Under-representation of older Indian persons with cancer in clinical trials

Introduction

The government of India has defined an ‘older person’ as someone aged 60 years and over, as reflected by the retirement age,1 and various programmes such as the National Program for Health Care of the Elderly2 and the National Pension Scheme.3 Other institutions in India as well as the National Cancer Registry Programme also follow this age cut-off of 60 years.4–8 This lower age cut-off used in India to define the geriatric age group as compared with that used in Western countries like the USA (65 years, as reflected by the American Society of Clinical Oncology guidelines)9 and Europe (70 years, as reflected by the European Society of Medical Oncology guidelines)10 is primarily based on the life expectancy of the population. As of 2024, the estimated life expectancy in India is 72.03 years, whereas it is 79.74 years in the USA, and 84.38 years in Switzerland.11 The life expectancy in other Southeast Asian countries is similar to that in India, and they, therefore, follow the same age cut-off to define older individuals.

India’s current projected population is 1.44 billion,12 with approximately 10% of the population aged over 60 years.13 As per data from the National Cancer Registry Programme of the Indian Council of Medical Research, 46.5% of the cancer cases registered with the 28 population-based cancer registries and 35.4% of the cases registered with the 98 hospital-based cancer registries were aged 60 years and over; globally, the proportion of older adults with cancer has been reported to be 66.3%.7 Thus, although approximately 10% of India’s population is in the geriatric age group, there is a disproportionately high number of older persons with cancer ranging between 35% and 46%. This group of older persons with cancer has traditionally been neglected in terms of both evaluation and therapy.14 15

Clinical trials serve as the cornerstone of evidence-based medicine. Our patient management decisions are based on the results of these clinical trials.16 It is imperative that the patients we routinely see in the clinic are adequately represented in these trials. Additionally, enrolment in a clinical trial is a mechanism for patients to gain access to new and potentially lifesaving medications, which may not be widely available. Unfortunately, some subsets of patients have been less represented than others in clinical trials. Racial minorities and women have been found to be under-represented in clinical trials.17 Similar under-representation has been noted for older persons with cancer. Various studies from the USA and Canada have reported that the representation of older persons with cancer in clinical trials ranges from 22% to 41%.18–24

High-income countries dominate the clinical research arena. An analysis of Cochrane reviews and randomised trials (797 reviews of 12 340 trials in 10 937 306 participants) done in non-communicable diseases (cardiovascular disorders, cancers, diabetes and chronic respiratory diseases) reported that 90% of clinical trials and over 80% of participants were from high-income countries. India, classified by the World Bank as a lower-middle-income country (LMIC),25 along with other LMICs contributed 4.95% of trials (n=438) and 11.68% of participants.26 Added to this dearth of clinical trials in LMICs, there is a complete lack of data on the representation of older persons with cancer in clinical trials in LMICs. We, therefore, performed this study to understand whether older patients with cancer have been adequately represented in clinical trials at our centre and whether we could identify any trend in enrolment over time, as well as any factors that impacted the enrolment of older Indian persons with cancer in interventional clinical trials.

Methods

Results

We included the data of 21 894 patients enrolled in 150 interventional clinical trials that were conducted between September 2003 and April 2023. None of the studies had been conducted exclusively in older adults, and none had been conducted in the geriatric oncology clinic. Most of the multicentric studies were pharma-sponsored global studies. The details of the trials included are provided in table 1. The median age of the participants was 51 years (IQR 43–59). There were 5132 patients (23.4%) who were ≥60 years, 2678 (12.2%) aged ≥65 years and 1045 (4.8%) ≥70 years. Between 2003 and 2022 (excluding 5 years for which data were not available in the hospital registry, ie, 2009–2011 and 2015–2016) there were 323 356 adult patients (aged ≥18 years) registered at our institution. Of these, 97 002 (30%) were aged ≥60 years (figure 1).

Request permission

The distribution of older patients with cancer according to the primary tumour, both overall and in those enrolled in clinical trials is depicted in online supplemental figure 1. The greatest differences between the proportion of patients enrolled in clinical trials as compared with their occurrence in the hospital registry were for oesophagogastric and gynaecological cancers. Interestingly, there was an excess enrolment of older adults in the uro-oncological trials, that is, prostate, bladder and kidney cancers as compared with the proportion in the hospital registry for those tumours.

