Association between expedited review designations and the US or global burden of disease for drugs approved by the US Food and Drug Administration, 2010–2019: a cross-sectional analysis

Abstract

Objectives

Pharmaceutical innovation can contribute to reducing the burden of disease in human populations. This research asks whether products approved by the US Food and Drug Administration (FDA) from 2010 to 2019 and expedited review programmes incentivising development of products for serious disease were aligned with the US or global burden of disease.

Design

Cross-sectional study.

Outcome measures

Association of FDA product approvals (2010–2019), first approved indications, designations for expedited review with the burden of disease (disability-adjusted life years (DALYs)), years of life lost (YLL) and years of life lived with disability (YLD) for 122 WHO Global Health Estimates (GHE) conditions in US and global (ex-US) populations.

Results

The FDA approved 387 drugs in 2010–2019 with lead indications associated with 59/122 GHE conditions. Conditions with at least one new drug had greater US DALYs (p=0.001), US YLL (p<0.001), global DALYs (p=0.030) and global YLL (p=0.004) but not US YLD (p=0.158) or global YLD (p=0.676). Most approvals were for conditions in the top quartile of US DALYs or YLL, but <27% were for conditions in the top quartile of global DALYs or YLL. The likelihood of a drug having one or more designations for expedited review programmes was negatively associated (OR<1) with US DALYs, US YLD and global YLD. There was a weak negative association with global DALYs and a weak positive association (OR>1) with US and global YLL.

Conclusions

FDA drug approvals from 2010 to 2019 were more strongly aligned with US than global disease burden. Designations for expedited review were not aligned with either the US or global burdens of disease and may inadvertently disincentivise development of products addressing global disease burdens. These results may inform policies to better align pharmaceutical innovation with the burdens of disease.

Strengths and limitations of this study.

• Extensively characterised dataset of drugs approved by US Food and Drug Administration over 10-year period.

• Logistic regression used to estimate the association (OR) between a product being designated for expedited approval and the burden of disease for Global Health Estimates (GHE) condition that includes the first approved indication.

• Drugs assigned WHO burden of disease metrics for GHE condition including ICD-10 code for lead indication, which may not capture follow-on indications or off-label use.

• GHE condition may include ICD-10 codes for which the drug is not indicated.

• GHE conditions do not delineate the disease burden associated with subsets or stages of disease.

Introduction

This study explores the relationships between the burden of disease in US and global populations, drugs approved by the US Food and Drug Administration (FDA) from 2010 to 2019, and expedited review programmes designed to promote development of drugs for ‘serious diseases’ with selected characteristics. Previous work suggests that the number of drug approvals in different therapeutic areas is generally aligned with their US burden of disease¹ but not their contribution to global disease burden.^1–6 In the context of evidence that pharmaceutical R&D is driven by the available market size and anticipated returns on investment,^{2 7–11} industry’s underinvestment in products for morbidities more prevalent outside the USA^{5 6} can be seen as a market failure in which the pharmaceutical industry fails to develop products for markets that are inadequate to justify the investment cost.^12–14 While alternative business models involving intergovernmental organisations, non-governmental organisations, public–private partnerships, non-profit entities and philanthropies have had notable success,^14–16 it is classically the role of government to rectify market failures through regulatory or economic policy.¹⁷

Over the past 40 years, regulations in the USA and EU have been implemented to redress market failures involving rare (orphan) diseases, paediatric disorders and serious diseases with attributes that make investment in these products unattractive.^18–21 The prototype for such regulations was the US Orphan Drug Act of 1983,²² which successfully promoted drug development for rare diseases,^{23 24} but had unanticipated consequences related to drug pricing and applications to precision medicine.²⁵ Subsequently, four expedited review programmes were legislated: ‘fast track’, ‘breakthrough’, ‘accelerated’ and ‘priority’ (online supplemental table 1) to incentivise development of products that address significant unmet medical needs, provide substantial improvement over existing therapies, require surrogate endpoints to establish efficacy, offer significant improvements in safety or effectiveness, or address paediatric, tropical, or infectious diseases.^{20 26–28} While questions have been raised about the efficacy, safety and innovativeness of products approved through these programmes,^{20 29–31} the majority of drugs coming to market now have at least one designation for expedited review programmes.²⁰ Additionally, the Qualified Infectious Disease Product designation, created through the 2012 ‘Generating Antibiotic Incentives Now’ (GAIN) Act, provides selected products with fast track and priority review designations as well as greater market exclusivity.²⁸

