Report: Diversity in Patient Recruitment for Clinical Trials

Report: Patient Diversity in Clinical Trials

The importance of maintaining patient demographic diversity in the recruitment and reporting of clinical trials.

The following is a snapshot of the full-report including citations that can be downloaded here.

 Contributors: Sarah Baker MSc(Biostatistics); Rosetta Aires B (Law).; Sophia Pulasis PhD. (Bioinformatics)


Patient diversity is an important feature of clinical trials recruitment in order for the study population to accurately represent the real-world patient populations that evidence shows are proportionally affected by the disease under investigation. This ensures that the insights gleaned in the clinical trial have external validity and that the findings can be generalised to the wider disease population, i.e. the population who will be utilising the therapeutic or diagnostic device under investigation once it attains regulatory approval.

This report, based upon a review of the current clinical trial literature, examines how well existing study populations in clinical trials reflect the real-world populations of their respective diseases. Demographic factors of sex, age, ethnicity are the primary focus. It discusses reasons why particular imbalances in these attributes may be observed and evaluates whether or not these imbalances can reasonably be ameliorated. The report then outlines methods and focus points to improve diversity in clinical trial recruitment for specific patient sub-groups where it is reasonable to do so.



Patient sex ratios in clinical trials recruitment

The reporting of safety and efficacy findings by sex difference

Sex ratio differences by disease type and disease burden

Pregnant women and Clinical Trials

Children and clinical trials

Elderly populations and clinical trials

Ethnic minorities and clinical trials

UK vs USA Factors behind insufficient diversity

FDA Guidelines on enhancing diversity (2020)

The role of bioinformatics in maintaining appropriate patient diversity in clinical trials.

Biostatistical approaches to analysing minority sub-groups that contain a very low number of subjects.


Anecdotal insights into barriers of participation in clinical trials

Recommendations for increasing clinical trial participation generally.


Patient demographics such as sex, ethnicity and age have been shown to influence the safety and effectiveness a therapeutic intervention in terms of how it performs in a given individual. This can be the case whether the therapeutic in question is a pharmaceutical, surgical or device. These demographic factors tend to be associated differences in physiological attributes and biomolecular processes such as organ tissue shape, size and function. Differences in biomolecular, genomic and other “omics” biomarkers also tend to occur as a function of either sex, ethnicity or age and these differences can be associated with differences in the effect of a drug via pharmacodynamic and pharmacokinetic mechanisms.

Elderly people carry 60% of the national disease burden and yet represent approximately 32% of participants in Phase II and III.[4]  Clinical trial participation of elderly people is low in research on cancer, Alzheimer’s disease, arthritis, and epilepsy, all diseases which disproportionately affect the elderly.[5] 

In many cases this discrepancy between clinical trial participants and disease burden is due to safety related eligibility criteria precluding the most frail or vulnerable patients, children or women of childbearing age. The exclusion of these groups is down to legislation in many contexts and the fact that certain clinical trials would not be safe or ethical for them to participate in. The down-side of this lack of representation in clinical trials is the increased challenge of gaining a complete picture of a drug or medical device’s safety and efficacy in these very populations.

Recent guidance by the NIHR and FDA emphasises the importance of sponsors striving towards a recruitment strategy whereby clinical study participants accurately reflect the intended real-world use population for a given therapeutic. In the cases where this is not achieved, any conclusions may confound clinicians as to the ideal treatment regimen for an individual patient or lead to a sub-optimal course of treatment being pursued. When pharmaceutical companies market a therapeutic without nuance, this portrays a skewed public health message that can further exacerbate healthcare inequities. [1, 43] Most recent figures for UK clinical trials show that only 85.7% of studies reported the age of participants, 90% reported participant sex and 60% reported ethnicity.[43] From both a clinical and statistical perspective this information is valuable and necessary and should be universally required by regulation. Having this data reduces the chance of unforeseen outcomes or treatment failures over the course of the regulatory approval process and beyond.

The importance of accounting for diverse demographics from a statistical perspective.

