International Country-Level Trends, Factors, and Disparities in Compassionate Use Access to Unlicensed Products for Patients With Serious Medical Conditions.

Importance: Compassionate use (CU) is a treatment option for patients with serious or life-threatening medical conditions that provides access to locally unlicensed medications (generally free of charge) when all available treatment options have been exhausted and enrollment in a clinical trial is not possible. Objective: To examine the disparity in CU access observed across countries and explore the key driving factors. Design Settings and Participants: This study analyzed all Novartis CU requests (for individual/named patients and cohort programs) received between January 1, 2018, and December 31, 2020, and investigated selected country-specific factors for association with request activity. Data analysis was performed from February 2021 to February 2022. Main Outcomes and Measures: Country-specific request activity was quantified using request counts and rates per million population and examined in stratified and multivariable analyses (negative-binomial regression) for association with the following covariates: existence of local CU regulations and their public availability, clinical trial activity, population size, and gross domestic product.
Results: During the 36-month observation period, 31 711 CU requests were received from 110 countries, 23 194 (73%) of which came from only 10 high-income countries. All high-income countries combined accounted for 27 612 (87%) of all requests, while lower-middle-income and low-income countries contributed only 1021 (3%). Of all requests, 29 870 (94%) were from countries with CU regulations made publicly available on the internet, and higher request activity was demonstrated in countries conducting more clinical trials. Presence and public availability of CU regulations, population size, gross domestic product, and clinical trial activity were independently associated with the CU request activity in multivariable analysis. Conclusions and Relevance: In this cohort study analyzing Novartis CU requests over a 3-year period, existence and public availability of CU regulations and local clinical trial activity were positively associated with higher CU request rates. The analysis also identified an association between macroeconomic factors and CU request activity, despite the generally free provision of unlicensed therapeutic products. Similar analyses of other comparable experiences are needed to supplement these initial observations. Ultimately, better understanding of factors associated with CU request activity would translate into improved early access to novel lifesaving products for patients with unmet medical needs around the world. Copyright 2022 Aliu P et al. JAMA Health Forum.

Introduction

Clinical trials are the preferred channel for access to unlicensed therapeutic products. However, there are patients with serious or life-threatening diseases for whom enrollment in a clinical trial is not possible and all available treatment options have been exhausted.[1,2,3] Outside of clinical trials, access to locally unlicensed treatments requires a complex interplay between health authorities (HAs), physicians, patients, and pharmaceutical companies.[1,2,4,5,6,7,8]

Compassionate use (CU) programs for patients with serious medical conditions provide patient access to locally unlicensed treatments (generally free of charge) when all available options have been exhausted and enrollment in a clinical trial is not possible.[9] These locally unlicensed treatments may include products still in clinical development, as well as those pending registration or already licensed in some other countries. Owing to the nature of the regulatory approval process, there is typically a gap of several months to years between generation of robust efficacy and safety data in clinical trials, local marketing authorization, and subsequently local patient access. While some of this time is spent on the drug’s benefit-risk assessment by HAs, there are many other factors that can cause substantial delays in access, including price negotiations, drug supply and logistics arrangements, and other operational matters.[10,11] This situation can create an ethical dilemma in which a disease-modifying therapy shown to be effective in clinical trials is not available to patients who need it. For many patients with serious or life-threatening medical conditions, waiting for months or years until licensing of the drug in their country means unnecessary suffering or even death. The purpose of CU is to close this gap and facilitate early access.[12]

For example, onasemnogene abeparvovec-xioi (Zolgensma), a gene therapy for spinal muscular atrophy, one of the most common genetic causes of infant death if left untreated, was first approved in the US in May 2019 and received conditional marketing authorization in the European Union 1 year later. With multiple local approval processes ongoing, Novartis established global access to Zolgensma via a CU program, which has been providing the therapy to patients free of charge in countries where it is yet to be licensed.[13]

Recently, CU gained attention during the COVID-19 pandemic as the scientific community considered investigational products or repurposing of approved therapies with antiviral or immunomodulating properties.[5,14,15] These considerations were influenced by in-country regulatory access mechanisms. There are substantial differences in the CU regulatory landscape across countries, with some introducing or updating their regulations as the landscape evolves.[16,17] However, many countries around the world still have limited to no CU regulations in place.[18,19] Recent research highlighted limited patient awareness, perception of an uncertain benefit-risk profile, and operational hurdles as some of the factors influencing CU activity.[20,21]