While the FDA does not have jurisdiction over product approvals outside the USA, US markets influence the development priorities of the global pharmaceutical industry³² and most new drugs are approved by the FDA before being approved by the European Medicines Agency or other regulatory authorities in other countries.^{30 33} As such, it has been observed that programmes designed to selectively incentivise product development in the USA may shape the portfolio of products available to global populations.^34–36 For example, most neglected tropical diseases are rare in the USA and qualify for FDA review and approval under the Orphan Drug Act.^{35 36} Additionally, products to prevent or treat certain tropical diseases may qualify sponsors for a priority review voucher that can be sold to sponsors of other (unrelated) drugs, reducing the net cost of developing drugs for tropical disease.^{37 38} While a number of drugs for tropical disease have taken advantage of this programme,^{39 40} recent analyses suggest there is insufficient evidence to assess its impact on promoting development of products for tropical disease.^{37 41}

Objectives

The objective of this study was to ask whether the FDA-approved drugs from 2010 to 2019 were aligned with measures of the US or global burden of disease at the outset of the decade and whether this alignment was promoted by the four expedited review programmes in the USA. Specifically, this study identified drugs approved by the FDA, designations for expedited review programmes, their first approved indications, the Global Health Estimates (GHE) conditions associated with these indications, and WHO metrics describing the US and global burden of disease associated with these conditions. The burden of disease is classically measured in disability-adjusted life years (DALYs),^{42 43} calculated as the sum of years of potential life lost (YLL) and years of healthy life lost to disability (YLD). YLL measures the disease burden associated with premature mortality calculated from the age of death to life expectancy. YLD measures the disease burden associated with morbidity, calculated as the product of a condition’s disability weight and its prevalence in a population.^{44 45}

The analysis describes the relationship between the number of product approvals for indications in each GHE condition and its US or global disease burden measured by DALYs, YLL or YLD. The analysis further estimates the relative likelihood (OR) of a product receiving one or more designations for expedited review based on metrics of disease burden for the condition comprising the first approved indication. These results are discussed in the context of the intended and unintended consequences of expedited review programmes and the continuing challenge of achieving global equity in pharmaceutical innovations that address humankind’s most burdensome diseases.

Methods

Study design

This cross-sectional study describes associations between the lead indications for drugs approved by the FDA from 2010 to 2019, designations of these products for expedited review, and WHO GHE metrics for the US and global burden of disease of conditions comprising these indications.

Data sources

FDA-approved products from 2010 to 2019 (New Drug Application (NDA) or Biologics License Application (BLA)) and dates of first approval were identified from annual FDA reports (CDER⁴⁶; CBER⁴⁷). Products derived from blood or tissue, diagnostic agents and vaccines were excluded. The 2010–2019 range excludes products developed and practices implemented during COVID-19.

Expedited and orphan designations for products approved by CDER were identified in the CDER annual reports or the Summary Basis for Regulatory Action for each product. The first approved indication was identified in the Full Prescribing Information for each product,⁴⁸ matched to corresponding ICD-10 code via the WHO ICD-10 lookup tool,⁴⁹ and categorised by GHE condition. Each GHE condition comprises one or more ICD-10 codes (online supplemental table 2).⁵⁰ For some analyses, GHE conditions were further categorised into 13 therapeutic areas adapted from the WHO Global Health Observatory (GHO) data repository disease hierarchy.⁵¹

DALYs, YLL and YLD for US and global (ex-US) populations for 2010 were obtained from the WHO GHO data repository GHE-2020 for all 132 available GHE conditions.⁵² Ten GHE conditions in the categories ‘sudden infant death syndrome’ and ‘injuries’ were excluded. The final dataset consisted of 122 GHE conditions.

