Diversity in Clinical Research & Precision Medicine: A Patient's Guide to Pharmacogenomics and Health Equity — Cost, Top Hospitals & Success Rates | MyMedicPlus

Precision medicine — the individualisation of treatment based on a patient's genetic profile, biomarkers, environment, and lifestyle — holds transformative promise. Yet its benefits cannot be equitably delivered if the clinical research underpinning treatment decisions has been derived predominantly from one population group. For much of the 20th century, clinical trials enrolled participants who were disproportionately white, male, and from high-income countries — a structural bias with real-world clinical consequences.

Pharmacogenomics is the study of how genetic variation between individuals and populations influences drug response, metabolism, and adverse effect risk. Enzymes of the cytochrome P450 (CYP) superfamily — particularly CYP2D6 and CYP2C19 — are responsible for metabolising approximately 25% of all commonly prescribed drugs, including antidepressants, antipsychotics, anticoagulants, opioids, and antiretrovirals. The frequency of functional polymorphisms in these genes varies dramatically between ancestral populations, meaning that a standard dose producing a therapeutic effect in one population may produce toxicity or therapeutic failure in another.

The failure to enrol diverse populations in clinical trials has several downstream consequences: treatment guidelines calibrated for the majority population may systematically under-treat or over-dose patients from different ancestral backgrounds; adverse drug reactions may go undetected in underrepresented groups during drug development; and biomarker associations validated in one population may not generalise to others.

Recognising this, the FDA issued final guidance in April 2024 requiring sponsors of Phase III clinical trials for diseases with differential prevalence or severity across race/ethnicity, sex, age, or geographic groups to submit Diversity Action Plans specifying enrolment targets and strategies for underrepresented populations. This represents the most significant regulatory step toward trial equity in the FDA's history.

While population diversity matters across all of medicine, several therapeutic areas have the most compelling evidence that under-representative research leads directly to suboptimal or harmful care for specific populations:

• Cardiovascular Disease: The landmark BiDil trial (A-HeFT, NEJM 2004) demonstrated that fixed-dose isosorbide dinitrate/hydralazine reduced mortality by 43% and hospitalisation by 33% in Black patients with heart failure — a finding that had been missed in racially homogeneous earlier trials. BiDil received FDA approval in 2005 as the first race-specific drug, illustrating both the potential and ethical complexity of race-stratified medicine. More broadly, hypertension prevalence, severity, and pharmacological response differ between ancestral groups.
• Oncology: Lung cancer mutation frequencies (EGFR mutations more prevalent in East Asian populations; KRAS G12C more prevalent in White patients) mean that targeted therapy eligibility — and therefore survival — is directly influenced by ancestral background. Breast cancer triple-negative subtype is more prevalent in Black women, yet this group has historically been under-enrolled in clinical trials.
• Psychiatry and Neurology: CYP2D6 and CYP2C19 polymorphisms profoundly affect the metabolism of antidepressants (SSRIs, TCAs), antipsychotics, and pain medications. Poor metabolisers (more prevalent in certain populations) on standard doses may experience severe toxicity; ultra-rapid metabolisers may derive no benefit. Without pharmacogenomic testing informed by population-appropriate reference data, prescribers may systematically mis-dose patients from minority communities.
• Infectious Disease: Antiretroviral pharmacokinetics, HBV e-antigen status, and malaria drug resistance profiles all vary by geographic ancestry. Treatment guidelines developed for European populations may require adaptation for patients from sub-Saharan Africa, South and East Asia.
• Metabolic Disorders: Type 2 diabetes manifests at lower BMI thresholds in South Asian and East Asian populations, yet risk algorithms and treatment thresholds have historically been calibrated on European cohorts.

One of the most important messages in health equity research is that clinical trial participation is open to a far broader population than most patients realise. Historical barriers — geographic, linguistic, cultural, structural, and educational — have discouraged participation from underrepresented communities. Understanding eligibility demystifies the process.

