Introduction

Until 2005, multinational clinical trials in India had to be one phase behind the phase running elsewhere. Phase parity was established in 2005 [1], leading to an explosion in the number of studies running in the country [2]. The rising number of multinational trials in India was part of the broader trend of the globalization of studies, which were increasingly moving out of ‘traditional’ trial nations such as the United States (US) and Western Europe, to ‘non-traditional’ ones such as Eastern Europe, Russia and Asia [3]. Various factors, such as (i) reduced costs in the new locations; (ii) tougher regulations in the US; (iii) the availability of ‘treatment naive’ patients in the new locations; (iv) the greater speed of enrollment in the new locations; and (v) a step to obtain marketing approval in these countries in future, drove the process [4]. However the move was also criticized on multiple counts such as (i) studies being conducted for diseases that were not the most pressing in the new locations; (ii) the possibility that poverty could lead to the undue inducement of participating in trials; (iii) the questionable practice of using Western informed consent frameworks in other societies; (iv) the ethics of conducting studies in populations that would not always be able to afford the new drugs [5]; (v) lack of clarity on the experience of the trialists in the new locations [4]; and (vi) the uncertainty about the generalizability of results obtained in the new populations [4].

Unfortunately, the rise in the number of trials in India was followed by controversies such as participants not providing informed consent, or not being compensated for serious adverse events [2]. Also, there have been cases in the past where about half the recruits in international studies were planned to be recruited from India, but in fact more than 80% were [6]. All of this has led to apprehensions that people in developing countries have sometimes been exploited in trials sponsored by organizations in the developed world [6–9]. However this question has never been systematically investigated. Therefore, for multinational trials that have run in India, we set out to determine whether there was too much (60% or higher) recruitment from India.

Methods

In India, trials are registered with Clinical Trials Registry-India (CTRI) [10]. Sample studies registered with CTRI, and their URLs, are available in S1 File. On 28 January 2022, we downloaded all the trials in CTRI using a script that we developed in R programming language (available at https://osf.io/bsncw as S2 File). The script web scraped and cleaned the data, and stored them in an SQLite database (available at https://osf.io/kr9f8 as S3 File). S4 File contains the schema of the database.

Of the 39,821 downloaded trials, we examined the 26,889 that were registered with CTRI between 1 Jan 2013 to 31 Dec 2020, inclusive (S5 File). It is known that older CTRI records had more errors [11] and that is why we focused on a more recent time period.

After the processing of these trials to obtain the sub-set of interest, we identified our main outcome variable, that is participation from India. We did this by examining either the planned fraction of recruitment from India, that is ’Sample Size from India’/’Total Sample Size’, or the actual fraction of recruitment from India, that is ’Final Enrollment numbers achieved (India)’/’Final Enrollment numbers achieved (Total)’.

Since we wished to determine whether there was too much recruitment from India, we defined ‘high recruitment’ as 60% or higher recruitment from India. Here we briefly mention our reason for choosing 60% as a cut off. Trials come with risks for the participants. They also come with potential (a) immediate benefit to the participants, or (b) subsequent benefit to those of the same ethnicity in case the intervention is subsequently approved and is affordable and accessible to similar populations. In a multinational trial, the fraction of participants from India should never be more than 50%, and proportionately lower based on how many countries participate in the trial. However, given that India’s population is almost 1.4 billion people, which implies the ready availability of patients, and the fact that it ought to be easier and quicker to conduct a trial in fewer countries, we believe that recruiting up to 60% of the participants from India is ethical.

Results

The processing of these 26,889 trials is shown in Fig 1. We used the field Type of trial to identify the subset of 19,361 interventional studies and Phase to identify the subgroup of 6437 phase 2 and phase 3 trials (S5 File). Next, we used the following pair of fields to identify the multinational studies: Recruitment Status of Trial (Global) and Recruitment Status of Trial (India). So, for instance, if either field was ‘not applicable’, then the study was rejected. However, given the types of errors that are known to be present in the CTRI records [11], we also used the following two pairs of fields to identify the studies that were unambiguously multinational: (a) Date of First Enrollment (India) and Date of First Enrollment (Global) and (b) Total Sample Size and Sample Size from India. For the trial to be considered multinational, in (a) both fields needed to have values, and the total sample size had to exceed the sample size from India. Thus, a trial was classified as Indian or multinational only if the data in all three pairs of fields were consistent with the relevant classification.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The processing of the 26,889 trials registered with CTRI between 1 Jan 2013 and 31 Dec 2020.

https://doi.org/10.1371/journal.pone.0284434.g001

Based on these criteria, we obtained 5580 Indian studies, 473 multinational ones and 384 other trials (S6 File). Of the 384 other cases, 22 were entirely foreign trials; 77 were terminated either in India or elsewhere; 1 was suspended in India; 256 appeared to be Indian trials based on data in some fields, but had ambiguous information in other fields; 18 appeared to be multinational trials based on data in some fields, but had ambiguous information in other fields; and 10 trials had other kinds of discrepancies. We provide three examples from these 384 cases in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Three samples from the 384 trials that had ambiguous data, and that were excluded from this study.

https://doi.org/10.1371/journal.pone.0284434.t001

We took forward the 473 multinational studies.