Statistical methods

Differences between the burden of disease metrics for conditions with one or more drug approvals or no drug approvals were analysed with Mann-Whitney tests. Significance is inferred with p<0.05. Bonferroni corrections of 8 were applied to account for multiple testing, with a corrected p<0.0065 corresponding to the p=0.05 threshold. Alignment of drug approvals with disease burden was assessed from the number of drug approvals associated with each quartile of GHE conditions ranked by US or global DALYs, YLD or YLL with Q4 representing the greatest disease burden and Q1 the lowest.

Probability estimates (OR) of a product receiving an expedited designation (dependent variable) as a function of the DALYs, YLL or YLD for the GHE condition associated with its indication (independent variable) were assessed using univariate binary logistic regression models (model 1):

Model 1: Y_i~ β₀+β₁X_i

OR=Exp(β₁)

where Y_i indicates designation or no designation for a drug i, β₀ represents the Y intercept, β₁ represents the slope coefficient, X_i represents the independent variable (DALYs, YLL or YLD) of the disease corresponding to drug i, and Exp(x) is e^x. Analysis was performed independently for each expedited designation and metric of disease burden. Model 1 was used for datasets comprising all drugs, drugs indicated for neoplastic disease and drugs indicated for non-neoplastic disease.

Mann-Whitney tests were performed comparing ranked DALYs, YLL or YLD for neoplasm versus non-neoplasm GHE conditions. Mann-Whitney and logistic regression were performed in SPSS Statistics V.27 (IBM). All other analyses were performed in Excel.

This manuscript follows Strengthening the Reporting of Observational Studies in Epidemiology guidelines where applicable.

Patient and public involvement

None.

Results

Drug approvals, expedited review designations and first approved indications

The FDA approved 387 drugs from 2010 to 2019 excluding vaccines and biological products derived from blood or tissues (online supplemental table 3). Of these, 227/387 (58.7%) were granted at least one designation for expedited review programmes including 49 (12.7%) for accelerated approval, 78 (20.2%) for breakthrough therapy, 142 (36.7%) for fast track and 207 (53.5%) for priority review. Across all approved products, there was an average of 1.2 designations for expedited review programmes. These fractions are consistent with observations by others (table 1).^{29 53 54} Seventeen products were associated with the GAIN Act, which includes priority review, and is not considered separately.

Table 1.

Number of drugs approved by the Food and Drug Administration (FDA), 2010–2019, and designations for expedited review by therapeutic area

Therapeutic area	Drugs	Accelerated	Breakthrough	Fast track	Priority	Average	>1
	n	Drugs, n (% drugs in class)
Neoplasms	97	38 (39.2%)	39 (40.2%)	46 (47.4%)	80 (82.5%)	2.1	90 (92.8%)
Endocrine, blood, immune disorders	73	3 (4.1%)	16 (21.9%)	28 (38.4%)	37 (50.7%)	1.2	39 (53.4%)
Neurological and sense organ	50	2 (4%)	7 (14%)	11 (22%)	22 (44%)	0.8	22 (44%)
Infectious and parasitic diseases	37	2 (5.4%)	10 (27%)	22 (59.5%)	27 (73%)	1.7	28 (75.7%)
Digestive and genitourinary	31	1 (3.2%)	–	13 (41.9%)	16 (51.6%)	1.0	16 (51.6%)
Respiratory	18	–	2 (11.1%)	5 (27.8%)	5 (27.8%)	0.7	6 (33.3%)
Skin and oral	16	–	1 (6.3%)	2 (12.5%)	4 (25%)	0.4	4 (25%)
Cardiovascular	16	1 (6.3%)	–	5 (31.3%)	7 (43.8%)	0.8	8 (50%)
Musculoskeletal	13	–	–	2 (15.4%)	2 (15.4%)	0.3	2 (15.4%)
Diabetes mellitus	12	–	–	–	1 (8.3%)	0.1	1 (8.3%)
Mental and substance use	12	–	1 (8.3%)	2 (16.7%)	2 (16.7%)	0.4	3 (25%)
Maternal, neonatal and congenital	4	–	–	1 (25%)	–	0.3	1 (25%)
Other	8	2 (25%)	2 (25%)	5 (62.5%)	4 (50%)	1.8	7 (87.5%)
Total	387	49 (12.7%)	78 (20.2%)	142 (36.7%)	207 (53.5%)	1.2	227 (58.7%)

FDA approved drugs and expedited designations for each drug are shown in online supplemental table 3.

