Ethnic representation in interventional clinical trials run in India

Summary

Globally, the need to enhance the diversity of trial participants is receiving increasingly urgent attention. We wanted to know whether trials run in India had adequately sampled the country's enormous ethnic diversity. We accessed the Clinical Trials Registry-India website to determine whether each interventional drug or biologic Phase 2 or 3 study, registered in a recent five-year period had run in each of six geographic zones. As regards Phase 3 trials conducted only in India, 61.4% ran in a single zone and just 6.8% were conducted in all six zones. Multinational Phase 3 trials had a better distribution since 3.6% had run in just one zone and 7.1% in all six. India's diverse ethnic groups are underrepresented in the majority of trials covered in this study. A trial that is conducted on non-representative groups and later discovered to be harmful or ineffective in parts of the population, is unethical. We propose various remedial steps.

Introduction

Several categories of individuals tend to be underrepresented in clinical trials. These include children, sexual minorities, etc. This has real consequences for medical care.¹ In the United States (US), guidelines, legislation and enrollment tactics have attempted to remedy the situation^1,2 and trials have become more inclusive over the past 25 years.³

Here we focus on ethnic diversity. The International Conference on Harmonization issued the ICH E5 guidelines in 1998, that described intrinsic (such as genetic or physiologic) and extrinsic causes (such as cultural practices) for ethnic differences between populations.⁴ In order to ensure that drugs developed in the West could be sold in Asia, for instance, without repeating a large trial in the new region, bridging studies were designed that would test for safety, efficacy, suitable dose etc. in the new region. The need to include various ethnic groups in clinical research has been emphasized in India as well. Both the Parliamentary Standing Committee on Health and Family Welfare in 2012,⁵ and the Ranjit Roy Chaudhury Committee in 2013⁶ mentioned the need to include different ethnicities in trials. In this study, we accessed the Clinical Trials Registry - India (CTRI) and investigated whether interventional drug and biological Phase 2 and Phase 3 studies had run in the six zones of the country, ie., Eastern, Western, Northern, Southern, Central, and North-Eastern. We also examined which zones, and which cities in each zone, had hosted the maximum number of trials.

We accessed the CTRI website (http://ctri.nic.in/Clinicaltrials/advancesearchmain.php) on 15 May 2021. We went on to process the data as shown in Fig. 1. Further details of the methodology are available in Appendix 1 File (https://osf.io/jxpkm), which refers to Appendix 2–5 Files, and Appendix 1–8 Tables.

Fig. 1.

Process of identifying the interventional drug and biological Phase 2 or Phase 3 trials that ran in India in a recent 5-year period.

Our dataset comprised interventional drug or biological Phase 2 or Phase 3 trials for which enrollment was complete, and that had run in India in a recent five-year period. Phase 3 trials included Phase 2/3 trials. Some results are presented briefly below. Data related to the Indian trials (ITs) and the Multinational trials (MTs) are presented in greater detail in Appendix 5 Table (https://osf.io/eyv3a) and Appendix 6 Table (https://osf.io/bwxaf) respectively.

The 163 Phase 3 ITs had 1049 sites, and the 28 Phase 3 MTs had 349 sites. The number (and percentage) of trials with sites in different numbers of zones is presented in Table 1.

Table 1.

For the Indian and multinational trials in Phase 3, the number of zones in which each trial had run.

Number of zones	163 Indian trials		28 multinational trials
	Number of trials	Percentage	Number of trials	Percentage
1	100	61.4	1	3.6
2	3	1.8	3	10.7
3	9	5.5	8	28.6
4	13	8.0	7	25.0
5	27	16.6	7	25.0
6	11	6.8	2	7.1
Total	163	100	28	100

For both the ITs and MTs, the Western and Southern zones have the most sites, and the Eastern, Central and North-Eastern, the least, as shown in Table 2.

Table 2.

For the Indian and multinational trials in Phase 3, the number and percentage of sites in each zone.

Trials in each zone	163 Indian trials		28 multinational trials
	Number of sites	Percentage	Number of sites	Percentage
Western	362	34.5	110	31.5
Southern	254	24.2	132	37.8
Northern	195	18.6	52	14.9
Central	113	10.8	22	6.3
Eastern	107	10.2	31	8.9
North-Eastern	18	1.7	2	0.6
Total	1049	100	349	100

For both sets of trials, 1–4 cities accounted for 50% or more of the trials in a given zone, as shown in Table 3.

