How cost-effective are new cancer drugs in the U.S.?

Figures & data

Figure 1. Number of drugs used to treat cancer ever approved by the FDA.

Figure 2. Drugs approved for Acute Lymphoblastic Leukemia.

Figure 3. Number of drugs ever approved for 4 types of cancer, 1994–2014.

Table 1. Number of drugs ever approved by the FDA, by cancer site and year, 1990–2014.

	All drugs							Priority-review drugs
	1990	1994	1998	2002	2006	2010	2014	1990	1994	1998	2002	2006	2010	2014
C00-C14 Head and Neck Cancer	3	3	4	4	5	5	5	2	2	3	3	3	3	3
C15 Esophageal Cancer	0	0	2	2	2	2	3	0	0	1	1	1	1	1
C16 Stomach (Gastric) Cancer	3	3	5	5	5	5	6	3	3	4	4	4	4	4
C18 Colon Cancer	2	2	4	5	8	8	11	2	2	4	5	5	5	6
C20 Rectal Cancer	2	2	4	5	8	8	11	2	2	4	5	5	5	6
C22 Liver Cancer	0	0	0	0	1	1	1	0	0	0	0	1	1	1
C25 Pancreatic Cancer	2	3	5	5	7	8	8	2	3	5	5	7	8	8
C34 Lung Cancer	5	7	11	11	15	15	21	5	7	10	10	13	13	16
C40-C41 Bone Cancer	3	3	3	3	3	4	4	3	3	3	3	3	3	3
C43 Melanoma	2	3	3	4	4	4	10	0	0	0	0	0	0	1
C44 Basal Cell Carcinoma	1	1	2	2	2	2	3	1	1	1	1	1	1	2
C45 Malignant Mesothelioma	0	0	0	0	1	1	1	0	0	0	0	1	1	1
C46 Kaposi Sarcoma	3	4	4	4	4	4	4	2	3	3	3	3	3	3
C49 Soft Tissue Sarcoma	3	3	3	4	4	5	5	3	3	3	4	4	4	4
C50 Breast Cancer	7	9	17	20	20	24	26	4	6	11	12	12	16	16
C51 Vulvar Cancer	1	1	1	1	1	1	1	0	0	0	0	0	0	0
C53 Cervical Cancer	1	1	2	2	3	3	3	0	0	1	1	1	1	1
C54.1 Endometrial Cancer	1	1	1	1	1	1	1	0	0	0	0	0	0	0
C56 Ovarian, Fallopian Tube, or Primary Peritoneal Cancer	5	6	8	8	9	9	10	4	5	7	7	7	7	8
C60 Penile Cancer	1	1	1	1	1	1	1	0	0	0	0	0	0	0
C61 Prostate Cancer	5	5	7	7	7	9	12	4	4	6	6	6	7	10
C62 Testicular Cancer	6	6	7	7	7	7	7	4	4	4	4	4	4	4
C64-C65 Kidney (Renal Cell) Cancer	3	4	4	4	7	10	12	3	3	3	3	5	7	7
C67 Bladder Cancer	3	3	3	3	3	3	3	2	2	2	2	2	2	2
C69.2 Retinoblastoma	1	1	1	1	1	1	1	1	1	1	1	1	1	1
C71 Brain Tumors	2	2	2	3	4	5	5	1	1	1	2	2	3	3
C73 Thyroid Cancer	1	1	1	1	2	2	4	1	1	1	1	2	2	4
C74.9 Neuroblastoma	3	3	3	3	3	3	3	3	3	3	3	3	3	3
C81 Hodgkin Lymphoma	12	12	12	12	12	12	13	9	9	9	9	9	9	9
C82-C85 Non-Hodgkin Lymphoma	13	13	14	16	21	25	29	11	11	11	11	15	18	21
C90 Multiple Myeloma and Other Plasma Cell Neoplasms	4	5	6	7	9	10	12	4	5	6	7	9	10	10
C91.0 Acute Lymphoblastic Leukemia (ALL)	8	10	10	11	14	14	16	8	8	8	9	12	12	13
C91.1 Chronic Lymphocytic Leukemia (CLL)	4	5	6	7	7	9	12	4	5	5	5	5	6	8
C91.4 Hairy Cell Leukemia	1	1	1	1	1	1	1	0	0	0	0	0	0	0
C92.0 Acute Myeloid Leukemia (AML)	8	8	8	9	9	9	9	8	8	8	9	9	9	9
C92.1 Chronic Myelogenous Leukemia (CML)	4	4	4	5	6	7	10	4	4	4	5	6	6	7

Table 2. Weighted least-squares estimates of models of potential years of life lost before age 75.