We first wished to identify completed trials, and chose those for which data were available both for ‘Date of Study Completion (India)’ and ‘Date of Study Completion (Global)’. We obtained 62 studies, which we termed the Completed Set (S7 File). Next, we selected cases where both ‘Date of Study Completion (India)’ and ‘Date of Study Completion (Global)’ listed ‘Applicable only for Completed/Terminated trials’. That is, these studies were not completed. We called these 362 trials the Not Completed Set (S8 File).

We then took the following steps to estimate the date by which the Not Completed Set studies ought to have been completed and the study details updated in the CTRI records, while allowing some margin for delays.

1. Of the ‘Date of First Enrollment (India)’ and ‘Date of First Enrollment (Global)’, we considered the later date as the overall ‘Date of First Enrollment’.
2. We then examined the ‘Estimated duration’ of the trial. This value is expressed in years, months and days. Where ‘days’ had a non-zero value, we substituted it with one month. Also, we converted the years to months, so that the final ‘Estimated duration’ was expressed in months.
3. In order to allow the sponsor a margin of time to update the registry record in case the trial was delayed, (a) we doubled the number of months (and termed it ‘2x months’) for studies whose estimated duration was up to two years; and (b) increased the number of months 1.5 times (termed ‘1.5x months’) for studies whose estimated duration was 2–5 years. We did not provide such a margin for trials whose estimated duration was longer than five years.
4. We added the number of months obtained from step (iii) to the overall ‘Date of First Enrollment’ from step (i), to yield the estimated ‘Date of Study Completion’.
5. In order to give the sponsor time to update the CTRI record, we added a further 6 month margin to the estimated ‘Date of Study Completion’ for each trial, to obtain the date of study filing with the registry.

This date of study filing gives the sponsor reasonable time to complete the trial and update the CTRI record. We went on to identify how many trials in the Not Completed Set had not updated their records by this date of study filing (S9 File).

We posed three questions of the Completed Set and two questions of Not Completed Set, as described below.

62 trials of the completed set: Ratio of actual to planned percent recruitment from India

Finally, for the Completed Set, we examined the ratio of the actual percent recruitment from India to the planned percent. Cases with a ratio greater than 1.0, and cases with incorrect data, are described below, and are available in Table 2 and S7 File.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Trials for which (i) the ratio of actual to planned percent recruitment from India was above 1, or (ii) incorrect data prevented the calculation of this ratio.

https://doi.org/10.1371/journal.pone.0284434.t002

These cases fall into four groups, as follows:

1. In six cases, actual enrollment was fairly close to planned recruitment, being up to 15% higher.
2. In two cases, actual enrollment was 16–60% higher than the planned recruitment.
3. The five cases in which the ratio of actual to the planned percent exceeded 1.6 are as follows. Three cases were listed in the previous section. To recapitulate, CTRI/2014/09/004965 was for breast cancer and CTRI/2018/05/013650 for macular degeneration (both sponsored by Samsung Bioepis Co Ltd, South Korea); and CTRI/2018/02/011794 was for HIV (National Institute of Allergy and Infectious Diseases, National Institutes of Health, US). The details of the other two studies are as follows: CTRI/2015/07/005984 for myocardial injury (Hamilton Health Sciences Corporation through its Population Health Research Institute, Canada and CTRI/2020/06/026192 for COVID-19 infection (Vicore Pharma AB; a Swedish pharma company). Notably, in three of these five cases a planned fraction of recruitment from India of between 12 and 18% changed to 100%.
4. Of the four trials whose data could not be used, two were described in the previous section: CTRI/2017/12/010935 (sponsored by Amgen Inc, US) and CTRI/2018/01/011307 (AstraZeneca AB, Sweden). In these cases, the enrollment from India exceeded total enrollment. In the other two cases, CTRI/2018/02/011983 (Eli Lilly and Company India Pvt Ltd, India); and CTRI/2018/02/012091 (mAbxience, Spain), the total enrollment was listed as 0.