To derive accurate insights from statistical analysis the resolution of the data matters. Too much resolution prevents precise generalisable insights about the population, and too little resolution creates paradox and bias. Commonly, not including sufficient demographic information in a statistical model goes wrong it can result in bias such as Simpson’s Paradox. This widespread phenomenon occurs when the direction of the overall trend between two variables is reversed once a third variable is accounted for.

An example in clinical studies might be a positive relationship between treatment effect and dose turning into a negative effect when the data is stratified by age group. In this situation the trend between treatment effect and dose will be negative for some or all of the age categories. More information about this particular form of bias can be found here. Of course the situation could be in the reverse where a treatment appears not to be effective based on the overall analysis but is found to be highly effective when the data analysis is stratified on a demographic or biomarker basis.

There are many forms of bias that can crop up when the data is collected in such a way as to provide an incomplete picture of the target population. This may even have short term benefits for the individual study however negative consequences such as costs and penalties of over-investing in the wrong therapy, proscribing to the wrong patient population or abandoning a clinical trial pre-maturely.

Trends and findings in the current literature

Patient Sex in clinical trials

  • Trends in sex ratios of enrolled patients differ based on clinical trial phase.
  • Females less prevalent early phases due to unknown or known adverse effects of novel therapeutics on female fertility or unborn foetus.
  • Under-representation of women in phase III when looking at disease prevalence
  • Males are under-represented as often as females when comparing enrolment ratios to disease burden proportions
  • There is a need for trial demographics to be recruited on a case-by-case basis, depending on the disease.

Differences in proportional representation of the sexes by clinical trial phase

It has been long bemoaned in journalistic sources that female representation in clinical trials is lacking, despite recent efforts to mitigate the gap. While this is a tantalising headline and has certainly been the case over the 20th century, the truth of these claims does not bare up under close expert scrutiny of the clinical trial literature.

US data from 2000-2020 suggests that trial phase has the greatest variation in enrolment sex ratio when compared to other factors, with median female enrolment being 42.9%, 44.8%, 51.7%, and 51.1% for phases I, I/II to II, II/III to III, and IV7. This shows that median female enrolment gradually increases as trial phases progress, with the difference in male/female enrolment between the final phases II/III to III and IV being <1%.

Additional US data on FDA approved drugs including trials from as early as 1993 report that female participation in clinical trials is 22%, 48%, and 49% for trial phases I, II, and III respectively8. While the numbers for participating sexes are almost equal in phases II and III, females make up only approximately one fifth of phase I trial populations in this dataset8. The difference in reported participation for phase I trials between the datasets could be due to a push towards increased female participation in dose-finding studies over more recent years. None-the-less care must be taken when interpreting clinical studies from this period.

The aim of a phase I trial is to evaluate safety and dosage, so it comes as no surprise that women, especially those of childbearing age, are often excluded due to potential high risks posed by a given therapeutic to female fertility and foetal development.

In these cases, women are included to a greater extent as trial phases progress and the potential risk of severe adverse events decreases. By the time a trial reaches phase III, it should ideally reflect the real-world disease population as much as possible. This is particularly the case where the standard treatment for the disease also poses a risk to fertility, which can be the case in oncology for example.

European data for phase III trials from 2011-2015 report 41% of participants being female9, which is slightly lower than female enrolment in US based trials. Further to this 26% of FDA approved drugs have a >20% difference between the proportion of females in phase II & III clinical trials and the prevalence of the disease in the female US population8, and only one of these drugs shows an over-representation of women.

The reporting of safety and efficacy results stratified by sex

  • Both safety and efficacy results tend to differ by sex.
  • Reporting of these differences is inconsistent and often absent
  • Higher rates of adverse events in women are possibly caused by less involvement or non-stratification in dose finding and safety studies.
  • There is a need to enforce analysis and reporting of sex differences in safety and efficacy data

Drug safety parameters very often differ as a function of sex. Study results have shown females to have a slightly higher percentage (p<0.001) of reported adverse events (AE) than males for both treatment and placebo groups in overall clinical trials9. Sub-group analyses of safety data are important in clinical trials and can offer insights into any unique risks that females may be subjected to during treatment. Despite this, sex specific safety analyses are available for only 45% of FDA approved drugs, with 53% of these reporting more side effects in females8. On average, females are at a 34% increased risk of severe toxicity for each cancer treatment domain, with the greatest increased risk being for immunotherapies (66%). Moreover, the risk of AE is greater in women across all AE types, including patient-reported symptomatic (female 33.3%, male 27.9%), haematologic (female 45.2%, male 39.1%) and objective non-haematologic (female 30.9%, male 29.0%)10. These findings highlight the importance of sex specific safety analyses and the fact that more sex stratified safety reporting is needed. More reporting will increase our understanding of sex-related AE and could potentially allow for sex-specific interventions in the future.