The first approved indications for products in this dataset were associated with 188 unique ICD-10 codes (online supplemental table 3) corresponding to 59/122 (48%) GHE conditions (online supplemental table 2). The GHE conditions were further categorised into 13 therapeutic areas (table 2). The largest fraction of drug approvals was for ‘neoplasms’ (n=97/387) followed by ‘endocrine, blood and immune disorders’ (n=73/387) and ‘neurological and sense organ’ (n=50/387). Drugs for infectious and parasitic disease included treatments for hepatitis C, HIV, malaria, Chagas disease, leishmaniasis, fascioliasis, onchocerciasis and tuberculosis, diseases of particular concern in the tropical and developing world. Table 2 shows the US and global burden of disease for all 122 GHE conditions categorised by therapeutic area measured by DALYs, YLL and YLD. Notably, the US burden estimates represent <6% of global DALYs, YLL and YLD.

Table 2.

US and global burden of disease estimates (‘000s) by therapeutic area as measured by DALYs, YLL and YLD (GHE-2020)

Therapeutic area*	Drugs†	GHE conditions	Global DALYs	Global YLD	Global YLL	US DALYs	US YLD	US YLL
Neoplasms	97	25	229 633	4683	224 951	14 108	623	13 485
Endocrine, blood, immune disorders	73	4	24 492	9973	14 519	1119	271	848
Neurological and sense organ	50	14	178 243	143 600	34 644	9106	5779	3327
Infectious and parasitic diseases	37	12	414 764	37 906	376 858	1658	324	1334
Digestive and genitourinary	31	15	139 286	30 384	108 902	5575	1739	3836
Respiratory	18	6	261 788	36 760	225 028	6802	2147	4654
Skin and oral	16	5	38 337	35 604	2734	1860	1767	93
Cardiovascular	16	6	379 934	28 353	351 580	15 339	1904	13 435
Musculoskeletal	13	5	91 154	87 532	3622	6415	6126	289
Diabetes mellitus	12	1	55 700	24 816	30 885	3506	1998	1508
Mental and substance use	12	11	159 681	146 887	12 793	11 907	9935	1972
Maternal, neonatal and congenital	4	11	339 320	24 149	315 171	3091	991	2099
Other	8	7	117 096	52 893	64 203	1379	579	799
Total	387	122	2 429 428	663 540	1 765 889	81 863	34 183	47 680

*Burden of disease (000’s) for each of the 13 therapeutic areas calculated by sum of the burden measured by DALYs, YLL or YLD for GHE conditions included in that therapeutic area.

†Number of drugs approved by the FDA, 2010–2019, categorised by the therapeutic area of the first approved indication. Drug approvals and burden of disease metrics for all 122 GHE conditions are provided in online supplemental table 2. Notably, the US burden estimates represent <6% of global DALYs, YLL and YLD (US DALYs/global DALYs=3.37%; US YLL/global YLL=2.70%; US YLD/global YLD=5.15%).

DALYs, disability-adjusted life years; GHE, Global Health Estimates; YLD, years of life lived with disability; YLL, years of life lost.

Products indicated for neoplasm were most likely to have expedited designations, with 90/97 (92.8%) having at least one expedited designation and an average of 2.1 designations per product. Products for infectious and parasitic diseases were also likely to have expedited designations, with 28/37 (75.7%) having at least one expedited designation and an average of 1.7 designations per product (table 1). Seven drugs were indicated for diseases explicitly listed by the FDA as tropical diseases eligible for a priority review voucher (online supplemental table 4). Therapeutic areas with the lowest average number of expedited designations were diabetes mellitus (0.1 designations per product), maternal, neonatal and congenital (0.3 designations per product), musculoskeletal (0.3 designations per product) and mental and substance use (0.4 designations per product) (table 1).