362 trials of the not completed set: Planned recruitment from India

We went on to consider the 362 studies in the Not Completed Set, that is cases where both the ‘Date of Study Completion (India)’ and the ‘Date of Study Completion (Global)’ listed ‘Applicable only for Completed/Terminated trials’. That is, these studies appeared not to be completed (S8 File). We determined the fraction of recruitment that was planned to be from India, and identified 25 cases (7% of 362) for which this fraction was 60% or more (Table 3 and S8 File).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Some details of the 25 ‘not completed’ trials where the planned recruitment from India was 60% or more of the total recruitment.

https://doi.org/10.1371/journal.pone.0284434.t003

Of these 25 studies, three were sponsored by Indian companies. Another one was by a multinational firm that acquired a local company that was involved with this trial. One study was sponsored by a public sector research institute set up by the Government of India. Two studies concerned conditions (pediatric tuberculous meningitis and post kala azar dermal leishmaniasis) that are more prevalent in India than in the West. Eight trials were sponsored by non-Western companies located in Brazil, Malaysia or Russia. The remaining 10 studies were sponsored by companies based in Australia, Austria, Italy, Switzerland, or the US. The latter 18 studies related to conditions that are not preferentially prevalent in India. As such 18 trials (5% of 362) had foreign sponsors that were potentially problematic.

Discussion

This work is concerned with the over-recruitment of participants from India in international trials. Undoubtedly trials conducted in India would have undergone review by Ethics Committees. However it is known that the processing of a trial proposal by an ethics committee may be perfunctory [12]. All trials in India have also received regulatory permission, if relevant. Here, too, it is known that permission may have been granted in a casual manner [13]. And, as mentioned earlier, both Indian and foreign researchers have commented on the ethical transgressions that have taken place in some trials run in India. For the 62 completed trials most of the trials with 60% or more planned or actual recruitment had foreign sponsors. Likewise, for the 362 trials that were listed as not completed, most of the cases with 60% or more planned recruitment had a foreign sponsor. Given the history of exploitation of trial participants in India, and given that all trials carry an element of risk, we believe that there is an unfair burden-benefit ratio on the Indian population if foreign sponsors, that are conducting trials in multiple countries, recruit too many patients from this country.

Before coming to an analysis of particular studies, we wish to comment on the low number of trials included in this study. It is known that trialists may register an interventional study as an observational one [11]. It is also known that records may have internal discrepancies of phase etc. As such, even though we used various filters in order to identify a well-defined subset of trials, we may have inadvertently either included a trial that we did not wish to, or excluded a valid trial. Any such error is a function of errors in the trial records. Having used the filters that we did, the number of trials came down significantly, leading to a dataset of just a few 100 trials.

We sought to determine whether, in multinational studies that had run in India, the fraction of Indian participants was actually high, or had been planned to be high. In Methods, we defined ‘High’ as 60% or higher recruitment from India, and explained the reasons thereof. In a multinational trial, where there is at least one country other than India, there ought to be a fair distribution of both the risks and potential benefits to the participants in the participating countries. As such, the fraction of participants from India should never be more than 50%. However, in the interest of each trial being conducted in a realistic time-frame, and without the extra costs of setting up sites in too many countries, we believe that recruiting 60% of the participants from India is ethical. We also found that doing a sensitivity analysis around 60% revealed a predictability that lends credence to this premise.

In this work, we have considered two sets of trials, ie (i) completed studies for which the data of final enrollment were available (Completed Set), and (ii) studies that are classified as not completed (Not Completed Set).

With regard to the Completed Set, we first identified five cases with 60% or more planned recruitment from India. Since one study (for cardiovascular risk) was sponsored by a domestic pharma company, without the participation of any foreign entity either as secondary sponsor or as ‘other source of monetary or material support’, the high planned recruitment from India was not a concern. The remaining four were concerned with uveitis, schizophrenia, COVID-19 infection and hemorrhoids. Given that these four conditions are well distributed around the world, we believe that planning to recruit such a large fraction of trial participants from India was not justified.

Next, for the Completed Set, we examined the ratio of actual to planned percent recruitment from India. Viewed from the view point of potential exploitation of participants from India, we believe that it was worth noting whether the actual recruitment was more than the planned, and if so, to what extent. In examining the extent of over-recruitment, we needed to use some arbitrary cut offs for what increase could be considered acceptable. We realize that if more patients are identified at some trial sites, who are willing to be part of the trial, then for practical reasons of seeing the trial progress in a timely manner, it is convenient to allow increased enrollment from those sites. We believed it acceptable if the actual was up to and including 15% of the planned recruitment. Only six (35% of the 17 potentially or actually problematic) trials were in this category, which is quite low. We considered it questionable if the actual was 16–60% higher than the planned recruitment, and we found it unacceptable if the actual was over 60% higher than the planned. These two categories–between 16–60% and over 60%–are arbitrary, and other researchers may prefer other cut offs. However, as noted, in three cases the recruitment from India rose from under 20% to 100%. Therefore, even changing the cut off would not alter the fact that certain ‘multinational’ trials became ‘Indian’ trials during the course of the trial. The Indian regulator, that is the Central Drugs Standard Control Organisation (CDSCO), needs to take cognizance of such increases in fraction of enrollment from India, and issue guidelines or rules on what is permissible.