Who can participate:

• Most interventional clinical trials enrol adults aged 18 and above who meet specific health and disease criteria. Many now have expanded age criteria to include older adults (previously systematically excluded) and adolescents where appropriate.
• Observational and biobank studies (such as the NHGRI All of Us Research Program) are open to all adults in the United States, including those without existing health conditions, with explicit priority on recruiting individuals from communities historically underrepresented in biomedical research.
• The All of Us program has enrolled over 800,000 participants as of 2025, with more than 50% from racial or ethnic minority groups — making it the most diverse biomedical database ever assembled.
• The UK Biobank (500,000 participants) and 100,000 Genomes Project (Genomics England) are comparable resources in the UK, though with recognised diversity limitations that ongoing expansion aims to address.

How to find trials: ClinicalTrials.gov lists over 450,000 registered studies globally. Patients can search by condition, location, and eligibility criteria. The FDA Diversity Action Plan requirement means that trials for conditions affecting minority groups must now proactively outreach to those communities. Additionally, community health centres, federally qualified health centres (FQHCs), and patient advocacy organisations serve as important trial referral sources for underserved populations.

Participation is always voluntary, and no incentive should be so large as to constitute coercion — a principle enshrined in the Belmont Report and all contemporary research ethics frameworks.

Patients and healthy volunteers can contribute to diversity in medical research through several distinct pathways, each with different implications for personal benefit and risk:

Observational Studies: Participants provide biological samples (blood, saliva, urine), complete questionnaires, and share electronic health record data. No experimental treatment is administered. Risk is minimal. The NHGRI All of Us Program is primarily observational and returns individual results to participants, including pharmacogenomic reports and ancestry-linked health predispositions.

Interventional (Phase I–IV) Clinical Trials: Participants receive an investigational drug, device, or procedure under controlled conditions. Phase I trials focus on safety and dosing in small cohorts; Phase II on efficacy signals; Phase III on randomised comparative effectiveness; Phase IV (post-marketing) on long-term safety and special populations. All phases require rigorous informed consent and ethics board (IRB/MREC) approval.

Pharmacogenomic (PGx) Testing as a Clinical Tool: Beyond research participation, patients can access clinical pharmacogenomic panels as a treatment optimisation service. The Clinical Pharmacogenetics Implementation Consortium (CPIC) has published evidence-based dosing guidelines linking specific CYP2D6, CYP2C19, TPMT, DPYD, and UGT1A1 genotypes to dose adjustments or drug avoidance recommendations. Population-specific allele frequencies from databases such as gnomAD and PharmGKB help clinicians interpret PGx results in diverse patients. PGx testing panels (covering 40–80 genes) are available through reference laboratories at costs ranging from USD 200–2,000, with variable insurance coverage.

Biobank Participation: Donating tissue samples or consenting to long-term data linkage enables population-scale genomic research. Participants in diverse biobanks directly enable the discovery of population-specific drug response associations that ultimately translate to better prescribing guidelines for their communities.

Investing in diversity in clinical research delivers measurable, tangible benefits for individual patients and for public health systems globally:

Individually Appropriate Drug Dosing: When pharmacogenomic reference data includes a patient's ancestral background, clinicians can select and dose medications with far greater precision. For example, CYP2C19 poor metaboliser status (more prevalent in East and South Asian populations) means that standard clopidogrel dosing may provide inadequate antiplatelet effect after cardiac stenting — a pharmacogenomically informed prescriber would switch to ticagrelor or prasugrel, reducing stroke and myocardial infarction risk.

Prevention of Adverse Drug Reactions: Adverse drug reactions (ADRs) account for approximately 6.5% of hospital admissions in the UK and are estimated to cost the US healthcare system over USD 30 billion annually. Many ADRs are pharmacogenomically predictable; expanding PGx research across diverse populations reduces the chance that population-specific risk profiles go undetected during drug development.