Sex stratified analyses for efficacy are available for 71% of FDA approved drugs, and of these, 11% were found to be more efficacious in males and 7% in females8. Alternatively, only 2 of 22 European Medicines Agency approved drugs examined were found to have efficacy differences between the sexes9. Nonetheless, it is important to stratify efficacy data of a novel therapeutic on disease population subgroups that may end up receiving it once approved.

Sex differences in terms of both safety and efficacy have been widely reported in the clinical trial literature. As a rule sex-based stratification of the data are important as they can reveal sex differences in dose optimisation, safety and adverse events which will enable clinicians to make more informed treatment decisions, on an individual basis, once a drug reaches the market. This has a knock-on effect of increasing the odds of successful phase IV and surveillance studies.

Sex differences in clinical trial enrolment by disease type and burden

  • Several disease categories have recently been associated with lower female enrolment.
  • An equal number of disease classifications have been associated with an under-representation of males, particularly with regards to their corresponding disease burden.

Sex differences by disease type

It is important to note that when classified by disease type, the sex ratio of clinical trial participation shows a more nuanced picture. Several disease categories have recently been associated with lower female enrolment, compared to other factors including trial phase, funding, blinding, etc7. Females comprised the smallest proportions of participants in US-based trials between 2000-2020 for cardiology (41.4%), sex-non-specific nephrology and genitourinary (41.7%), and haematology (41.7%) clinical trials7. Despite females being proportionately represented in European phase III clinical studies between 2011-2015 for depression, epilepsy, thrombosis, and diabetes, they were significantly under-represented for hepatitis C, HIV, schizophrenia, hypercholesterolaemia, and heart failure and were not found to be over-represented in trials for any of the disease categories examined9. This shows that the gap in sex representation exists even in later clinical trial phases when surveying disease prevalence, albeit to a lesser extent.  Examining disease burden shows that the gap is even bigger than anticipated and includes the under-representation of both sexes.

Sex Differences by Disease Burden

It is not until the burden of disease is considered that males are shown to be under-represented as often as females. Including burden of disease can depict proportionality relative to the variety of disease manifestations between men and women. It can be measured as disability-adjusted life years (DALYs), which represent the number of healthy years of life lost due to the disease. Despite the sexes each making up approximately half of clinical trial participants overall in US-based trials between 2000-2020, all disease categories showed an under-representation of either women or men relative to disease burden, except for infectious disease and dermatologic clinical trials7. Females were under-represented in 7 of 17 disease categories, with the greatest under-representation being in oncology trials, where the difference between the number of female trial participants and corresponding DALYs is 3.6%. Males were under-represented compared with their disease burden in 8 of 17 disease categories, with the greatest difference being 11.3% for musculoskeletal disease and trauma trials.7 Male under-representation in mental health and trauma research is significant as in the US suicide, violence, and drugs are associated with higher morbidity. Males were found to be under-represented to a similar extent to females, suggesting that the under-representation of either sex could be by coincidence. Alternatively, male under-representation could potentially be due to the assumption of female under-representation leading to overcorrection in the opposite direction.  It should be noted that these findings would benefit from statistical validation, although they illustrate the need for clinical trial participants to be recruited on a case-by-case basis, depending on the disease.

Pregnant Women and Clinical Trials

There is concern about including pregnant women in many clinical trials due to the unknown effects that drugs may have on the developing foetus. The lack of long-term developmental safety information for almost any drug further compounds this concern.[11] In the majority of cases it is not ethical or safe to include pregnant women in the early phase clinical trial process and guidelines are laid out in the FDA regulations that govern research in pregnant women (CFR).[12]

Women have historically been excluded from pharmaceutical research for many reasons, both cultural and scientific. One of these reasons is the increased risk to female fertility compared to male fertility of certain types of pharmaceutical, particularly oncological agents. As discussed in the previous section this can lead to a lack of knowledge around the ideal dose and safety concerns for women taking the drug once it is regulated. What also needs to be considered is the off-label use of the drug once it has reached the market.