	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value
	Years 1999, 2014					Years 1999, 2010
	Model 1					Model 5
Δδ	0.0795	0.0292	2.72	0.0065		0.0799	0.0221	3.62	0.0003
	Model 2					Model 6
N_APP_1948_t	−0.0230	0.0066	−3.51	0.0005		−0.0325	0.0060	−5.41	<.0001
Δδ	−0.0895	0.0399	−2.25	0.0247		−0.0699	0.0306	−2.29	0.0222
	Model 3					Model 7
N_PRI_1948_t	−0.0183	0.0108	−1.70	0.0893	0.4251	−0.0266	0.0070	−3.82	0.0001	0.1933
N_STD_1948_t	−0.0308	0.0088	−3.50	0.0005		−0.0505	0.0154	−3.28	0.0011
Δδ	−0.0950	0.0394	−2.41	0.0159		−0.0777	0.0317	−2.45	0.0141
	Model 4					Model 8
N_APP_1948_t-4	−0.0252	0.0059	−4.25	<.0001	0.0912	−0.0333	0.0054	−6.19	<.0001	0.0939
N_APP_t-4_t	−0.0152	0.0085	−1.79	0.0728		−0.0487	0.0077	−6.35	<.0001
Δδ	−0.1020	0.0411	−2.48	0.0130		−0.0283	0.0391	−0.72	0.4695
						Model 9
N_APP_1948_t						−0.0340	0.0057	−5.98	<.0001
N_APP_t_t + 4						0.0076	0.0084	0.90	0.3677
Δδ						−0.0528	0.0285	−1.85	0.0643
						Model 10
N_APP_1948_t-4						−0.0337	0.0059	−5.71	<.0001	0.1532
N_APP_t-4_t						−0.0483	0.0079	−6.14	<.0001
N_APP_t_t + 4						0.0026	0.0097	0.27	0.7890
Δδ						−0.0248	0.0355	−0.70	0.4845

All models include ln(CASES), AGE_DIAG, and cancer-site fixed effects; estimates of these parameters are not shown to conserve space. Δδ denotes the difference between initial and final year fixed effects, e.g. δ₁₉₉₉ – δ₂₀₁₄. Standard errors are clustered within cancer sites. Observations are weighted by mean potential years of life lost before age 75.

Figure 4. Correlation across cancer sites between number of drugs approved during 2000–2014 and 1999–2014 change in log of potential years of life lost before age 75 (PYLL75), controlling for changes in incidence and mean age at diagnosis.

Table 3. Weighted least-squares estimates of models of potential years of life lost before age 65.

	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value
	Years 1999, 2014					Years 1999, 2010
	Model 1					Model 5
Δδ	0.1387	0.0281	4.93	<.0001		0.1487	0.0374	3.97	<.0001
	Model 2					Model 6
N_APP_1948_t	−0.0264	0.0055	−4.82	<.0001		−0.0322	0.0058	−5.52	<.0001
Δδ	−0.0631	0.0481	−1.31	0.1894		−0.0510	0.0377	−1.35	0.1756
	Model 3					Model 7
N_PRI_1948_t	−0.0250	0.0088	−2.85	0.0044	0.8031	−0.0307	0.0073	−4.19	<.0001	0.7360
N_STD_1948_t	−0.0288	0.0097	−2.98	0.0029		−0.0369	0.0150	−2.46	0.0141
Δδ	−0.0659	0.0473	−1.39	0.1634		−0.0543	0.0425	−1.28	0.2018
	Model 4					Model 8
N_APP_1948_t-4	−0.0281	0.0052	−5.38	<.0001	0.1280	−0.0327	0.0055	−5.94	<.0001	0.2695
N_APP_t-4_t	−0.0202	0.0069	−2.94	0.0033		−0.0437	0.0080	−5.43	<.0001
Δδ	−0.0749	0.0499	−1.50	0.1333		−0.0120	0.0518	−0.23	0.8170
						Model 9
N_APP_1948_t						−0.0332	0.0059	−5.66	<.0001
N_APP_t_t + 4						0.0056	0.0104	0.53	0.5941
Δδ						−0.0360	0.0402	−0.90	0.3696
						Model 10
N_APP_1948_t-4						−0.0332	0.0060	−5.49	<.0001	0.3209
N_APP_t-4_t						−0.0434	0.0079	−5.51	<.0001
N_APP_t_t + 4						0.0028	0.0111	0.25	0.7989
Δδ						−0.0070	0.0487	−0.14	0.8863

all models include ln(CASES), AGE_DIAG, and cancer-site fixed effects; estimates of these parameters are not shown to conserve space. Δδ denotes the difference between initial and final year fixed effects, e.g. δ₁₉₉₉ – δ₂₀₁₄. Standard errors are clustered within cancer sites. Observations are weighted by mean potential years of life lost before age 65.