In 2016, Novartis strengthened its CU processes and launched a publicly available company policy.[9] An online request system was introduced in December 2019 to streamline request management.[9] This end-to-end system is accessible to all physicians worldwide, offering an easy-to-complete form and the ability to submit and handle a request through any desktop or mobile device and provide supporting documents where needed (Figure 1).[9] Despite this globally accessible online system, the CU requests (for individual/named patients and cohort programs) received over the study period revealed variability in request activity across countries, highlighting the disparity in patient access to novel but locally unlicensed innovative therapies. This study examines this further and explores factors that might be associated with the disparity observed in CU access globally.

Figure 1.

Novartis Compassionate Use (CU) Pathway—A 5-Step Process

aWe consider granting access to investigational or preapproval products when all of the following criteria are met:

• The patient has a serious or life-threatening disease or condition, and no comparable or satisfactory alternative therapy is available.

• The patient is ineligible for enrollment into or unable to access ongoing clinical trials.

• The patient meets any other important medical criteria established by the medical experts working on the product development program.

• Sufficient data exist to believe the potential benefit of treatment outweighs the potential risk of the disease or condition to be treated.

• We have an adequate supply of the investigational product and providing the investigational product will not interfere with ongoing clinical trial(s) or with the overall development program.

• Such access provision is allowed as per local laws and regulations.

EC indicates ethics committee; HA, health authority.

Methods

Data

All CU requests received from any country or territory (hereafter referred to as “country”) between January 1, 2018, and December 31, 2020, were included in the analysis. The numbers of requests were evaluated and stratified by country.

Ethics committee and/or HA approval was obtained for CU for all patients according to local regulations. Written patient informed consent was obtained by the treating physician for every patient treated via CU included in this study.

Study Design

This study examined country-specific request counts and rates against selected factors that could be associated with CU request activity:

Presence of CU regulations in the country and their public availability: Data were obtained from the Cortellis database[22] and HA websites when available. In this case, public availability of a country’s regulation was taken as a proxy for public awareness.

In-country clinical trial activity: Information was obtained from the ClinicalTrials.gov website[23] with a search on phase 1 to 3 interventional trials (recruiting, enrolling by invitation, active not recruiting, and completed) for all medicines (start date on or after January 1, 2010, until December 31, 2020).

Macroeconomic factors and population: Country gross domestic product (GDP) and gross national income (GNI) per capita (categorized as low-income, lower-middle-income, upper-middle-income, and high-income economies) and population size data were retrieved from the World Bank Open Data,[24] based on the 2019 information (if not available, data were retrieved from Statista.com).

The study also examined these factors for the countries from which Novartis did not receive any requests during the study period. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for cohort studies.[25]

Statistical Analysis

Country-specific average annual CU request rates per million population were calculated as the number of CU requests received during the analysis period (2018-2020) divided by the population size in millions and the length of the analysis period in years. Distributions of request rates were summarized with means, SDs, medians, and IQRs and examined with stratification on the following variables: existence and public availability of CU regulations; number of phase 1 to 3 trials conducted in the country (January 1, 2010, through December 31, 2020); and population size, GDP, and GNI per capita. In the stratified analysis, numerical variables of population size, GDP, and number of trials were grouped in quartiles.

Association of the country-specific attributes with the request activity was examined in multivariable analysis based on a log-linear negative binomial regression model. Details of model specification are provided in the eMethods in the Supplement. Briefly, the outcome variable was defined as the request count per country during the analysis period. Covariates included existence and public availability of CU regulations, population size, GDP, and number of trials. Continuous covariates were log-transformed because of their strongly right-skewed distributions. Because GDP and population size jointly determine the GNI, the GNI categories were not included in the model separately, to avoid multicollinearity and model overfitting. Request counts predicted by the model were tabulated along with the observed request counts and prediction errors (observed minus predicted) for each country to assess model fit and evaluate variability in the request activity not explained by the model. Mean, median, and IQR of the absolute prediction error were calculated. Countries with the prediction error of greater than 30 requests and more than 3-fold in either direction (both conditions satisfied) were flagged as outliers. Data analysis was performed from February 2021 to February 2022 using SAS, version 9.4 software (SAS Institute).