A well-known example where this has gone awry when female specific data is lacking is the thalidomide disaster of the 1960’s. Specifically this illustrates the horrendous effects of administering an off-label drug to pregnant women which has not been stringently tested in this population,[13]  Thalidomide, now used as an immunomodulatory medicine in the treatment of cancer, was initially marketed as a sedative and used off-label to treat anxiety, depression, coughs colds, insomnia and morning sickness in pregnant women. The drug was later found to cause death and severe limb deformities in newborns, as well as having dangerous side-effects in adults including peripheral neuropathy, blood clots and severe cardiovascular events. [14] 

This case study was a major turning point which emphasised the importance of safety studies in humans and the need to include females in clinical trials. Of course the ethical dilemma of subjecting an unborn foetus to unknown risks from a novel therapeutic agent remains controversial.

 There are various factors in favour of the exclusion and inclusion of pregnant women in clinical trials. This issue was addressed by the National Commission for the Protection of Human Subjects of Biomedical and Behavioural Research. It recommended that “non-therapeutic research on the pregnant woman or on the foetus in utero may be conducted or supported, provided it will impose minimal or no risk to the foetus, the woman’s informed consent has been obtained, and the father has not objected”.[15] 

Early phase studies observe how a drug works in the body, and would provide vital information on the effects on the foetus; however this would pose the greatest risk to the foetus and would be unethical to carry out. A new grading system to monitor complications during clinical trials involving pregnant women has been co-developed by UCL researchers and an international team of experts. The grading is vital in determining what dose of medication can be safely offered to mothers.[16] Even so, this would only be applicable for lower risk therapeutics. It makes little sense to put a foetus at risk in an early dose or safely study when clinical efficacy has not yet been established in non-pregnant subjects.

Children and Clinical Trials

Paediatric clinical trials are more challenging to conduct than clinical studies in adults. This is due to legal and ethical concerns, and complexities in the enrolment, consent, and access of children to clinical research.[17] The side effects of a drug could be more detrimental to children than adults, as they have developing bodies and organs, side-effects may stunt their growth and produce long lasting damage.[18]

In many cases it is not sufficient to carry out research with adults and apply the findings to children. In addition to the concerns mentioned above, disease processes in children might differ from those in adults. Some childhood diseases have no close analogies in adults, therefore to understand these it is necessary to carry out research with children directly. [19] In many cases there are statistical methods, such as Bayesian extrapolation, that are cautiously applied in order to extrapolate from adult subjects to children.

Children included in clinical trials require special protection because they are less likely than adults to be able to communicate their needs effectively or defend their own interests. Children may not have the capacity to fully understand risks or give consent and therefore the consent of a guardian is required in most cases. The requirements for consent, where participants in clinical trials are children depend on the form of study and where in the UK it is taking place.[20] “The Medicines for Human Use (Clinical Trials) Regulations prohibit children under the age of 16 from giving consent to take part in a Clinical Trial of an Investigational Medicinal Product (CTIMP).Young people over 16 are presumed to be capable of giving consent on their own behalf to participate in Clinical Trials of Investigational Medicinal Products (CTIMPs)”.[21]

Elderly Persons and Clinical Trials

 There are nearly 2 million elderly people on 7 or more prescription medicines in the UK.[22] Despite this poly-pharmacy is almost never studied. Elderly people carry 60% of the national disease burden but represent approximately 32% of participants in Phase II and III clinical trials

.[23] Moreover older people suffer the greatest health disparities, enduring disproportionately high rates of cancer, cardiovascular disease, dementia, arthritis, and Parkinson’s disease.[24] Despite this it would be unethical to include elderly participates in the earlier stages of the trial process as older people may be frailer and their metabolism may process drugs differently to younger people. According to The Medicines for Human Use Regulations (2004)[25] elderly people cannot participate in a clinical trial, if the trial would have an adverse effect on their health.