Discovery of Population-Specific Biomarkers: Genome-wide association studies (GWAS) conducted in diverse cohorts identify disease risk loci and drug response variants that are not present at detectable frequency in European reference populations. The expansion of GWAS to diverse ancestries through initiatives like the H3Africa consortium has already identified novel hypertension and diabetes loci.

Reduction of Health Disparities: When treatment guidelines are based on diverse evidence, the structural health disparities that result from applying "majority population" medicine to minority patients can be meaningfully reduced. This is the central objective of the FDA 2024 Diversity Action Plan mandate and the NHGRI All of Us Program.

Acceleration of Drug Development: Diverse trial populations that better represent the eventual market of a drug yield more generalisable efficacy and safety data, reducing the likelihood of post-marketing safety signals emerging disproportionately in specific populations.

There are two distinct categories of risk to consider in this context: the risks of participating in clinical research, and the often-greater but less visible risks of not having one's population represented in the research base.

Risks of the Representation Gap (Systemic Risks):

• Systematic mis-dosing: Standard drug doses calibrated in predominantly white European populations may be too high (toxicity) or too low (therapeutic failure) for patients with different CYP450 genotype frequencies
• Undetected adverse effects: Rare but serious ADRs that manifest predominantly in specific ancestral groups (e.g., abacavir hypersensitivity in HLA-B*57:01 carriers, more prevalent in certain South Asian subgroups) may not be detected during drug development in non-diverse trial populations
• Perpetuation of health disparities: Evidence-based guidelines built from non-diverse evidence systematically disadvantage patients whose biology or social determinants of health differ from the reference population

Risks of Research Participation (Individual Risks):

• Observational studies/biobanks: Minimal physical risk; primary concerns are privacy and data security. Well-designed studies use de-identification, data access committees, and consent frameworks addressing re-identification risk
• Interventional trials (Phase I): Exposure to investigational drugs with incompletely characterised safety profiles; possibility of unexpected adverse effects; careful dose escalation and stopping rules mitigate but do not eliminate this risk
• Interventional trials (Phase II–IV): Uncertainty about comparative efficacy vs standard of care; random assignment may result in receiving placebo or comparator rather than the experimental agent; time burden of more frequent study visits
• Psychological risks: Receiving unexpected pharmacogenomic or disease-risk genetic results through research return programmes may cause anxiety; genetic counselling support should be available

Participating in clinical research does not end at the last study visit. Understanding what follows helps patients make an informed decision about joining a study and prepares them for ongoing engagement with the research process.

Results Return: The approach to returning individual research results varies by study design. Interventional clinical trials typically do not return individual-level results during the blinded phase; unblinding occurs at trial completion. The NHGRI All of Us Program has a pioneering policy of returning genomic health-related results directly to participants, including pharmacogenomic reports (CYP2D6/CYP2C19 status) and ancestry-linked disease risk information, accompanied by educational resources and access to a genetic counsellor.

Long-Term Data Use: Participants should understand that biological samples and linked data contributed to biobanks may be used in future research beyond the original study purpose, subject to consent frameworks and ethics board approval. Broad consent models (used by UK Biobank and All of Us) permit wide secondary use; narrow consent models restrict data to specific pre-specified purposes. Participants can typically withdraw from future data use at any time, though previously collected data may not be retrievable from existing analyses.

Pharmacogenomic Follow-Up in Clinical Practice: Patients who receive clinical PGx test results should receive a formal results review with a clinical pharmacist, pharmacogeneticist, or appropriately trained physician. Results should be documented in the electronic health record in a format accessible to all future prescribers (a "PGx passport" concept) — since pharmacogenomic variants are lifelong and relevant to multiple future drug exposures.

Community Feedback: Ethical research practice increasingly includes returning community-level results — aggregate findings and public health implications — to the communities from which participants were drawn, as part of the reciprocal relationship between researchers and research participants.