Associations between drug approvals and US or global burden of disease

Table 3 compares US and global DALYs, YLL and YLD for 59 conditions associated with at least one new drug approval versus 63 conditions with no approval. US DALYs and YLL were significantly higher for conditions with at least one new drug approval (US DALYs, U=1193, p=0.001; US YLL, U=1144, p<0.001). Global DALYs and YLL were also higher for conditions with at least one new drug approval (global DALYs, U=1436, p=0.030; global YLL, U=1304, p=0.004), though the difference in global DALYs was not significant after Bonferroni correction.

Table 3.

Mann-Whitney analysis of differences in disease burden between GHE disease states with or without drug approvals, 2010–2019

Burden metric	Approved drug	All GHE disease states					Non-neoplasm GHE disease states
		N	Median (‘000s)	Mean rank	Mann-Whitney U	Asymp. sig. (two-tailed)	N	Median (‘000s)	Mean rank	Mann-Whitney U	Asymp. sig. (two-tailed)
Global DALYs	No	63	7212	54.8	1436	0.030	53	7255	43.7	883	0.040
	Yes	59	11 197	68.7			44	13 503	55.4
Global YLD	No	63	1741	60.2	1777	0.676	53	3127	46.7	1044	0.377
	Yes	59	1341	62.9			44	3950	51.8
Global YLL	No	63	1129	52.7	1304	0.004	53	812	43.0	846	0.020
	Yes	59	6203	70.9			44	6196	56.3
US DALYs	No	63	143	50.9	1193	0.001	53	132	41.4	761	0.003
	Yes	59	523	72.8			44	513	58.2
US YLD	No	63	47	57.1	1583	0.158	53	57	45.0	955	0.126
	Yes	59	70	66.2			44	124	53.8
US YLL	No	63	15	50.2	1144	<0.001	53	8	40.7	724	0.001
	Yes	59	287	73.6			44	114	59.0

DALYs, disability-adjusted life years; GHE, Global Health Estimates; YLD, years of life lived with disability; YLL, years of life lost.

There was no significant difference in the YLD for conditions with drug approvals and without an approved drug (US YLD, U=1583, p=0.158; global YLD, U=1777, p=0.676). When the analysis was repeated using only non-neoplastic conditions, US DALYs and YLL remained significantly higher for conditions with at least one approval (US DALYs, U=761, p=0.003; US YLL, U=724, p=0.001). There was no significant difference between conditions with or without at least one drug approval across any of the global burden of disease measures (table 3).

Figure 1 illustrates the relationship between US and global DALYs for the 122 GHE conditions ranked by US DALYs, the boundaries of quartiles 1–4 of US to global DALYs, and the number of new drugs for indications in each condition. While the US and global DALYs are highly correlated (R=0.615, p<0.001), there are also substantive differences (online supplemental table 5). For example, multiple neoplasms, cardiovascular diseases, drug and alcohol use disorders, skin diseases, digestive diseases, osteoarthritis and neurological conditions were in Q4 (the top quartile) of the US disease burden but were in Q2 or Q3 of the global burden. Conversely, tuberculosis, meningitis, maternal conditions, rheumatic heart disease, and parasitic and vector diseases were in Q1 (the lowest quartile) of the US disease burden but were in Q3 or Q4 of the global burden (online supplemental table 2). Figure 1 also illustrates the number of drug approvals for indications in each condition, illustrating the greater number of drugs for conditions in the highest quartiles of US disease burden and the distribution of these conditions across all four quartiles of the global burden.

Figure 1.

Relationship between Food and Drug Administration (FDA) drug approvals and the US or global burden of disease of Global Health Estimates (GHE) conditions associated with their lead indications. The 122 GHE conditions in the GHE-2020 study were ranked by US disability-adjusted life years (DALYs). Bars indicate disease burden measured in US DALYs (left, blue) or global DALYs (right, orange). Bars are coloured by the number of products approved by the FDA, 2010–2019, with first approved indications for that condition. Empty bars indicate no approvals. Darker bars indicate larger numbers of approvals. Grid lines indicate quartiles of GHE conditions ranked by disease burden measured in DALYs with Q4 representing conditions with the highest disease burden and Q1 representing conditions with the lowest disease burden.