As a result, much clinical research of relevance to elderly patients examines individuals who are younger than those who typically have the disease in question, specifically in phase I of the trial process. An example is heart failure. Among 251 trials investigating treatments for heart failure, 64 excluded patients by an upper age limit. Overall 109 trials on heart failure had 1 or more exclusion criteria that could limit the inclusion of older individuals.[26] Exclusion of older individuals from ongoing trials regarding heart failure continues to be prevalent, this is also apparent for type 2 diabetes, colorectal cancer, hypertension and Alzheimer’s disease. This type of exclusion is generally due to the health of older people and the ethical concerns surrounding elderly people participating in a clinical trial.[27]

It was noted that between 2008 and 2015 the number of emergency hospital admissions caused by adverse drug reactions increased by 53 percent. 1 in 50 cases proved to be fatal[28], thus supporting the need for better representation. There needs to be more testing and inclusion of older people within the clinical trial process to avoid such adverse events once a drug reaches the market, however this needs to be done as ethically and safely as possible.

There is now a growing need associated with the global burden of disease to prevent, diagnose and treat chronic diseases[29] as the use of medical products, including devices, increases with age.[30] Increasing the inclusion of elderly participants in different phases of the clinical trial process can increase the number of risk modifiers. Where it is not possible or ethical to recruit the most frail of patients statistical extrapolation techniques such as Bayesian extrapolation can be applied to fill in the gaps.

In addition to ethical concerns a further reason for under-representation of elderly patients is that elderly people may be less motivated to participate in clinical trials as the process can be more physically and mentally draining with age, they tend to be less mobile impacting travel to the trial site, and may have an increased concern with side effects. Elderly patients may also not meet inclusion criteria as a result of poly-pharmacy. Thus increasing the inclusion of elderly people entails a different approach and clinical trials should be designed in a way that elderly people feel more inclined to participate and that ensures their needs are met.[31]

Part of this solution should include attention to patient transport in a way that mitigates cost and increases convenience. With cars and public transport now untenable for many, a patient chauffeur service may make participating in clinical trials more attractive to those with reduced mobility or other age and disability related concerns.

In some situations home-based participation via digitisation and decentralisation may provide a solution. The limits to this approach being it would require both a therapeutic amenable to low-supervision as well as sufficient techno-literacy in the participant or carer.

Ethnic Minorities and Clinical Trials

A recent review found that 13% of over all trial participants commencing studies between 2007-2017 were of BAME ethnicity, which roughly aligns with census indications of the proportion of BAME individuals in the UK population.[43] This report by the NIHR also uncovered 68 distinct categories of ethnicity used across the clinical trials. This points to the need have a standardised set of around five ethnic categories so that statistical analysis are both viable and able to be compared across different studies.

Medical treatments can have varying effects for people of different ethnic backgrounds, and hence it is essential that clinical trials include participants that are representative of the different ethnic groups that comprise our population. This approach enables subgroup analysis of clinical trial data, which improves the external validity of the study findings, particularly in terms of the safety and efficacy of the therapeutic when it is used across ethnic sub-groups.[32 Under-representation of BAME participants has been identified with respect to treatment development for cardiovascular disease, cancer and dementia.[33] This is despite the fact that these and many other diseases disproportionately affect BAME groups in the UK and globally.

Among black and minority ethnic (BAME) patients, distrust in research and medical professionals was a common reason given for not wanting to participate.[35] Concerns about privacy and confidentiality have increased over the past decade. This portrays that the lack of diversity is down to attitudinal factors such as a fear of participating in clinical trials.

An under-researched area in the context of diversity is methods to accommodate language needs of ethnic minorities who are under-represented, some of whom might have low first language literacy. Engaging with a large proportion of groups to make recruiting easier is unlikely to be simple or cheap.[34] Simply translating a document into a target language is unlikely to be an effective solution for this complex challenge. Such an approach is important for working with communities to facilitate understanding and develop trust.