The cases where we could not use the data were also problematic. In earlier publications [11, 14], we documented some of the fields for which CTRI records have missing or misleading data. To be noted, similar issues of missing data have been reported for the registry of the US, ClinicalTrials.gov, as well [15, 16]. The implications of such erroneous information vary with the field. For multinational trials, missing or out-of-date information in the fields ‘Date of Study Completion (India)’, ‘Date of Study Completion (Global)’, Total Sample Size and Sample Size from India, prevents audits of the fraction of trial participants recruited from India. Since over-recruitment from India is an issue of concern, it is important that CTRI or CDSCO, finds a way to ensure that these fields are updated in a timely manner.

With regard to the Not Completed Set trials, we first identified the 25 studies for which the planned recruitment from India was 60% or more of the total enrollment. It seemed perfectly justified that a large fraction of participants was recruited from India, in situations where (i) the sponsors were local companies (or a multinational company that had taken over a local company, and the local unit was involved in the trial), (ii) the sponsor was a local public sector research institution or (iii) the conditions being investigated were particularly prevalent in India. However, it is difficult to justify trials that were sponsored by those based in Brazil, Malaysia or Russia, since these are populous nations, and it ought to be possible to find trial participants domestically. Likewise, organizations based in Western nations ought to be able to find recruitees in many countries if the medical condition is found around the world.

Further, in the Not Completed Set studies, neither the global nor the Indian ‘Date of Study Completion’, had been updated in 362 (77% of the 473) CTRI records. Clearly, some of these were ongoing studies. However it was clear that others must have been completed, but that the CTRI records had not been updated to reflect this. This is despite that fact that the New Drugs and Clinical Trials Rule requires the status of a trial, such as whether it is ongoing or completed, to be reported to the Regulator every six months [17]. Although there do not appear to be any guidelines for updating this field at CTRI, in the US there is a requirement that the trial record at ClinicalTrials.gov be updated within 30 calendar days of the completion of a trial [18, 19]. In the interest of transparency, it would be advisable for CTRI to have a similar rule.

Here, too, we had to define arbitray margins, as outlined earlier, in order to identify trials that were most likely to have been completed. Some of the trials that we identified as ‘completed’ are likely to be false positives. There are also likely to be some false negatives. Although other researchers may have changed some of these margins, they too will have some false positives or negatives. Given the absence of ‘Date of Study Completion’ data, it is impossible to correctly identify which trials have actually been completed. Nevertheless, we believe that it is important to provide some estimate of completed trials for which completion dates have not been listed, to establish the phenomenon, and its relevance to the issue of recruitment from India.

Even after providing various types of margins for each record to be updated, we found that 128 (27%) records had not been updated. Due to the large fraction of outdated records, there was no way to determine–for these trials–whether the phenomenon of excess recruitment from India was common or not. The large fraction of outdated records is cause for concern, and CTRI needs to find ways to ensure that they are updated. Although it is known that registry staff check the records when they are first submitted [20], it is not known whether they check updates to records. Nevertheless, there must be best practices followed by the highly rated public registries such as ClinicalTrials.gov; Clinical Research Information Service, Republic of Korea; Australian New Zealand Clinical Trials Registry; Iranian Registry of Clinical Trials; and the German Clinical Trials Register [21], which CTRI could learn from to ensure updated study records.

Finally, we wish to point out that not only is over-representation from India a potential problem, but under-representation is also a problem, and we have highlighted the latter issue recently [22]. However the reasons that over- and under-representation are problematic are different. In the former situation, it is the worry of exploitation of Indian participants, and in the current work, we have focussed on this aspect. In the latter situation, it is that without adequate ethnic diversity, the trial results may not be generalizable to a larger population. Further, for many patients, drugs for rare disease can be completely unaffordable, and participation in a trial provides a way to potentially benefit from a novel therapy. So, although under-representation in multinationals is also a potential problem, we did not focus on that angle in this study.

Conclusion

In summary, for trials registered in a recent eight year period, we have identified several problematic trials for which the planned or actual recruitment from India in Phase 2 or Phase 3 multinational studies was 60% or more. We also identified trials in which the actual recruitment was significantly higher than the planned recruitment. Further, the records of several studies that were probably completed were not updated in CTRI in a timely manner.

It is not our intention to question the practice of running multinational trials in India. Nor are we inferring malafide intentions of the CTRI staff, or of CDSCO. We merely wish to point out that CDSCO should be particularly alert to the planned, or actual, over-recruitment of participants from India. Further, CTRI, alone or in collaboration with CDSCO, needs to ensure that multinational trial records are updated for the enrollment fields in particular, in a timely manner.

Supporting information