For most of modern medicine's history, treatment decisions have been guided by population averages. A physician diagnoses hypertension, consults a guideline, and prescribes a first-line agent that clinical trials showed worked for most patients in a study cohort. The problem is that "most patients" is not every patient. Drug response varies enormously across individuals, shaped by genetics, environment, diet, microbiome, and prior exposures. Precision medicine is the effort to move past those averages and match patients to the therapies most likely to benefit them while minimising harm.
This guide covers what precision medicine is, how it works at a technical level, where it is already changing clinical practice, and why the field's dependence on genomic data from a narrow slice of the global population threatens to limit its reach.
Precision Medicine Defined: Beyond One-Size-Fits-All
Precision medicine is a clinical and research framework that uses molecular, environmental, and behavioural data to stratify patients into subgroups that respond differently to treatment. Instead of prescribing the same drug at the same dose to every patient with a given diagnosis, precision medicine asks: which patients are likely to respond, which are likely to experience adverse effects, and which need a different approach entirely?
The concept has a long history. Blood typing before transfusion, introduced in the early 20th century, is a rudimentary form of precision medicine. What has changed is scale. The completion of the Human Genome Project in 2003, the subsequent collapse in sequencing costs (from billions of dollars for the first human genome to below 200 dollars today), and the growth of large biobanks and electronic health record systems have made it feasible to characterise patients at a molecular level and link those profiles to clinical outcomes.
The term gained mainstream policy recognition in 2015 when President Obama launched the Precision Medicine Initiative, committing 215 million dollars in initial funding and establishing the All of Us Research Program through the National Institutes of Health (NIH). That programme aims to enroll at least one million participants from diverse backgrounds, collecting genomic data, health records, lifestyle surveys, and wearable device data to build a research resource that reflects the full range of human variation.
How Precision Medicine Works: From Genomics to Treatment
At its core, precision medicine relies on identifying biomarkers: measurable biological characteristics that predict how a patient will respond to a specific intervention. The most mature biomarker category is genomic, but the field increasingly incorporates proteomic, metabolomic, transcriptomic, and environmental data.
Genomics and Pharmacogenomics
Pharmacogenomics studies how genetic variation affects drug metabolism, efficacy, and toxicity. It is the most clinically advanced branch of precision medicine. The core insight is straightforward: enzymes encoded by genes in the cytochrome P450 family (and other drug-metabolising pathways) vary substantially across individuals, and that variation has direct consequences for drug dosing.
Take CYP2D6, a liver enzyme responsible for metabolising approximately 25 percent of all prescribed drugs, including codeine, tamoxifen, and several antidepressants. Individuals carry different CYP2D6 allele variants, which place them into one of four metaboliser categories: poor, intermediate, extensive (normal), or ultra-rapid.
A poor metaboliser given codeine will fail to convert it to morphine, the active analgesic compound, and will experience no pain relief. An ultra-rapid metaboliser given the same dose will convert codeine to morphine far faster and in greater quantities than intended, risking respiratory depression. This is a potentially fatal outcome that has been documented in case reports, including in children. Same drug, same dose, radically different clinical outcomes, determined almost entirely by a single gene.
The Clinical Pharmacogenetics Implementation Consortium (CPIC) now publishes evidence-based guidelines for dozens of gene-drug pairs, covering therapeutics in oncology, cardiology, psychiatry, and pain management. The US Food and Drug Administration (FDA) lists pharmacogenomic biomarkers in the labelling of more than 300 drugs.
Multi-Omics Integration
Genomics alone does not capture the full picture. A patient's transcriptome (which genes are actively expressed), proteome (which proteins are present and at what concentrations), and metabolome (the small molecules produced by cellular processes) all contribute to disease risk and treatment response.
Multi-omics approaches, integrating data from two or more of these layers, are increasingly used in oncology. Tumour profiling may combine whole-exome sequencing with RNA expression analysis to identify which mutations are present and which are actively driving tumour growth. This distinction matters for therapy selection: a mutation that is present in the DNA but not expressed at the RNA level may be a passenger rather than a driver, and targeting it would be ineffective.
Precision Medicine vs Personalised Medicine: What Is the Difference?