Table 3.

For the Indian and multinational trials in Phase 3, the top cities that accounted for 50% or more of the trials in a given zone.

Zone	Top cities for 163 Indian trials	Top cities for 28 multinational trials
Central	Lucknow (48.7%), Kanpur Nagar (20.4%)	Lucknow (31.8%), Indore (22.7%)
Eastern	Kolkata (57.9%)	Kolkata (64.5%)
North-Eastern	Kamrup (83.3%)	Kamrup (50%), Jorhat (50%)
Northern	Delhi (37.9%), Jaipur (31.3%)	Delhi (48.1%), Jaipur (19.2%)
Southern	Hyderabad (21.3%), Bangalore (21.3%), Srikakulam (8.3%)	Hyderabad (19.7%), Bangalore (15.9%), Chennai (10.6%), Coimbatore (6.8%)
Western	Ahmedabad (19.9%), Pune (16.6%), Mumbai (15.5%)	Pune (25.5%), Ahmedabad (19.1%), Mumbai (18.2%)

Also, we have analyzed Census 2011 data and found that the Scheduled Tribe population was 10% or more of the total population in 238 (37.2%) of the 640 districts in India. The distribution of such districts per zone is presented in Table 4.

Table 4.

Per zone, the number of districts with a Scheduled Tribe population that is 10% or more of the total population, and the fraction per zone.

Zone	Number of districts	Percentage
North-Eastern	69	29.0
Central	53	22.3
Eastern	44	18.5
Northern	28	11.8
Western	25	10.5
Southern	17	7.1
Islands	2	0.8
Total	238	100

We see that the fraction of tribal population was higher in zones with the fewest trials, ie North-Eastern, Central and Eastern.

Finally, we characterized each (a) Phase 2 and Phase 3 and (b) Phase 3 trial from the viewpoint of type of study, intervention, year of registration, the health condition investigated, the sponsor, and the PI. Details are available in Appendix 7 Table (https://osf.io/p7vkw) and Appendix 8 Table (https://osf.io/2ghnm) for the ITs and MTs respectively, and are discussed in Appendix 6 File (https://osf.io/r7mhp).

Discussion

Why should studies sample a range of ethnicities? Around 20% of the drugs approved in recent years in the US have had differential outcomes based on ethnicity.⁷ The difference in outcomes between different groups may be in the efficacy of the drug, or in the incidence of side effects, or their nature and severity.⁸ Some of these differential outcomes are due to genetic factors, and others to lifestyle.⁹ For these reasons, and due to the growing number of Asians living in Western nations, the regulators in those countries also seek wider ethnic representation in trials, and welcome data from the Asia–Pacific region, for example.¹⁰

The Government of India's Indian Council of Medical Research has indicated that adequate sampling of ethnicities requires enrollment from six zones of the country.¹¹ Therefore, we set out to examine whether each Phase 2 or Phase 3 interventional drug or biologic study had run in each of these six zones. Although each zone had a mix of ethnicities, the dominant ethnicity per zone varied.¹² In this study, we assumed that each zone had a different ethnicity from the others. As such, this was a top-level assessment of whether each study had sampled a broad enough set of ethnicities.

For the Phase 3 ITs, it is astounding that 61.4% ran in a single zone, and a minuscule 6.8% ran in all six zones. Although the MTs did better, both were a far cry from the ideal situation, where each study would have run in all six zones. For both the ITs and the MTs, the Southern and Western zones have the most sites, and the Eastern, Central and North-Eastern, the least. Whereas some of these data are understandable – the North-Eastern region has fewer medical facilities than big metropolises such as Bangalore and Mumbai,¹³ and is also known to have poor connectivity¹⁴ – some of them are not as clear.

Going beyond the number of zones sampled, we would have liked to determine the fraction of participation from each zone in a given study. However, CTRI provided the enrollment per study and not per site, and so we were unable to pursue that angle.