The costs associated with pharmacogenomic testing and precision medicine vary considerably depending on the type of test, clinical indication, insurance coverage, and healthcare system.

Clinical PGx Panel Testing:

• Single-gene tests (e.g., CYP2D6 or CYP2C19 alone): Typically USD 100–300 in the United States; NHS England offers targeted PGx testing free at point of care for specific clinical indications (e.g., TPMT before thiopurine therapy, DPYD before fluoropyrimidines)
• Comprehensive PGx panels (40–80 gene panel covering major drug-metabolising enzymes and transporters): USD 300–2,000 at US reference laboratories (e.g., Mayo Clinic Laboratories, GeneSight, Genomind)
• Whole-genome sequencing with PGx interpretation: USD 800–3,000; falling rapidly with technology advances

Insurance Coverage: Coverage of PGx testing by US insurance payers is variable and rapidly evolving. Medicare covers BRCA1/2 testing and tumour genomic profiling for specific oncology indications, but coverage for preemptive (preventive) PGx panels is inconsistent. The FDA 510(k)-cleared PGx tests are increasingly covered for psychiatric medications (e.g., antidepressant selection) as evidence of clinical utility grows. In the UK, NHS covers clinically indicated PGx tests (TPMT, DPYD, UGT1A1) — broader preemptive panels remain primarily private.

Clinical Trial Participation Costs: Participants in registered interventional clinical trials generally do not pay for the investigational drug; study sponsors cover drug costs, protocol-required tests, and often provide transport reimbursement. Standard-of-care costs during a trial are typically billed to insurance in the usual manner. The US National Cancer Institute (NCI) mandates that insurance companies cover routine clinical costs of cancer trial participation.

Return on Investment: Pharmacogenomically guided prescribing has been shown to reduce adverse drug reactions, hospitalisation, and trial-and-error prescribing — representing significant healthcare system savings estimated at USD 5–10 of savings per USD 1 invested in preemptive PGx testing by several health economic analyses.

While precision medicine and diverse pharmacogenomic databases represent the gold standard, clinical practice must account for situations where population-appropriate PGx data or testing is unavailable. Several alternative strategies help clinicians make safe and effective prescribing decisions in data-sparse environments.

Therapeutic Drug Monitoring (TDM): Measuring plasma drug concentrations at steady state allows dose adjustments based on actual pharmacokinetic behaviour rather than assumed metaboliser phenotype. TDM is particularly valuable for narrow therapeutic index drugs (lithium, cyclosporin, aminoglycosides, phenytoin, tacrolimus) where population-level dosing tables may not predict individual concentration-response relationships in diverse patients.

Empirical Dose Titration: Starting at the lower end of the dosing range and titrating up based on clinical response and tolerability — the principle of "start low, go slow" — partially mitigates the risk of mis-dosing in patients whose metaboliser status is unknown. This approach is recommended by CPIC guidelines for CYP2D6-metabolised drugs when genotyping is unavailable.

Adverse Effect Monitoring Protocols: Enhanced clinical monitoring for known population-prevalent adverse effects — even when the underlying pharmacogenomic basis has not been characterised — can detect emerging toxicity before it becomes severe. Structured adverse event reporting to pharmacovigilance systems (FDA MedWatch, MHRA Yellow Card) contributes to the growing safety database for diverse populations.

Advocating for Participation in Research: The single most impactful alternative to unavailable pharmacogenomic data is encouraging participation in research by patients from underrepresented communities. Patients who speak with their physicians about clinical trial availability, contribute to biobanks, or enrol in observational studies such as All of Us directly contribute to the evidence base that will improve precision medicine for future generations from their communities.

Global Initiatives: The H3Africa Consortium, the Indian Genome Variation database, the TOPMed programme, and regional pharmacogenomics consortia in Latin America and Southeast Asia are actively building population-diverse reference databases that will progressively close the representation gap in precision medicine.