Table 4. Weighted least-squares estimates of models of the observed 5-year survival rate (Equation (2)(2) $ln\left(\mathrm{S}\mathrm{U}\mathrm{R}\mathrm{V}\mathrm{\_}\mathrm{O}\mathrm{B}{\mathrm{S}}_{\mathrm{s},\mathrm{t}}/\left(1-\mathrm{S}\mathrm{U}\mathrm{R}\mathrm{V}\mathrm{\_}\mathrm{O}\mathrm{B}{\mathrm{S}}_{\mathrm{s},\mathrm{t}}\right)\right)=\beta \mathrm{N}\mathrm{\_}\mathrm{A}\mathrm{P}\mathrm{P}\mathrm{\_}1948\mathrm{\_}{\mathrm{t}}_{\mathrm{s},\mathrm{t}}+\gamma ln\left(\mathrm{C}\mathrm{A}\mathrm{S}\mathrm{E}{\mathrm{S}}_{\mathrm{s},\mathrm{t}}\right)+\pi ln\left(\mathrm{S}\mathrm{U}\mathrm{R}\mathrm{V}\mathrm{\_}\mathrm{E}\mathrm{X}{\mathrm{P}}_{\mathrm{s},\mathrm{t}}/\left(1-\mathrm{S}\mathrm{U}\mathrm{R}\mathrm{V}\mathrm{\_}\mathrm{E}\mathrm{X}{\mathrm{P}}_{\mathrm{s},\mathrm{t}}\right)\right)+{\alpha }_{\mathrm{s}}+{\delta }_{\mathrm{t}}+{\epsilon }_{\mathrm{s},\mathrm{t}}$(2) ).

Parameter	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value
	Years 2000, 2008
	Model 1
δ2000 – δ2008	−0.0971	0.0557	−1.74	0.0810
	Model 2
N_APP_1948_t	0.0240	0.0087	2.77	0.0056
δ2000 – δ2008	−0.0158	0.0528	−0.30	0.7654
	Model 3
N_PRI_1948_t	0.0471	0.0098	4.82	<.0001	0.0005
N_STD_1948_t	−0.0389	0.0200	−1.94	0.0519
δ2000 – δ2008	−0.0294	0.0491	−0.60	0.5501
	Model 4
N_APP_1948_t-4	0.0353	0.0093	3.78	0.0002	0.0003
N_APP_t-4_t	0.0212	0.0091	2.33	0.0200
δ2000 – δ2008	−0.0443	0.0569	−0.78	0.4362
	Model 5
N_APP_1948_t-8	0.0257	0.0089	2.88	0.0040	0.4546
N_APP_t-8_t	0.0214	0.0100	2.14	0.0327
δ2000 – δ2008	−0.0301	0.0594	−0.51	0.6121

all models include ln(CASES), ln(SURV_EXP/(1 – SURV_EXP)), and cancer-site fixed effects; estimates of these parameters are not shown to conserve space. Standard errors are clustered within cancer sites. Observations are weighted by number of patients diagnosed.

Figure 5. Correlation across cancer sites between number of drugs approved during 2001–008 and 2000–2008 change in log-odds of surviving at least 5 years after diagnosis, controlling for changes in incidence and expected survival rate.

Table 5. Weighted least-squares estimates of models of the number of hospital days (Equation (3)(3) $ln\left(\mathrm{H}\mathrm{O}\mathrm{S}\mathrm{P}\mathrm{\_}\mathrm{D}\mathrm{A}\mathrm{Y}{\mathrm{S}}_{\mathrm{s},\mathrm{t}}\right)=\beta \mathrm{N}\mathrm{\_}\mathrm{A}\mathrm{P}\mathrm{P}\mathrm{\_}1948\mathrm{\_}{\mathrm{t}}_{\mathrm{s},\mathrm{t}}+\gamma ln\left(\mathrm{C}\mathrm{A}\mathrm{S}\mathrm{E}{\mathrm{S}}_{\mathrm{s},\mathrm{t}}\right)+\pi \mathrm{A}\mathrm{G}\mathrm{E}\mathrm{\_}\mathrm{D}\mathrm{I}\mathrm{A}{\mathrm{G}}_{\mathrm{s},\mathrm{t}}+{\alpha }_{\mathrm{s}}+{\delta }_{\mathrm{t}}+{\epsilon }_{\mathrm{s},\mathrm{t}}$(3) ).