In 38 trials (25.3%), the eligibility criteria specified an upper age limit, which was set at 60 years in one trial (0.7%), 65 years (8%) in 12, 70 years in 14 (9.3%), 75 years in seven (4.7%) and 80 years in four trials (2.7%). Eastern Cooperative Oncology Group performance status of 2 was an exclusion criterion in 38 (25.3%) trials while 95 (63.3%) trials excluded patients with uncontrolled comorbidities. In two trials (1.3%), illiteracy was an exclusion criterion. The various trials at our institute were planned and written by different investigators at different time points and for different malignancies. The eligibility criteria were decided either by the individual investigator in the case of investigator-initiated trials or by a central steering committee in the case of sponsored trials. The actual eligibility criteria, including the decision regarding an upper age limit for enrolment, would have depended on multiple factors including the primary malignancy, stage of the disease, type of study, therapy planned, the disease management group, department, ethics committee and various other factors, including the discretion and overall philosophy of the planning committee. This is probably why different age cut-offs for the upper age limit were decided for various trials.

The proportion of older persons with cancer enrolled in various types of studies, based on different factors, is depicted in figure 2. Of the 150 trials included, there were 87 trials (58%) conducted in 10 497 patients (47.9%) on systemic therapy. We broadly divided systemic therapy into oral (hormonal therapy, oral targeted therapy, oral chemotherapy; 17 trials (20%)), intravenous cytotoxic therapy; 57 trials (66%) and intravenous non-cytotoxic therapy (immunotherapy, targeted therapy; 13 trials (15%)). The most toxic therapy being administered was used for classification. For example, a randomised trial in which oral metronomic chemotherapy was compared with intravenous chemotherapy was coded as an intravenous chemotherapy trial (as enrolled patients would have to be fit for intravenous cytotoxic chemotherapy). Similarly, studies in which oral medications were added to standard intravenous chemotherapy were categorised as intravenous cytotoxic chemotherapy studies. There was a significantly higher proportion of patients aged ≥60 years in trials investigating oral therapy (339 of a total of 1101; 34%), compared with intravenous cytotoxic chemotherapy (1974 of a total of 8355; 24%) and intravenous non-cytotoxic anticancer therapy (263 of 1131; 23%); p<0.001 (figure 3I).

Request permission

Request permission

On univariate logistic regression (figure 3), the factors associated with significantly lower odds of enrolling older patients with cancer were age-specific eligibility criteria (OR 0.45; 95% CI 0.42 to 0.49; p<0.001), female sex of the participant (OR 0.59; 95% CI 0.55 to 0.63; p<0.001), trials conducted in the curative setting (OR 0.63; 95% CI 0.58 to 0.68; p<0.001), trials across various treatment intents (eg, supportive care) (OR 0.73; 95% CI 0.64 to 0.83; p<0.001), investigator-initiated trials (OR 0.86; 95% CI 0.75 to 0.99; p=0.033), single-centre studies (OR 0.43; 95% CI 0.38 to 0.47; p<0.001) and the absence of quality of life (QoL) as a trial endpoint (OR 0.59; 95% CI 0.55 to 0.63; p<0.001).

To assess the time trend of recruitment of older Indian patients with cancer in interventional clinical trials (figure 4), we included 9807 patients (44.8%) who had been recruited in 65 trials (43.3%) between 2003 and 2019, as we had the enrolment dates for these. Although the proportion of older patients with cancer registered at our hospital remained relatively stable at 28%–29% throughout the course of the study, there was a steady increase in the proportion of older patients recruited, from 8% in 2003 to 22% in 2019. Comparing the proportion of older patients in clinical trials to the proportion overall in the hospital registry, there was a significant increase from the start to the end of the study; p<0.001. There did not appear to be a significant time trend in the proportion of interventional clinical trials that had age-specific exclusion criteria (online supplemental figure 2).

Request permission

Discussion

We found that although older patients with cancer comprised approximately one-third of the hospital registrations, they constituted only 23% of the participants in interventional clinical trials. Thus, there is a gap between the proportion of older persons with cancer registered at our hospital and the proportion enrolled in interventional clinical trials, which suggests that the evidence generated by these trials may or may not be entirely applicable to them. Older persons with cancer constitute a unique demographic with special challenges including issues with functionality, pharmacodynamics and pharmacokinetics with resultant dosing and toxicity issues, organ dysfunction and comorbidities, polypharmacy and drug interactions. Thus, deciding the optimal treatment for these vulnerable individuals is difficult and must be based on robust evidence. Under-representation in clinical research compounds the challenges in making appropriate evidence-based treatment decisions for these patients. Designing clinical trials exclusively for older patients with cancer would be ideal, however, this would be impractical and perhaps, impossible to do in all clinical situations and would involve excessive expense, manpower and infrastructure. The next best option would be to include these older patients in general clinical trials that recruit a broad base of patients. Hearteningly, our study revealed that the gap between the representation of older patients with cancer in clinical trials and that in the overall pool of patients seen in the hospital has been significantly shrinking over time, despite the fact that there have not been any formal interventions over the study period at our institution that could have contributed to this increase in enrolment. This increase is possibly attributable to a gradual recognition of the importance of geriatric oncology resulting from advocacy and a general increase in knowledge about older patients with cancer and reflected in the establishment of a dedicated geriatric oncology clinic at our centre in 2018.28