The terms precision medicine and personalised medicine are often used interchangeably, but they carry different connotations in the research community. The distinction is worth clarifying because it affects how stakeholders think about data requirements, scalability, and evidence standards.
Personalised medicine implies an entirely individualised approach, a unique treatment for each patient. In its purest form, this describes therapies like chimeric antigen receptor T-cell (CAR-T) therapy, where a patient's own immune cells are extracted, genetically engineered, and re-infused. Each treatment is literally manufactured for one person.
Precision medicine is broader and more pragmatic. It groups patients into subpopulations based on shared biological characteristics and matches those groups to targeted therapies. The individual patient is still the unit of care, but the evidence base is built at the subgroup level. This is why precision medicine is sometimes described as "the right treatment, for the right patient, at the right time." That framing is aspirational, but it is grounded in stratification rather than true one-to-one customisation.
The National Research Council's 2011 report, "Toward Precision Medicine," deliberately chose the term precision over personalised to avoid the implication that every treatment would be unique. In practice, most clinical applications fall under the precision umbrella: companion diagnostics, biomarker-guided dosing, and targeted therapies validated in molecularly defined patient subgroups.
Key Applications Across Therapeutic Areas
Oncology
Cancer treatment has been the primary proving ground for precision medicine. Targeted therapies such as imatinib (Gleevec) for BCR-ABL-positive chronic myeloid leukaemia and trastuzumab (Herceptin) for HER2-positive breast cancer demonstrated decades ago that matching therapy to tumour biology could dramatically improve outcomes. Imatinib raised five-year survival in CML from roughly 30 percent to above 85 percent.
More recently, comprehensive genomic profiling, sequencing hundreds of cancer-related genes simultaneously, has become standard of care in many tumour types. The FDA has approved several tumour-agnostic therapies, meaning drugs that target a specific molecular alteration regardless of where in the body the cancer originated. Pembrolizumab (Keytruda), for instance, was approved for any solid tumour with high microsatellite instability (MSI-high), irrespective of tissue type. This represents a shift from organ-based to biology-based oncology.
Cardiovascular Disease
Pharmacogenomics is reshaping cardiovascular prescribing. Clopidogrel (Plavix), one of the most widely prescribed antiplatelet agents, is a prodrug that requires activation by the CYP2C19 enzyme. Patients carrying loss-of-function CYP2C19 alleles (roughly 2 to 15 percent of European-ancestry populations, and up to 40 percent in some East Asian populations) metabolise clopidogrel poorly, leaving them at elevated risk of cardiovascular events despite treatment. CPIC guidelines recommend alternative agents such as ticagrelor or prasugrel for these patients.
Warfarin dosing is another well-established example. Variants in VKORC1 and CYP2C9 explain a substantial proportion of the inter-individual variation in warfarin dose requirements. Genotype-guided dosing algorithms, validated in multiple trials, reduce the time to therapeutic range and lower the risk of bleeding complications.
Rare Diseases
Precision medicine has had a strong impact on rare disease diagnosis and treatment. Whole-exome and whole-genome sequencing can now identify the genetic cause of a rare disease in approximately 25 to 40 percent of previously undiagnosed cases, often ending years-long diagnostic odysseys. Programmes such as the UK's 100,000 Genomes Project and the Undiagnosed Diseases Network (UDN) in the United States have demonstrated the clinical utility of this approach at scale.
Gene therapies, perhaps the most direct application of precision medicine, are now approved for conditions including spinal muscular atrophy (onasemnogene abeparvovec) and certain inherited retinal dystrophies (voretigene neparvovec).
The Role of Real-World Data in Precision Medicine
Clinical trials remain the gold standard for establishing the efficacy of precision medicine interventions, but they have inherent limitations: controlled environments, narrow eligibility criteria, and relatively short follow-up periods. Real-world evidence (RWE) derived from electronic health records, insurance claims, patient registries, and wearable devices fills important gaps.
Real-world data is particularly valuable in precision medicine for three reasons:
Validating biomarker-outcome associations at scale. A clinical trial may demonstrate that a biomarker predicts treatment response in a few hundred patients. Real-world data can confirm whether that association holds across tens of thousands of patients in routine clinical practice, with all the complexity that entails: comorbidities, polypharmacy, variable adherence.