Parameter	Estimate	Std. Err.	Z	Pr > |Z|	Difference p-value
	Years 1997, 2013
	Model 1
δ1997 – δ2013	0.3913	0.0741	5.28	<.0001
	Model 2
N_APP_1948_t	−0.0195	0.0110	−1.77	0.0769
δ1997 – δ2013	0.2789	0.0801	3.48	0.0005
	Model 3
N_PRI_1948_t	−0.0351	0.03	−1.17	0.2439	0.5592
N_STD_1948_t	0.0037	0.04	0.09	0.9268
δ1997 – δ2013	0.2838	0.078	3.64	0.0003
	Model 4
N_APP_1948_t-4	−0.0224	0.0101	−2.22	0.0265	0.0854
N_APP_t-4_t	0.0211	0.0243	0.87	0.3848
δ1997 – δ2013	0.2293	0.0859	2.67	0.0076
	Model 5
N_APP_1948_t-8	−0.0291	0.0124	−2.34	0.0192	0.0610
N_APP_t-8_t	−0.0028	0.0125	−0.23	0.8213
δ1997 – δ2013	0.2115	0.0947	2.23	0.0256

all models include ln(CASES), AGE_DIAG, and cancer-site fixed effects; estimates of these parameters are not shown to conserve space. Standard errors are clustered within cancer sites. Observations are weighted by mean number of hospital days.

Table

Parameter	Estimate	Std. Err.	t Value	Pr > |t|
δROW – δUSA	−0.591	0.093	6.37	<.0001

Table A1. U.S. sales of top cancer drugs ($ millions).

Molecule	2010	2011	2012	2013	2014
All cancer drugs	$21,385	$22,413	$23,299	$24,942	$28,971
RITUXIMAB	$2,673	$2,879	$3,037	$3,199	$3,351
BEVACIZUMAB	$2,993	$2,581	$2,543	$2,618	$2,790
IMATINIB	$1,363	$1,560	$1,674	$1,872	$2,333
DENOSUMAB	$27	$478	$932	$1,229	$1,552
ABIRATERONE ACETATE	$0	$174	$454	$767	$1,010
LENALIDOMIDE	$311	$431	$537	$528	$974
EVEROLIMUS	$108	$182	$426	$714	$841
DASATINIB	$185	$291	$380	$549	$725
LEUPRORELIN	$631	$650	$639	$651	$692
IPILIMUMAB	$0	$339	$507	$541	$687

Source: IMS Health.

these figures represent the value of drugs sold at ‘list prices’; rebates are not accounted for.

Table A2. Data on premature mortality, incidence, and hospitalization, by cancer site.