Various other studies have examined the representation in clinical research studies of older persons with cancer. In 2003, Lewis et al reported that 32% of participants in cooperative group clinical trials sponsored by the National Cancer Institute between 1997 and 2000 were ≥65 years, as compared with 61% of incident cancer cases in the American population.21 It is pertinent to note that we compared the representation of older adults with cancer in clinical trials to their proportion in our hospital-based cancer registry. However, data from the National Cancer Registry of India suggest that approximately 46.5% of patients with cancer in the population-based cancer registries and 35.4% in the hospital-based cancer registries are ≥60 years, which is relatively similar to the proportion noted in our hospital registry.7 Talarico et al had reported that 36% of patients enrolled in drug registration trials of the US Food and Drug Administration (USFDA) between 1995 and 2002 were ≥65 years, compared with 60% in the US population.19 A follow-up analysis by Singh et al for new drug approval trials by the USFDA between 2005 and 2013 revealed that hearteningly, 41% of participants were ≥65 years, compared with 56% in the US population, possibly suggesting a narrowing of the representation gap, similar to what was suggested by our study.22 However, patients aged ≥70 years continued to be significantly under-represented: 24% in USFDA drug registration trials compared with 42% in the general US population. In our study, the proportion of patients ≥70 years was very low, both in the hospital-based cancer registry (9%), as well as in the interventional clinical trials (5%). Hutchins et al reported that 25% of patients enrolled in Southwest Oncology Group (SWOG) treatment trials between 1993 and 1996 were ≥65 years, compared with 63% in the US population.23 Unger et al updated the analysis to include trials from 1997 to 2002 and reported that 31% of patients enrolled in SWOG trials were aged ≥65 years, compared with 61% in the Surveillance, Epidemiology and End Results (SEER) Programme, again suggesting that enrolment of older persons with cancer in clinical trials has increased over time.24 All of these studies were conducted in the USA and therefore used an age cut-off of 65 years to define the geriatric age group. Comparing our study results in which we used an age cut-off of 60 years to these American studies may seem inappropriate, however, this was a comparison of enrolment of a geriatric cohort in two different populations, regardless of the age cut-off used. In general, it is the physiological age rather than the chronologic age that determines the geriatric age cut-off.29²⁹ In the absence of studies done in countries that have used 60 years as the age cut-off, we considered that this was the best method of comparing the data to give us an idea regarding the difference in representation of older patients with cancer in clinical trials conducted around the world.

We found that several factors impacted the recruitment of older patients with cancer in clinical trials. Some of these factors were actionable, for example, studies without an upper age limit in the eligibility criteria and multicentric studies were far more likely to recruit older persons. This suggests that through some simple methods (doing away with upper age eligibility criteria) and some more complex methods (expanding single centre studies to make them multicentric), it may be possible to make studies more inclusive and increase the representation of older persons with cancer in clinical trials. Some factors that predicted less recruitment of older patients like sex of the participant (older women with cancer were most under-represented in clinical trials), curative setting trials and trials evaluating intravenous systemic therapy were not immediately actionable, and these would perhaps require continued advocacy and change in the mindset of investigators as well as patients. Kimmick et al conducted a randomised trial to evaluate whether it was possible to increase the recruitment of older patients in Cancer and Leukaemia Group B (CALGB) trials through a geriatric educational intervention. They found that the educational intervention failed to result in a significant increase in enrolment, and they suggested that much more would need to be done to bring about a change in physician and patient behaviour.20 Sedrak et al recently published a systematic review on barriers to clinical trial participation of older patients with cancer, and interventions to improve recruitment.30 They characterised the barriers to participation as system level (stringent eligibility criteria, language used in the consent form, appropriate trial availability), provider level (reluctance to enrol older patients for various reasons, time and personnel constraints, lack of awareness, etc), patient level (lack of knowledge, transportation issues, time constraints, concerns about efficacy and toxicity, treatment preferences, financial problems, emotional issues and lack of self-belief) and caregiver level (caregiver concerns and caregiver burden). The authors suggested various methods to increase the trial participation of older persons with cancer, including ‘geriatricisation of trial design’ (designing trials specifically for older patients, expanding the trial design and various other trial design modifications like adaptive design, prospective cohort, embedded study), measurement of relevant endpoints including QoL (which we also found to be impactful in our study) and overall treatment utility, expansion of the trial eligibility criteria, addressing barriers at the clinical trial site or at the stakeholder level, designing pragmatic studies and leveraging real-world data. Recently, Bertagnolli and Singh published a call to action with various recommendations to eliminate the under-representation of older patients in cancer clinical trials.31 By studying the representation and factors affecting this representation at our institution, we hope to be taking the first step towards closing this evidence gap. Implementing any institution-wide change to increase the representation of older adults with cancer in interventional clinical trials would be challenging, however, perhaps the easiest and most feasible would be a mandate that the IEC questions the requirement for an upper age limit in the eligibility criteria, as long as the study intervention is not contraindicated in older adults with cancer.