Figure 2A shows the number of drugs for conditions in each quartile of disease burden measured by DALYs, YLL or YLD for both US and global populations (data breakdown in online supplemental table 6). The majority of drug approvals were for diseases indicated for conditions in the top quartile (Q4) of US DALYs. This was not evident for the global burden of disease, where <27% of products were approved for conditions in the top quartile of global DALYs, YLL or YLD (figure 2A, online supplemental table 6). Similarly, Mann-Whitney analysis shows that conditions with at least one drug approval had significantly higher US DALYs and YLL than conditions without approved products but lower global burden of disease metrics than those without an approved product (table 3).

Figure 2.

Association of Food and Drug Administration (FDA) drug approvals and designations for expedited review across quartiles of US and global burden of disease. (A) Number of FDA drug approvals, 2010–2019, is shown for quartiles of Global Health Estimates (GHE) conditions ranked by US (blue) or global (orange) burden of disease. The burden of disease is measured by DALYs, YLL or YLD as shown. The number of FDA approvals is shown for neoplastic (patterned fill) and non-neoplastic conditions (solid fill). (B) The proportion of drugs approved, 2010–2019, with at least one expedited designation across quartiles of GHE conditions ranked by disease burden. The burden of disease for 122 GHE conditions is measured by DALYs, YLL and YLD and was divided by quartile with Q1 representing conditions with the lowest disease burden and Q4 representing conditions the highest disease burden. DALYs, disability-adjusted life years; YLD, years of life lived with disability; YLL, years of life lost.

These analyses demonstrate a strong association between drug approvals and the US burden of disease measured by DALYs or YLL and a demonstrably weaker association between drug approvals and global disease metrics. These analyses also show that these associations are largely limited to measures of premature mortality (YLL) and that there is little association between drug approvals and measures of disability (YLD).

Figure 2A also illustrates differences in the association of drugs for neoplastic disease and non-neoplastic disease and quartiles of US or global disease burden. Drugs for neoplasms are predominantly associated with conditions in the top two quartiles of US DALYs and YLL, the lowest two quartiles of US and global YLD, and widely across quartiles of global DALYs and YLL. This suggests that the large number of approvals for neoplastic indications contributes to the overall absence of an association between drug approvals and either US or global YLD. This is particularly pronounced for global YLD, where most neoplastic conditions are in the lowest quartile (figure 2A). Nevertheless, there were fewer products for conditions in the top quartile of both US and global YLD than for conditions in the third quartile.

Association between expedited review designations and burden of disease

Figure 2B shows the relationship between the fraction of approvals having one or more expedited designations and the quartiles of US or global DALYs, YLL or YLD. Figure 3A,C shows the calculated ORs of a product having at least one designation for expedited review as a function of US or global burden of disease metrics associated with the condition comprising the lead indication.

Figure 3.

ORs representing the likelihood of receiving an expedited designation as a function of burden of disease. The OR of an approved drug having a designation for expedited review was calculated as a function of burden of disease metrics for Global Health Estimates conditions comprising the first approved indication. Data are shown for analyses of DALYs, YLL or YLD associated with the US burden of disease (A, B) or global burden of disease (C, D) using either the full data set (A, C) or excluding drugs indicated for neoplasm (B, D). Analyses were performed separately for accelerated, breakthrough, fast track or priority designations or at least one designation (‘1+designation’). (A) OR calculated for 387 products approved in 2010–2019 and US burden of disease metrics. (B) OR calculated for 290 products indicated for non-neoplastic conditions and US burden of disease metrics. (C) OR calculated for 387 products approved in 2010–2019 and global burden of disease metrics. (D) OR calculated for 290 products indicated for non-neoplastic conditions and global burden of disease metrics approved in 2010–2019. Data are shown as the average and 95% CI. Filled symbols represent data with p<0.0065 corresponding to the significance threshold of p=0.05 after a Bonferroni correction of 8 to account for multiple testing. DALYs, disability-adjusted life years; YLD, years of life lived with disability; YLL, years of life lost.