Results

Analysis of CU Requests

Over the 36-month period, 31 711 CU requests were processed from 110 countries (6879 requests in 2018, 13 390 in 2019, and 11 442 in 2020). The top 30 countries with the highest request rates per million population are shown in Table 1 (full distribution is provided in the eTable 1 in the Supplement). Requests were spread among several therapeutic areas (Figure 2), with oncology conditions (12 782 requests) representing 40% of the total request activity. The top 5 compounds during the study period were erenumab (5557 requests [18%]), clofazimine (4454 [14%]), alpelisib (3277 [10%]), ruxolitinib (2810 [9%]), and dabrafenib/trametinib (1742 [5%]), with the top 5 indications being chronic migraine (5264 [17%]), mycobacteria infection (3647 [12%]), metastatic breast cancer (2692 [8%]), COVID-19 (2185 [7%]), and multiple sclerosis (899 [3%]). Table 2 shows the number of CU requests and countries in relation to existence and public availability of regulations, clinical trial activity, population, GDP, and GNI per capita.

Table 1.

Number of Compassionate Use Requests in 2018, 2019, and 2020 for the First 30 Countries With the Highest Request Rates per Million Population in Decreasing Order

Country or territory	Request year			All	Request status at data lock, %				Rate per millionb
	2018	2019	2020		Approved	Canceleda	Pending	Denied
Belgium	358	2426	521	3305	93.5	3.2	2.3	1.0	95.9
Ireland	232	346	39	617	89.3	6.3	4.4	0.0	41.7
Australia	1238	985	577	2800	92.4	6.0	1.5	0.2	36.8
Israel	133	243	224	600	75.3	16.7	7.8	0.2	22.1
Croatia	66	121	79	266	86.8	6.0	4.5	2.6	21.8
Netherlands	145	787	149	1081	90.5	6.9	2.4	0.2	20.8
Canada	684	636	767	2087	83.2	12.1	4.0	0.7	18.5
Slovenia	21	46	32	99	79.8	17.2	3.0	0.0	15.8
Italy	285	659	1804	2748	85.7	9.7	3.8	0.8	15.3
Andorra	0	3	0	3	0.0	66.7	33.3	0.0	13.0
France	216	1238	923	2377	96.8	2.1	0.7	0.5	11.8
Switzerland	17	90	150	257	84.8	6.6	7.8	0.8	10.0
Poland	215	449	424	1088	77.8	12.4	9.5	0.3	9.6
Malta	0	6	8	14	85.7	7.1	7.1	0.0	9.3
Spain	264	689	355	1308	85.5	10.9	2.8	0.8	9.3
Taiwan	67	147	383	597	73.7	13.4	10.9	2.0	8.4
Slovak Republic	14	101	21	136	85.3	6.6	5.9	2.2	8.3
Hong Kong	22	60	69	151	80.8	9.3	9.3	0.7	6.7
Cyprus	5	4	15	24	87.5	8.3	0.0	4.2	6.7
Greece	8	63	137	208	89.9	3.4	6.3	0.5	6.5
Singapore	14	50	41	105	73.3	12.4	13.3	1.0	6.1
Luxembourg	0	5	6	11	90.9	0.0	9.1	0.0	5.9
US	1671	2164	1504	5339	87.5	8.2	2.1	2.2	5.4
United Kingdom	184	590	287	1061	81.0	9.3	9.3	0.4	5.3
Austria	12	48	76	136	92.6	3.7	3.7	0.0	5.1
Namibia	19	13	6	38	76.3	18.4	0.0	5.3	5.1
Lebanon	22	30	49	101	74.3	11.9	10.9	3.0	4.9
Lithuania	0	13	26	39	94.9	0.0	5.1	0.0	4.7
Costa Rica	5	25	27	57	80.7	15.8	1.8	1.8	3.8
Iceland	2	0	2	4	75.0	25.0	0.0	0.0	3.7

Request status “canceled” means that the request was canceled by the requesting physician.

Average annual request rate per million = (total request count in 3 years/population size in millions)/(3 years).