Our study was limited by the fact that it was conducted at a single institution, and therefore, may not be reflective of the representation of older adults with cancer in clinical trials across India. For the purpose of this study, we sent emails to all the investigators from the institution and requested them to share the data of the studies that they had conducted; not all the investigators responded. We included the data of the investigators who responded and shared their data. We were unable to access the information regarding the studies that were not included in the analysis, or the total number of studies done in the institution. However, considering the large number of patients and trials included in the study as well as the fact that we included the data from multiple primary disease sites, and from different departments, we consider that the studies included in our analysis were likely to be representative of the overall pool of patients from our institution. The data from our hospital-based cancer registry are still in the process of being extracted, and were not available for 5 years (2009–2011 and 2015–2016). We will attempt to expand this study and obtain more complete data in the future. We compared the representation of older patients with cancer to their proportion in our hospital-based cancer registry, as we considered this to be the true denominator. However, other similar studies have used the proportion of older patients with cancer in the general population as the denominator. It appears that the proportion of older patients with cancer in the Indian population (between 35% and 46%) is only slightly higher than the proportion registered at our centre (30%), and therefore, using the denominator of the proportion in the general population would likely not have substantially changed the interpretation of our findings. We were able to obtain the dates of recruitment for little under half of the total cohort, and we used this data to determine the time trend for enrolment of patients in clinical trials at our centre. This may have introduced some bias in the time trend assessment. We found a 7% gap between the enrolment of older patients with cancer in interventional clinical trials to the proportion in the hospital-based cancer registry. Whether the gap between the proportion registered and the proportion enrolled truly signifies under-representation would have required a comparison to similar data from other institutions in India (which we were unable to find in the published literature), or a comparison to the enrolment of a younger cohort of patients, which was beyond the scope of our study. Unfortunately, an evaluation of the reasons why the number of older persons with cancer enrolled at our institute over the past few years has not increased commensurate with the increase in the number of older persons in India, was beyond the scope of our study. We found that pharma-sponsored studies had better representation of older patients with cancer than investigator-initiated studies. Most of the pharma-sponsored studies were global studies. It was our observation that these studies usually did not have an upper age limit in the eligibility criteria, which is probably the reason for the better representation of older adults as compared with investigator-initiated studies. The reason for the difference in the eligibility criteria between pharma sponsored and investigator-initiated studies was unclear. It is possible that sociocultural factors may have influenced enrolment, especially considering the fact that the lowest representation of older adults with cancer in interventional clinical trials was noted in women and for patients with gynaecological malignancies. However, it was beyond the scope of our study to characterise these factors. We hope to be able to investigate this in a more systematic manner in the future. Although we evaluated the factors that impacted the representation of older patients with cancer in clinical trials, it was beyond the scope of our study to assess whether addressing those factors may have resulted in higher recruitment.

Conclusion

There is a gap between the representation of older Indian adults with cancer in interventional clinical trials, as compared with their proportion in the hospital registry. Various factors may affect this under-representation, some of which are actionable and could be modified to increase recruitment. The representation of older adults with cancer has significantly increased over time. Adequate representation of older adults with cancer in clinical research is a critical factor in deciding the optimal cancer-directed therapy for this group of vulnerable individuals.