Within each zone, the trials cluster in certain cities, but it is again not always clear why these particular cities were popular. Some cities are easily understood: Delhi, Mumbai, Hyderabad and Bangalore are large cities with a large number of hospitals. However, for example, Indore (Central zone), Jaipur (Northern zone) and Kamrup (North-Eastern zone) could probably have been substituted by Bhopal, Chandigarh and Guwahati respectively. It is known¹⁴ that there are particular criteria that sponsors use to identify suitable sites in India. However, it is not known why particular cities were not popular amongst sponsors.

It is concerning that most interventional drug or biologic trials did not adequately sample the very wide range of ethnicities in India, even at a top-level consideration of these ethnicities. In large trials in particular, it is important to have wider ethnic representation so that the possibility of different treatment effects in different ethnic groups can be explored. Providing patients with ineffective drugs, exposing them to unnecessary toxicity, denying particular groups of patients the opportunity to benefit from a trial and post-trial access, and wasting the efforts of volunteers are all ethical transgressions. We realize that increasing the ethnic diversity of trials may increase trial costs. Whether such increased costs would lead to fewer treatments being developed is a question that would need separate investigation. Further, we should also bear in mind there are costs to developing drugs that have safety or efficacy concerns in parts of the population.

Although both the 59th Parliamentary Standing Committee Report and the Ranjit Roy Chaudhury Committee report emphasized the need to sample a wide variety of ethnicities, this is not mandated by the Government of India's New Drugs and Clinical Trial Rules, 2019.¹⁵ Also, unlike many other registries, and in contradiction to WHO recommendations, CTRI does not have a field for the reporting of trial results.¹⁶ Since 2017, the US Food and Drug Administration (USFDA) has required that the race and ethnicity of each participant, if captured, needs to be submitted. This led to a big upsurge in the reporting of such data in registry records.¹⁷ As such, if the government required it, this data could also be captured by CTRI. However, to its credit, and consistent with WHO requirements, CTRI does have a field asking for a data sharing plan for Individual Participant Data (IPD). Although stating a willingness to share IPD does not always translate to sharing this data,¹⁸ if sponsors do share IPD, some ethnicity data may become available.

In earlier work we have made various recommendations that would make CTRI world class.¹⁶ For instance, the managers of CTRI could easily include a ‘Results’ field in a trial record. These results could require disaggregated data based on ethnicity and other demographic indicators. The Indian government could also make explicit its expectations in terms of wider ethnic representation—and reporting—in Phase 3 trials in particular. In the medium term, the Indian regulator should—like the USFDA—make public the regulatory documents pertinent to a drug's approval.¹⁶ These would contain summary demographic details of those who participated in the trial(s). Furthermore, such documents need to be searchable and user-friendly. In the medium to long term, under-served regions of the country need to have more medical facilities. Whereas the primary reason for such facilities would be regular medical care, they could also be used to improve the diversity of ethnic representation in trials.

Finally, this study had certain limitations. First, we only examined whether each trial had run in six zones of the country. Ideally, one would have liked to determine whether it had sampled ethnicities at much greater granularity. However, such information is unavailable in the public domain and may not exist at all. Second, due to the many categories of Type of Study, we probably missed some of the relevant drug or biologic studies. We do not know whether our results are mirrored by the trials that were left out. And third, we examined records that were registered only in a recent five-year period, and do not know whether earlier records would have similar results.

Conclusion

In this study, we examined trials registered with CTRI in a recent 5-year period to investigate whether each Phase 2 or Phase 3 interventional drug or biologic trial had run in the six zones of the country, thereby sampling the diverse ethnicities of India. Most had not. We have suggested various steps that could be taken to improve the situation. Also, similar analyses probably need to be carried out in other countries in which particular regions are dominated by specific ethnicities, if wide ethnic representation in each trial has not been ensured by other means.

Contributors

GS – conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, validation, visualization, writing - original draft preparation, writing - review & editing. JM – Data curation, investigation, methodology, software, visualization, writing - review & editing. IC – data curation, formal analysis, investigation, methodology, software, validation, writing - review & editing. MP – investigation, software, writing - review & editing. MK - formal analysis, investigation, methodology, writing - review & editing. SV – formal analysis, writing - review & editing. RV – formal analysis, methodology, validation, writing - review & editing. All authors (a) contributed to, and approved, the final version of the manuscript, (b) had access to all the data in the study and (c) had final responsibility for the decision to submit for publication.

Data sharing statement

https://osf.io/4pdsy/.

Declaration of interests

We declare no competing interests.