There was a negative trend in the fraction of drugs with at least one expedited review designation across quartiles for either US or global DALYs (figure 2B). This trend was also evident with ORs significantly <1 for a product having at least one expedited review designation as a function of US DALYs (OR=0.526, 95% CI 0.335 to 0.826, p=0.005) (figure 3A) and a negative association with global DALYs (OR=0.673, 95% CI 0.428 to 1.057, p=0.085) (figure 3C).

There was a positive trend in the fraction of drugs having at least one expedited designation across quartiles of either US or global YLL (figure 2B). This was also evident by the OR of >1 for a product having at least one designation for expedited review as a function of US or global YLL (US: OR=1.274, 95% CI 0.924 to 1.758, p=0.140; global: OR=1.501, 95% CI 1.150 to 1.958, p=0.003) (figure 3A,C).

In contrast, there was a negative trend in the fraction of drugs having at least one expedited designation across quartiles of US or global YLD (figure 2B). This trend was also evident in the statistically significant ORs <1 for a product having at least one expedited designation as a function of US or global YLD (US: OR=0.243 95% CI 0.168 to 0.352, p<0.001; global: OR=0.215, CI 0.149 to 0.311, p<0.001) (figure 3A,C). ORs and full binary logistic regression variables for US and global expedited designations are shown in figure 3 and online supplemental table 7, respectively. ORs for each individual designation were similar to ORs for having at least one expedited designation (figure 3A–D).

Similar patterns were observed when considering the 290 drugs approved across 97 non-neoplastic conditions (figure 3B,D). For these drugs, the OR of having at least one expedited review designation was significantly <1 as a function of US DALYs and not significantly <1 as a function of global DALYs (US: OR=0.440 95% CI 0.269 to 0.719, p=0.001; global: OR=0.751, CI 0.454 to 1.243, p=0.265). As a function of US and global YLD, the OR of non-neoplastic drugs having at least one expedited review designation was significantly <1 (US: OR=0.375 95% CI 0.098 to 0.131, p<0.001; global: OR=0.315, CI 0.197 to 0.505, p<0.001).

Analysis of the 97 products indicated for neoplastic disease was limited by the paucity of products without expedited designations and, consequently, large SEs. These results are shown in online supplemental table 8.

Discussion

This research examined the association between the first approved indication for FDA-approved drugs from 2010 to 2019 and the US or global burden of disease for the associated GHE disease condition. Specifically, this analysis asked two questions; first, whether drug approvals for each GHE condition were related to the US or global burden of disease measured by DALYs, YLL or YLD; and second, whether products with indications having greater US or global burden of disease were more likely to be eligible for incentives associated with expedited review programmes.

Drug approvals and the US or global burden of disease

These analyses demonstrated a strong association between drug approvals and conditions with higher US DALYs. This was evident in the observation that the majority of all approvals had a first indication associated with conditions in the top quartile of US DALYs and that conditions with at least one new drug approval had significantly higher US DALYs than those without new approvals. In contrast, <27% of drug approvals were indicated for conditions in the top quartile of global disease burden, and there was a less pronounced difference in the global burden of disease for conditions with or without new drug approvals.

These results are consistent with previous studies showing that US drug approvals are more closely associated with the US burden of disease than the global burden of disease.1 2 4–6 55 Since the global (ex-US) burden of disease represents >96% of the total DALYs associated with the 122 GHE conditions, these results suggest there is little alignment between US drug approvals and the composite health needs of global human populations.

An unexpected finding in this work was the disparity between the strong association observed between US drug approvals and the disease burden associated with premature death (YLL) and the absence of any association between drug approvals and years of life lived with disability (YLD). While this disparity was partially related to the large number of drugs for neoplastic conditions with high YLL and lower YLD, the disparity was still evident when considering only drugs for non-neoplastic conditions.