Figure 2.

Distribution of Compassionate Use Requests Received per Therapeutic Area

aOncology includes hematology (n = 3831) and solid tumor (n = 8951) requests.

Data set from January 1, 2018, to December 31, 2020. Total requests = 31 711, including 1 request with therapeutic area unassigned (1 non-Novartis product).

Table 2.

Stratified Analysis of Compassionate Use Requests per Select Country Attributes

Variable	Level	Countries or territories, No.	Requestsa	Request rate per million
			No. (%)	Mean (SD)	Median (IQR)
Regulations	Public	54	29 870 (94.2)	8.0 (15.1)	2.2 (1.1-9.3)
	Nonpublic	23	1320 (4.2)	1.3 (2.3)	0.2 (0.0-1.6)
	None	122	521 (1.6)	0.3 (1.0)	0.0 (0.0-0.0)
Population size quartiles	Upper	50	20 033 (63.2)	1.8 (4.1)	0.1 (0.0-1.1)
	Upper-middle	50	9768 (30.8)	4.5 (14.8)	0.0 (0.0-1.9)
	Lower-middle	50	1838 (5.8)	2.7 (6.9)	0.1 (0.0-2.2)
	Lower	49	72 (0.2)	1.0 (2.6)	0.0 (0.0-0.5)
GDP quartiles	Upper	50	29 950 (94.4)	7.1 (15.6)	1.4 (0.3-6.7)
	Upper-middle	50	1601 (5.0)	1.8 (4.1)	0.1 (0.0-1.6)
	Lower-middle	50	153 (0.5)	0.7 (1.8)	0.0 (0.0-0.2)
	Lower	49	7 (<0.1)	0.4 (1.9)	0.0 (0.0-0.0)
GNI per capita categories	High	65	27 612 (87.1)	7.1 (13.9)	2.1 (0.3-8.3)
	Upper-middle	56	3078 (9.7)	0.6 (1.1)	0.1 (0.0-0.9)
	Lower-middle	53	1014 (3.2)	0.1 (0.4)	0.0 (0.0-0.0)
	Low	25	7 (<0.1)	0.0 (0.0)	0.0 (0.0-0.0)
No. of trials, by quartile	Upper	50	30 272 (95.5)	7.7 (15.5)	2.1 (1.1-8.3)
	Upper-middle	50	1294 (4.1)	1.5 (3.9)	0.1 (0.0-1.0)
	Lower-middle	50	80 (0.3)	0.2 (1.0)	0.0 (0.0-0.0)
	Lower	49	65 (0.2)	0.7 (2.4)	0.0 (0.0-0.0)

Abbreviations: GDP, gross domestic product; GNI, gross national income.

Total requests = 31 711.

Existence and Public Availability of CU Regulations

A total of 77 countries were observed to have CU regulations, 54 of which had the regulations available on their government or HA websites (Table 2). These countries contributed the overwhelming majority (29 870 of 31 711 [94%]) of CU requests (Table 2). Among the countries from which Novartis did not receive any CU requests during the study period, only 6 (7%) had CU regulations in place (Algeria, Bulgaria, Congo, Cuba, Guinea, and Latvia).

Clinical Trial Activity

A positive association between in-country clinical trial activity and CU request activity was observed in stratified analysis (Table 2). Concerning the countries from which Novartis did not receive any CU requests during the study period, only 5 (6%) had each conducted more than 100 clinical studies (January 1, 2010, until December 31, 2020: Bulgaria, 1258 studies; Puerto Rico, 843; Latvia, 435; Moldova, 159; Bangladesh, 110), with an overall country median of 2 trials conducted (data not shown).

Macroeconomic Factors and Population Size

More CU requests were received from countries with a high GDP compared with low- and middle-income countries (Table 2). Of all requests, 23 194 (73%) came from only 10 high-income countries (US, 5339; Belgium, 3305; Australia, 2800; Italy, 2748; France, 2377; Canada, 2087; Spain, 1308; Poland, 1088; the Netherlands, 1081; United Kingdom, 1061). Overall, high-income countries accounted for 27 612 (87%) requests, while lower-middle-income and low-income countries contributed only 1021 (3%) (Table 2). Larger population size was not associated with higher request activity in the upper 2 quartiles of population size, which accounted for the majority of requests (Table 2) because the highest request rates per million population were observed in a few high-income countries with relatively small population sizes (Table 1). Some countries with very large population (>200 million) showed very low request numbers (eg, China, 5 requests; Nigeria, 5; Indonesia, 10; Brazil, 427; India, 504), even though the majority had CU regulations in place, albeit available publicly only in Brazil and India.