	Potential years of life lost before age 75		Potential years of life lost before age 65		Average annual number diagnosed before age 75 (SEER 9)		Mean age at diagnosis of patients diagnosed before age 75 (SEER 9)		Number of hospital days
	1999	2014	1999	2014	1988–1997	2003–2012	1988–1997	2003–2012	1997	2013
C00-C14 Head and Neck Cancer	74,052	87,161	34,112	36,161	2,166	2,849	58.0	57.2	159,120	139,442
C15 Esophageal Cancer	99,905	121,933	40,205	45,788	821	959	62.1	61.6	132,260	103,241
C16 Stomach (Gastric) Cancer	95,716	94,763	43,796	43,253	1,199	1,449	61.4	59.7	262,018	199,976
C18 Colon Cancer	286,278	301,178	116,093	129,363	5,376	5,265	62.7	60.1	1,044,400	689,071
C20 Rectal Cancer	48,443	60,020	21,343	26,350	1,710	2,110	61.1	57.5	364,882	268,129
C22 Liver Cancer	109,239	220,317	49,824	83,537	869	2,245	59.3	58.6	81,381	141,369
C25 Pancreatic Cancer	194,392	269,310	75,797	96,095	1,627	2,508	62.8	61.7	276,671	265,989
C34 Lung Cancer	1,089,130	1,010,583	393,295	336,483	10,884	10,652	63.0	62.9	1,183,100	829,940
C40-C41 Bone Cancer	26,401	29,409	18,501	19,729	250	292	35.3	37.8	36,229	42,784
C43 Melanoma	88,005	85,268	47,315	40,233	3,634	5,515	50.5	53.7	15,321	11,540
C44 Basal Cell Carcinoma	15,193	19,996	6,528	7,616	291	439	50.0	53.4	28,348	27,511
C45 Malignant Mesothelioma	14,113	12,243	4,843	4,103	153	138	62.8	63.2	18,289	14,035
C46 Kaposi Sarcoma	590	570	435	345	262	170	39.1	43.0	1,439	0
C49 Soft Tissue Sarcoma	49,563	61,264	29,373	34,339	521	719	46.5	48.9	51,342	58,920
C50 Breast Cancer	427,324	401,888	214,099	186,728	14,100	17,705	56.6	56.4	326,240	157,356
C51 Vulvar Cancer	3,385	6,000	1,435	2,530	205	266	57.0	57.2	18,898	10,365
C53 Cervical Cancer	72,735	70,125	43,145	39,890	1,108	930	46.4	47.6	91,193	39,409
C54.1 Endometrial Cancer	19,773	31,253	7,393	11,158	2,429	3,837	58.6	58.5	159,234	100,397
C56 Ovarian, Fallopian Tube, or Primary Peritoneal Cancer	115,547	115,525	51,707	48,060	1,454	1,548	56.3	55.9	188,761	156,756
C60 Penile Cancer	1,565	2,388	690	1,038	56	87	60.2	60.4	3,157	4,730
C61 Prostate Cancer	76,492	83,051	16,342	20,171	12,716	14,350	65.7	62.9	389,198	153,372
C62 Testicular Cancer	12,311	12,710	8,836	9,110	731	852	34.2	34.9	9,287	12,436
C64-C65 Kidney (Renal Cell) Cancer	92,658	103,563	40,028	40,348	1,990	3,796	58.0	57.7	210,836	229,344
C67 Bladder Cancer	49,151	60,795	16,286	19,515	3,059	3,602	62.7	62.7	244,471	191,431
C69.2 Retinoblastoma	422	205	257	110	21	26	4.9	3.7	381	895
C71 Brain Tumors	176,630	199,026	97,895	101,896	1,291	1,546	45.9	46.9	223,031	221,945
C73 Thyroid Cancer	9,078	12,773	3,968	5,113	1,605	4,281	44.5	47.7	45,057	46,112
C74.9 Neuroblastoma	9,095	8,178	5,645	4,918	35	37	21.4	17.8	14,985	13,315
C81 Hodgkin Lymphoma	26,015	15,628	17,000	9,463	683	744	36.1	38.0	41,761	44,139
C82-C85 Non-Hodgkin Lymphoma	170,143	121,130	79,183	50,615	3,439	4,171	54.6	56.4	456,126	386,511
C90 Multiple Myeloma and Other Plasma Cell Neoplasms	62,555	63,338	22,540	20,973	882	1,370	62.3	61.2	162,886	213,160
C91.0 Acute Lymphoblastic Leukemia (ALL)	33,325	31,268	23,035	20,743	326	446	19.7	22.0	113,448	156,189
C91.1 Chronic Lymphocytic Leukemia (CLL)	16,876	13,585	5,101	4,065	668	1,157	62.8	62.2	47,874	32,660
C91.4 Hairy Cell Leukemia	610	360	275	120	73	89	54.7	54.3	0	0
C92.0 Acute Myeloid Leukemia (AML)	66,432	77,876	35,282	36,561	495	710	53.0	54.3	248,775	390,516
C92.1 Chronic Myelogenous Leukemia (CML)	22,590	7,940	13,065	4,145	240	329	52.0	51.4	54,344	33,370

Table A3. Observed and expected 5-year survival rates and number of patients diagnosed, by cancer site, 2000 and 2008.