It is also important to have a clear definition of the terms “ethnicity” versus “nationality”. Anecdotally, there are many instances of the NHS conflating these two concepts. For example, persons with Australian nationality being classified in the UK as having Australian ethnicity despite having British or European ancestry. This runs the risk of incorrect conclusions being drawn about Australia’s indigenous population who are well documented to respond differently to many psychoactive agents and are typically highly under-represented in clinical trials. This “white-washing” is of particular concern in the context of safety and efficacy studies and also may affect diagnostic categorisation as well as the insights derived from genome-wide studies.

UK vs USA factors behind insufficient Diversity

A Health Technology Assessment (2005) has found that NHS policy principles of equality and diversity were not formally promoted,[36] despite concluding that estimates of efficacy and validity would be compromised if drugs are not tested on different socio-demographic groups. With growing evidence of ethnic health inequalities it is important to understand and tackle the root causes. Apart from the ethical and social arguments, there are rational scientific, clinical, health and economic reasons to include different populations in research. [37]

People in the U.S without insurance can be linked to the lack of involvement in the healthcare system generally. Special care is often inaccessible and can be expensive, further limiting who can and cannot afford care and therefore who is exposed to clinical trial opportunities.[38] The issue of insurance is not a distinctive problem in the UK thanks to the public health services that we have.

Difficulty finding transportation to reach clinical trial sites, overly stringent inclusion and exclusion criteria, language barriers and distrust toward medical research also all contribute to the lack of diversity in clinical trials today.[39] 

FDA Guidelines on Enhancing Diversity in clinical trials recruitment and analysis (2020)

There have been several studies done exploring the reports submitted to FDA with specific focus on the diversity of patients participating in clinical trials. Unfortunately these studies found that for most clinical trials there was no information included about both safety and effectiveness or sensitivity and specificity for all applicable demographic subgroups.[41]

The FDA recently issued guidelines on enhancing the diversity of clinical trial populations

● Ensure trial participation is less burdensome for patients (reducing the frequency of study visits, using digital health technology tools, working with mobile medical professionals, and offering compensation for costs associated with participation)

● Enrol participants who reflect the characteristics of clinically relevant populations with regard to age, sex and ethnicity.

● Include ethnic minorities in clinical trials and ensure the analysis of clinical trial data includes stratification by key demographic factors.

● Provide clinical study information and resources in multiple languages[42]

FDA Takes Important Steps to Increase Racial and Ethnic Diversity in Clinical Trials 2022

The FDA issued new guidance for developing plans to enrol more participants from under-represented racial and ethnic populations in the U.S. into clinical trials. There are biological differences which exist in how people respond to certain therapies. For example, variations in genetic coding can make a treatment more or less toxic for one racial or ethnic group than another. These variations can also make drugs like antidepressants and blood-pressure medications less effective for certain groups. Individuals from these populations are frequently under-represented in biomedical research despite having a disproportionate disease burden for certain diseases relative to their proportional representation in the general population.[43]

The role of bioinformatics in maintaining appropriate patient diversity in clinical trials.

  • Clinical data can be mined to understand disease demography
  • Examination of genomic data can identify genomic biomarkers associated with disease
  • These insights should be included in trial planning & patient recruitment

Publicly available data can be used in large-scale bioinformatics analyses before clinical trial design to tailor the participant population to the disease in question as each disease has its own distinct demography. Hospital records spanning long periods of time can offer insights into the demographic makeup of a specific disease. Examining these records and looking for trends in patient populations can allow researchers to build a picture of how factors such as sex, age, and race are distributed across patients with the disease. Further investigation may even give insights on more specific circumstances, including geographic regions, which could indicate specific areas or hospitals to sample from.

Having this information before carrying out a clinical trial enables the recruitment of patients that better represent the disease population for the clinical trial. It should be noted, however, that there are limitations that need to be considered and not every disease population can be included in a clinical trial. Conditions that affect vulnerable groups including children and pregnant women cannot be examined so easily due to the increased risks and ethical considerations associated with trial participation.

Having a trial population that proportionally represents the disease population in terms of demographic features means that findings during the trial related to therapeutic effectiveness can be generalised and applied to the disease population. As a result, the therapeutic will have more predictable effects on patient’s post-market approval, and less adverse effects or reasons for recall.