Multivariable Regression Analysis

In the multivariable analysis, request activity was positively associated with presence of CU regulations (public or nonpublic vs none), higher GDP, and more clinical trials, and was negatively associated with population size (Table 3). Request counts predicted by the model based on the country attributes (regulations, population size, GDP, and clinical trials) were tabulated along with the observed request counts and prediction errors for each country in eTable 2 in the Supplement. Mean absolute prediction error was 207.8 (first quartile = 0.5, median = 3.4, third quartile = 54.4). The mean was much larger than the median because for several countries, the observed request count was considerably different from what was predicted by the model based on the country attributes. The top 5 countries flagged as outliers with the largest absolute differences in the observed vs predicted request counts were the US (5339 observed vs 19 100 predicted), China (5 observed vs 4371 predicted), Japan (95 observed vs 3279 predicted), Germany (272 observed vs 3164 predicted), and Belgium (3305 observed vs 587 predicted). While the US contributed the largest number of CU requests among all countries, the observed request count for the US was still less than that expected by the model based on the country’s attributes, including its exceptionally large GDP and very high clinical trial activity. It is also worth noting that despite the large absolute number of requests, the request rate per million population in the US was considerably lower than that of many other countries (Table 1).

Table 3.

Multivariable Analysis

Covariate	β̂ (95% CI)	Rate ratioa (95% CI)
Regulations
Public vs none	0.76 (0.52 to 0.97)	2.13 (1.69 to 2.64)
Nonpublic vs none	0.72 (0.55 to 0.90)	2.06 (1.74 to 2.47)
Log2(population)	−0.28 (−0.34 to −0.23)	0.75 (0.71 to 0.79)
Log2(GDP)	0.83 (0.76 to 0.92)	2.30 (2.14 to 2.51)
Log2(No. of trials)	0.13 (0.09 to 0.18)	1.14 (1.09 to 1.20)

Abbreviation: GDP, gross domestic product.

Rate ratio = request rate ratio per 1-unit increase in the value of the corresponding covariate, holding the values of other covariates constant. A 1-unit increase on the log2 scale corresponds to doubling of the covariate value. Model intercept: β̂0 = −2.4952.

Discussion

To our knowledge, this is the first study on a large data set of CU requests to explore association between selected factors and the CU request activity at the country level. Existence and public availability of CU regulations, local clinical trial activity, and economic development were positively associated with CU request activity and may have significant influence on patient access. The CU channel continues to expand; year-on-year, there have been first-time requests from new countries. This expansion is particularly important to ensure that innovative medicines reach patients earlier in the context of slow regulatory approval processes as well as widening gaps between market authorization and patient access in many countries.[10,11]

In this study, existence and public availability of the CU regulations was associated with a higher request activity. Indeed, when patients and health care professionals can easily obtain information related to CU regulations, the likelihood of this access pathway being used appears to be higher than in settings with no regulations.

Compassionate use has evolved to become a complex enterprise involving several stakeholders,[1,2,4,5,6,7,8] including pharmaceutical companies, HAs, physicians, patients, caregivers, and patient advocacy groups, as well as, increasingly, social media.[26] Two specific issues that have been highlighted in the context of CU are the challenges for HAs to monitor usage of locally unlicensed products and for physicians and patients to obtain information about the products and their availability.[27] Examples of local efforts to increase awareness and facilitate this access pathway include the US 21st Century Cures Act,[28] the Reagan-Udall Foundation Navigator,[29] the health care professional portal introduced by the French National Agency for the Safety of Medicines and Health Products (ANSM),[30] and the US Food and Drug Administration’s Project Facilitate.[31]