The under-representation of drugs indicated for conditions with the highest burden of disability is paradoxical given that the burden of disability measured by US YLD represents 42% of the total US disease burden measured by US DALYs. Evidence suggests that new drug launches can significantly reduce broad measures of disability⁵⁶ and that individuals with disabilities filled an average of five times more prescriptions than those without disabilities,⁵⁷ generating sales representing 32% of the US pharmaceutical market.⁵⁷

Expedited review designations and the burden of disease

It has been suggested that expedited review programmes designed to incentivise development of products for serious diseases and unmet needs²⁶ could also incentivise product development for conditions with the greatest burden of disease.^{35 36 58} The present analysis does not support this association. These data show no significant association between the likelihood of a drug having at least one designation for expedited review and the burden of disease associated with the approved indications, whether or not drugs for neoplastic disease are included in the analysis. Particularly notable was the strong negative association between each of the expedited review designations and YLL in US and global populations. These results suggest a dissociation between programmes designed to incentive development of products that ‘… address unmet medical need[s] in the treatment of a serious or life-threatening conditions’²⁶ and the need for products that address the greatest burden of disease in US or global populations. While the large number of drugs for rare diseases in this dataset (43% had Orphan Drug Designation) could contribute to the absence of a positive association between expedited review designations and disease burden, such an analysis would require different analytical methods that do not aggregate indications into GHE disease states.

Given the importance of US markets in shaping the global portfolio of products, these results raise concerns that US policies for incentivising drug development for certain classes of serious disease through expedited review may unintentionally disincentivise product development for conditions contributing the greatest burden of human disease. To the extent that the failure of industry to address the global disease burden represents a market failure in which the anticipated economic returns from these markets are inadequate to justify investments in product development, policies that selectively reduce the timelines or costs of developing products addressing lesser disease burden may make the return on investing in products that address the greatest burden of disease even less attractive. These results further suggest that such disincentives may be particularly acute for products addressing long-term morbidities over those addressing premature death. Additional research is required to assess the impact of expedited review programmes on product development for different therapeutic areas, specific diseases or individual indications. Research should also be directed towards developing evidence-based solutions to rectify any unanticipated consequences of policies for expedited review programmes that further complicate development of products addressing the health needs of global human populations.

Limitations

First, this study considers the burden of disease at the level of GHE conditions, many of which comprise multiple indications based on their ICD-10 codes. This may lead to overestimation of the burden of disease when the drug’s indications comprise a fraction of the potential indications associated with the GHE condition. Second, the study design focuses explicitly on the first approved indication for each drug, many of which subsequently receive supplemental indications or have off-label applications. This may lead to underestimation of the burden addressed for drugs that have indications for more than one GHE condition. Third, the statistical inferences in this work relate to a population of GHE conditions; no inferences should be drawn to individual disease states or subsets. Fourth, this work looks explicitly at the burden of disease measured by WHO global burden of disease metrics and may not fully account for comorbidities or economic costs associated with disorders or their treatments. Fifth, this analysis does not assess the availability of FDA-approved drugs or the association between a product’s availability and the burden associated with its indication in any country outside the USA. Thus, no inferences can be made regarding the association between the burden of disease and new drug approvals in any one nation or region. Finally, this analysis explicitly excluded the years of the COVID-19 pandemic and does not reflect changes in practice implemented in response to the pandemic.

Conclusion

Drug approvals from 2010 to 2019 were more strongly aligned with US disease burden than global disease burden and more strongly associated with YLL than YLD. The likelihood of an approved product receiving one or more designations for expedited review was not positively associated with either the US or global burden of disease and may inadvertently disincentivise product development for diseases that contribute the most to the global disease burden. These results may inform policies to better align pharmaceutical innovation with burdens of disease.

Significance is inferred with a Bonferroni correction of 8, such that a calculated p value of 0.0065 is equivalent to a threshold of 0.05.

Data availability statement

Data are available in a public, open access repository. All data used in writing this article are publicly available from sources referenced in the text. All extracted data is provided in supplemental materials. Raw data from statistical outputs are provided in online supplemental materials.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study did not involve research on human subjects and was exempt from ethics review under the Declaration of Helsinki.