This approach can be taken a step further by examining the genomic profiles of the disease population. Many diseases are associated with unique genomic biomarkers, often in the form of gene expression and mutations. These markers can be used in both the diagnosis of a disease and as indications for risk factors for becoming ill eventually. Existing data and literature can be mined for biomarker information associated with the disease under study, and if used to inform patient recruitment, could result in even more representative trial populations.

In the future, in silico trials and virtual patient populations based on specific disease demographics may be used to complement clinical trials. By using patient data from clinical databases to create a computational model that reflects age, sex, ethnicity, and disease burden, biological variability amongst individuals could be simulated. Including interactions between anatomy, physiology, and blood biochemistry in the model could further predict the impact of a therapeutic. Although far from current reality, this approach has been applied in a brain aneurysm medical device study with encouraging results, illustrating its potential use in the future. 

Biostatistical approaches to analysing minority sub-groups that contain a very low number of subjects.

Particular care needs to be taken when considering statistical comparison of a very small n sub-group with another group or groups of of more typical size. Most frequently this can occur when participants who are transgender, intersex or from an ethnicity or age group that is under-represented in the context of the study comprise a sub-group category where n is <6. From a statistical perspective it is not possible to derive robust generalisable insights from these individuals on a stratified basis. As all representative diversity is welcome and valuable, these individuals should be included in the overall analysis but in certain circumstances may need to be excluded from subsequent sub-group analyses. They can sometimes be added to larger sub-groups if their data is not significantly outlying or skewing the results in that group. In any case, care must be taken and all decisions should be recorded clearly in the statistical methodology and results.

A further example where standardised categorisation is needed was outlined above with the 68 non-unified ethnic categories across multiple studies. Meaningful statistical analysis cannot be conducted until the number of categories, and hence the resolution of the data, is reduced in a standardised way. It would generally not be advisable to use more than 5 distinct ethnic categories, unless the sample size of the study was very, very large and this is typically not the case in clinical trials. The reason that standardisation of categories is important is so that comparisons can be made across studies. There also needs to be a clinically sound rationale to the chosen ethnic categories, such as indications of disease or treatment response variation on a biomolecular basis.


Clinical trials of devices in pharmaceutical trials for adults have in the past under-represented women, BAME , children and patients over the age of 65 and have shown bias towards healthier patients to participate.  If looking at different studies done in the past years, there are studies that have been submitted to the FDA that may have included diverse participants but did not include adequate demographic diversity or sub-group analyses of the study data, which in itself results in a loss of resolution in terms of the insights gleaned from the study pertaining to these subgroups. .[6]  This represents a wasted opportunity to derive maximal value in terms of the immense resources, financial and otherwise, already invested into the study. When looking at the lack of representation; children, pregnant women and the frail elderly will not always participate in clinical trials as there is legislation that prevents this in cases where it would be fundamentally unethical and unsafe. For some pharmacologic agents this includes women of child bearing age in early phases of drug research such as dose-finding and safety studies, due to unknown or known effects on female fertility. A common example of this is chemotherapeutic agents in oncology.

There is room for improvement in the proportional representation of both sexes in clinical trials and knowing a disease demographic is vital to planning a representative trial. Assuming the under-representation is on the side of female rather than male may lead to incorrect conclusions and paradoxical actions to redress the balance. Taking demographic differences in disease burden into account when recruiting trial participants is needed. Trial populations that more accurately depict the real-world populations will allow a therapeutic to be tailored to the patient. As a caveat it is also crucial to keep in mind that limitations to including pregnant women, women of child bearing age, children, or than frail elderly at early phase clinical trials for certain high risk or highly novel therapeutics must be maintained and a risk assessment should be conducted in each instance.

Efficacy and safety findings highlight the need for clinical study data to be stratified by sex, age and ethnicity so that respective safety and efficacy estimates can be determined. This enables more accurate, sex/age-appropriate dosing that will maximise treatment efficacy and patient safety, as well as minimise the chance of adverse events. This also reduces the risks associated with later off label use of drugs and may avoid modern day tragedies resembling the thalidomide tragedy. Moreover, efficacy and adverse events should always be reported separately for men and women, as the evidence shows their distinct differences in response to therapeutics.

In line with the trend towards personalised and recision medicine, it will be more common to see clinical trial results stratified by patient biomarker and genomic factors as the future unfolds.