	Observed 5-year survival rate		Expected 5-year survival rate		Number of patients diagnosed
Cancer site	2000	2008	2000	2008	2000	2008
C15 Esophageal Cancer	13.5%	14.1%	86.2%	88.0%	2,754	2,994
C16 Stomach (Gastric) Cancer	18.8%	26.0%	83.9%	86.2%	5,055	4,972
C18 Colon Cancer	50.6%	54.6%	80.7%	83.9%	22,409	21,575
C20 Rectal Cancer	54.7%	59.4%	85.6%	89.1%	9,208	8,998
C22 Liver Cancer	10.4%	16.5%	90.8%	91.6%	3,149	5,253
C25 Pancreatic Cancer	4.2%	6.3%	86.2%	88.6%	6,663	8,226
C34 Lung Cancer	12.6%	15.2%	86.2%	86.6%	38,808	38,678
C40-C41 Bone Cancer	66.4%	68.0%	96.1%	96.7%	619	691
C43 Melanoma	81.1%	82.2%	90.1%	89.7%	10,889	13,949
C45 Malignant Mesothelioma	5.0%	8.3%	80.7%	87.4%	647	653
C49 Soft Tissue Sarcoma	59.4%	58.7%	91.0%	91.2%	1,927	2,295
C50 Breast Cancer	79.5%	81.8%	90.0%	91.4%	45,901	47,537
C51 Vulvar Cancer	65.4%	60.0%	86.0%	86.7%	737	778
C53 Cervical Cancer	68.1%	65.0%	95.8%	96.5%	3,557	3,248
C56 Ovarian, Fallopian Tube, or Primary Peritoneal Cancer	40.3%	43.4%	92.1%	93.0%	4,869	5,014
C60 Penile Cancer	61.4%	57.1%	82.3%	83.8%	246	258
C61 Prostate Cancer	80.6%	85.2%	81.9%	86.4%	49,820	53,557
C62 Testicular Cancer	94.2%	94.4%	98.5%	98.4%	2,049	2,237
C64-C65 Kidney (Renal Cell) Cancer	56.9%	66.6%	88.3%	90.0%	7,148	10,573
C67 Bladder Cancer	62.6%	62.8%	80.3%	81.4%	11,874	12,633
C69.2 Retinoblastoma	78.7%	76.2%	89.4%	90.5%	557	562
C71 Brain Tumors	28.9%	29.9%	97.2%	97.2%	4,084	4,520
C73 Thyroid Cancer	92.2%	93.9%	96.1%	96.1%	5,170	9,228
C81 Hodgkin Lymphoma	81.2%	84.9%	97.1%	96.9%	1,931	2,170
C82-C85 Non-Hodgkin Lymphoma	54.6%	62.4%	88.0%	88.6%	11,964	13,585
C90 Multiple Myeloma and Other Plasma Cell Neoplasms	28.9%	42.4%	85.5%	87.4%	3,644	4,241
C91.0 Acute Lymphoblastic Leukemia (ALL)	62.3%	66.5%	99.1%	99.1%	1,145	1,311
C91.1 Chronic Lymphocytic Leukemia (CLL)	64.0%	68.3%	81.9%	83.1%	2,550	3,153
C92.0 Acute Myeloid Leukemia (AML)	16.6%	23.6%	94.8%	95.3%	2,303	2,454
C92.1 Chronic Myelogenous Leukemia (CML)	45.5%	57.0%	90.6%	90.9%	1,100	1,209

Table A4. Calculation of the cost per life-year gained in 2014 from cancer drugs approved during 2000–2014.

Line	Description	Basis	Value
(1)	Total expenditure (before rebates) on cancer drugs in 2014	IMS Institute for Healthcare Informatics (2015)	$32,300,000,000
(2)	Fraction of total expenditure on post-1999 drugs	Author’s calculations based on IMS MIDAS data	67.7%
(3)	Expenditure (before rebates) on post-1999 cancer drugs in 2014	(1) * (2)	$21,867,100,000
(4)	Rebates on post-1999 drugs, assuming 25% rebate rate (Herper (2012))	25% * (3)	$5,466,775,000
(5)	Reduction in old drug cost (assumed to be 25% of new drug cost) (Lichtenberg (2014c))	25% * ((3) – (4))	$4,100,081,250
(6)	Hospital cost reduction	Author’s calculations based on estimates of models 1 and 5 in Table 6.	$4,770,000,000
(7)	Net expend–all ages	(3) – ((4) + (5) + (6))	$7,530,243,750
(8)	Net expend–below age 75 (assumed to be 75% of net expenditure for all ages)	75% * (7)	$5,647,682,813
(9)	Life-years before age 75 gained, controlling for incidence and mean age at diagnosis	Author’s calculations based on estimates of models 1 and 2 in Table 5a.	719,133
(10)	Cost per life-year gained before age 75	(9)/(8)	$7,853

Figure A1. Number of cancer and other new molecular entities launched worldwide, 1982–2015.

Figure A2. Mean ratio of the quantity of a cancer drug sold n years after approval to the quantity of the same drug sold 9 years after approval.