Stakeholders will need to work hand in hand to bring further improvements to the regulatory CU landscape globally.[1,2,4,5,6,7,8] The existence of CU regulations and policies at a country level, as well as the need to make these available transparently to the public, are important factors to facilitate patient access.[18,19,32,33,34] In this study, more CU requests were received from countries with higher clinical trial activity. Clinical research influences patient outcomes in a health care system, and its subsequent benefits include improvements in infrastructure and health care process strengthening overall.[35,36,37] Clinical trial investigators, having gained experience with a treatment during a trial, are more likely to request the same therapy via CU for patients ineligible for enrollment or unable to join the trial. Conversely, CU experience can contribute to the overall evidence base and support the design of future clinical trials.[38] In addition, with utilization of treatments outside of controlled trial settings, data collected would be complementary to the trial data, informing future real-world use.[39] However, CU regulations in some countries exclude patients who are unable to fulfill the eligibility criteria for clinical trials, thereby restricting them from early access to treatments.[17,32]

In some instances, more patients have received unlicensed therapeutic products via CU than in clinical trials. For example, historically with the imatinib CU program, 7380 patients in 34 countries were able to access imatinib via CU prior to local approval compared with 1027 patients in 6 countries in the 3 multicenter studies used as the basis for regulatory filing.[40]

Globally, low- and middle-income countries display the highest burden of diseases. Clinical research is essential to identify and address a population’s unmet medical needs, and therefore more research should be performed in resource-limited settings according to the principles of good clinical practice.[41] Clinical trials are generally conducted in a limited number of countries, resulting in an overall patient enrollment not necessarily representative of the global patient population for that specific disease. This health inequity has been highlighted recently and is an important concern because people of different races and ethnicities may react differently to certain medical products.[42,43]

Countries from which Novartis received the highest absolute numbers of CU requests all had a high-level GDP.[24] Similarly, most countries in the upper quartile of request rates per million population were high-income economies per their GNI per capita.[24] Interestingly, a parallel can be drawn with countries that have orphan drug policies. Countries/areas with orphan drug policies are statistically wealthier, and these policies have been established in only approximately 20% of the low-income countries, indicating a considerable policy gap.[44]

Looking at the lower-middle-income and low-income countries in the data set, the majority of the requests were for novel products with ongoing development programs. Patient access to unlicensed therapeutic products via CU, in particular for novel or personalized medicines, may be influenced by the overall country’s health care system capabilities, which are constrained or limited in many low-income and lower-middle-income countries.[45] Nevertheless, personalized medicines have already been effectively used and monitored successfully in patients in these countries.[46]

The results suggest an association between economic development and CU request activity. This reinforces the hypothesis that economic growth is a factor associated with access to unlicensed therapeutic products via CU. Low-income countries might lack the necessary medical organizational infrastructure and facilities to ensure CU access awareness and process efficiency. Although the provision of therapeutic products is generally free of charge in CU, ancillary costs may have to be borne by the patient, health care professionals, or third-party payers, depending on the local circumstances. These additional costs may affect CU activity, especially in countries where health care costs are paid out of pocket.[47] Overall, there is a need for the active involvement and collaboration of several stakeholders, including regulators, industry, physicians, and patients, for the CU access process to better address patients’ needs and to specifically support countries in building up their CU access capabilities.

Limitations

This analysis was limited to the experience of a single company, and results may differ in CU programs offered by other organizations. Another limitation of this study is that it did not include an in-depth assessment of the individual country regulations. Further work is required to better understand the different components in the various local regulations to provide a framework for consideration for countries looking to update or put in place CU regulations. In addition, this study did not assess the time interval between availability of trial results, local product approval, and/or reimbursement, which could play a role in CU level activity. Finally, as with any observational study, associations observed in the analyses may not represent true causal effects owing to residual confounding by various factors not captured in the study.

Conclusions

In this cohort study analyzing Novartis CU requests over a 3-year period, existence and public availability of CU regulations and local clinical trial activity were positively associated with higher CU request rates. The analysis also identified an association between macroeconomic factors and CU request activity, despite the generally free provision of unlicensed therapeutic products.

Mechanisms for CU will continue to evolve and be used by more patients, physicians, and regulators. Similar analyses of other comparable experiences are needed to supplement these initial observations. Ultimately, better understanding of factors associated with CU request activity would translate into improved early access to novel lifesaving products for patients with unmet medical needs around the world.