at £421. Reported in Tables 64 and 68 of the company’s submission, pages 203 and 205 of the committee papers.
Rambucirumab plus docetaxel adverse event costs	£807	A weighted average assuming: neutropenia affected 48.8% at £356; febrile neutropenia affected 16.0% patients at £2,070; fatigue affected 14.0% patients at £381; nausea/vomiting affected 1.3% patients at £1,975; diarrhea affected 4.6% patients at £1,848; rash affected 0.8% patients at £657; dyspnea affected 3.8% patients at £571; leukopenia affected 13.7% patients at £435; anemia 2.9% patients at £1,006 and hypertension affected 5.6% patients at £421. Reported in Tables 64 and 68 of the company’s submission, pages 203 and 205 of the committee papers.
Docetaxel adverse event (dis)utility	−0.0029	A weighted average assuming: neutropenia affected 39.8% at −0.090 per day for 7 days; febrile neutropenia affected 10.0% patients at −0.090 per day for 4 days; fatigue affected 10.5% patients at −0.073 per day for 21 days; nausea/vomiting affected 1.9% patients at −0.048 per day for 3 days; diarrhea affected 3.1% patients at −0.047 per day for 3 days; hair loss affected 25.2% patients at −0.045 per day for 21 days; rash affected 0.7% patients at −0.032 per day for 21 days; dyspnea affected 8.3% patients at −0.073 per day for 21 days; leukopenia affected 12.5% patients at −0.090 per day for 7 days; anemia affected 5.7% patients at −0.090 per day for 21 days and hypertension affected 2.1% patients at −0.073 per day for 21 days. Reported in Tables 56 and 68 of the company’s submission, pages 194 and 205 of the committee papers.
Rambucirumab plus docetaxel adverse event (dis)utility	−0.0031	A weighted average assuming: neutropenia affected 48.8% at −0.090 per day for 7 days; febrile neutropenia affected 16.0% patients at −0.090 per day for 4 days; fatigue affected 14.0% patients at −0.073 per day for 21 days; nausea/vomiting affected 1.3% patients at −0.048 per day for 3 days; diarrhea affected 4.6% patients at −0.047 per day for 3 days; hair loss affected 25.8% patients at −0.045 per day for 21 days; rash affected 0.8% patients at −0.032 per day for 21 days; dyspnea affected 3.8% patients at −0.073 per day for 21 days; leukopenia affected 13.7% patients at −0.090 per day for 7 days; anemia 2.9% patients at −0.090 per day for 21 days and hypertension affected 5.6% patients at −0.073 per day for 21 days. Reported in Tables 56 and 68 of the company’s submission, pages 194 and 205 of the committee papers.
Discount rate	10.9%	Please note that the reported discount rate in Table 42 of the company’s submission, page 168 of the Committee Papers was 3.5%. However the ERG established that due to a calculation error, the effective error rate in the company submission was 10.9%.
PFS hazard ratio	0.762	Reported in Figure 28 of the company’s submission, page 178 of the committee papers.
OS hazard ratio	0.857	Reported in Figure 24 of the company’s submission, page 172 of the committee papers.
PFS curve parameters	Lognormal distribution with coefficients of 1.18 and 0.96	No coefficients were provided in the company’s submission, thus coefficients for alternative distributions were produced by estimating individual patient level data from digitised versions of the company Kaplan-Meier plots (Figures 24 and 28). The distributions that produced curves and extrapolations that visually best matched the curves reported in the company’s submission were then selected. OS was extrapolated using a loglogistic distribution in line with the company’s submission (page 177 of the committee papers). However, the lognormal distribution generated using our estimated coefficients was a better visual fit for the PFS curve than the generalised gamma that was used in the company’s submission (page 183 of the committee papers). It should be noted that the company’s submission found that the assumption of proportional hazards did not hold for the PFS curves, and thus modelled each treatment arm separately. For simplicity, our replication assumed proportional hazards, but we recognise that this does not fully align with the company’s submission.
OS curve parameters	Loglogistic distribution with coefficients of 8,85 and 1.56

Abbreviations: ERG, evidence review group; OS, overall survival; PFS, progression-free survival.

Table S3 Model results for Replication 1, where inputs matched to the company’s submission

	Treatment		Comparator		Incremental
	Company’s	Replication 1	Company’s	Replication 1	Company’s	Replication 1
PFS treatment cost	£24,037	£24,166	£1,118	£1,125	£22,919	£23,040
PFS background costs	£1,618	£1,601	£1,242	£1,157	£376	£444
PPS background costs	£8,821	£8,892	£7,979	£8,175	£842	£717
AE costs	£807	£807	£656	£656	£151	£152
Total Cost	£35,283	£35,466	£10,995	£11,113	£24,288	£24,353
PFS QALYs	0.341	0.401	0.262	0.290	0.079	0.111
PPS QALYs	0.478	0.484	0.433	0.445	0.045	0.039
AE QALYs	−0.003	−0.003	−0.003	−0.003	0.000	0.000
Total QALYs	0.816	0.882	0.692	0.732	0.124	0.150
ICER					£194,919	£162,430

Abbreviations: AE, adverse event; ERG, evidence review group; ICER, incremental cost-effectiveness ratio; PFS, progression-free survival; PPS, post progression survival.

Table S4 Artificially deflated utilities applied to second replication

Health state	Utility values
	Reported in submission	Deflated: treatment	Deflated: comparator
Pre progressed	0.706	0.6	0.64
Post progressed	0.599	0.592	0.583

Abbreviations: AE, adverse event; ICER, incremental cost-effectiveness ratio; PFS, progression-free survival; PPS, post progression survival.

Table S5 Model results for Replication 2, where utilities were artificially deflated

	Treatment		Comparator		Incremental
	Company’s	Replication 2	Company’s	Replication 2	Company’s	Replication 2
PFS treatment cost	£24,037	£24,166	£1,118	£1,125	£22,919	£23,040
PFS background costs	£1,618	£1,601	£1,242	£1,157	£376	£444
PPS background costs	£8,821	£8,892	£7,979	£8,175	£842	£717
AE costs	£807	£807	£656	£656	£151	£152
Total Cost	£35,283	£35,466	£10,995	£11,113	£24,288	£24,353
PFS QALYs	0.341	0.341	0.262	0.263	0.079	0.078
PPS QALYs	0.478	0.478	0.433	0.433	0.045	0.045
AE QALYs	−0.003	−0.003	−0.003	−0.003	0.000	0.000
Total QALYs	0.816	0.816	0.692	0.693	0.124	0.123
ICER					£194,919	£198,299