Extending evidence to underrepresented populations. Clinical trials disproportionately enroll white, male, middle-aged patients from high-income countries. Real-world data drawn from diverse clinical trial populations can show whether a biomarker validated in one population performs equivalently in others.
Supporting post-market surveillance for targeted therapies. Precision medicine drugs often receive accelerated approval based on surrogate endpoints in small, biomarker-selected populations. Real-world data provides the long-term outcomes evidence that regulators increasingly require for continued approval.
Why Genetic Diversity Matters: The Africa Gap
Precision medicine's promise depends on a basic assumption: that the biomarkers used to stratify patients are valid across the populations being treated. That assumption is under strain.
As of the most recent analyses, the vast majority of participants in genome-wide association studies (GWAS) have been of European descent, with individuals of African ancestry comprising roughly two percent. This skew has real consequences. Polygenic risk scores, algorithms that estimate disease risk based on the combined effect of many genetic variants, perform significantly worse when applied to non-European populations. A cardiovascular risk score developed primarily from European-ancestry data may misclassify African-descent patients, leading to under-treatment of those at genuine risk or over-treatment of those who are not.
The underlying biology makes this gap particularly consequential. African populations harbour the greatest genetic diversity of any continental group, a consequence of humanity's evolutionary origins on the continent. This diversity means that allele frequencies, linkage disequilibrium patterns, and gene-gene interactions differ in ways that cannot be extrapolated from European-ancestry data alone.
Pharmacogenomic variation follows the same pattern. The distribution of CYP2D6 metaboliser phenotypes varies significantly across African populations: certain alleles that are rare in European populations occur at much higher frequencies in West African and East African groups. Countries like Nigeria and Kenya, with their growing digital health infrastructure, are increasingly positioned to contribute the clinical data needed to validate pharmacogenomic findings across diverse populations. Prescribing guidance developed without accounting for this variation risks both inefficacy and toxicity in African patients.
Several initiatives are working to close this gap. The Human Heredity and Health in Africa (H3Africa) Consortium, funded by the NIH and the Wellcome Trust, has supported genomic research across more than 30 African countries since 2012, building both datasets and local research capacity. The African Genome Variation Project has provided finer-grained maps of genetic diversity across the continent. Uganda's MRC/UVRI and LSHTM research unit and South Africa's H3ABioNet bioinformatics network are training the next generation of African genomicists.
Yet gaps persist, particularly in linking genomic data to clinical outcomes. Genomic sequence data alone has limited precision medicine value without corresponding phenotypic data: diagnoses, lab results, treatment histories, and outcomes. This is where platforms like Kapsule play a role, providing access to structured, de-identified clinical records that can be linked to genomic datasets, enabling the population-diverse research that precision medicine requires to work globally.
The Future of Precision Medicine: Trends to Watch
Artificial Intelligence and Machine Learning
AI is increasingly central to precision medicine as a tool for pattern recognition in high-dimensional datasets. Machine learning algorithms can identify complex multi-omics signatures that predict treatment response more accurately than any single biomarker. Deep learning models are being applied to histopathology images, radiology scans, and longitudinal EHR data to generate diagnostic and prognostic predictions. The main challenge is validation: AI models trained on data from one population may not generalise to another, which reinforces the need for diverse training datasets.
Liquid Biopsies
Circulating tumour DNA (ctDNA) analysis, detecting cancer-derived DNA fragments in a blood draw rather than a tissue biopsy, is expanding precision oncology's reach. Liquid biopsies can detect minimal residual disease after surgery, monitor treatment response in near-real-time, and identify resistance mutations as they emerge. For patients in regions where tissue biopsy infrastructure is limited, liquid biopsies offer a less invasive path to molecular profiling.
Decentralised Clinical Trials and Digital Biomarkers
The shift toward decentralised trial designs, using wearable devices, remote monitoring, and telemedicine to collect data outside traditional clinical sites, is generating new categories of biomarkers. Continuous glucose monitoring data, actigraphy-derived sleep patterns, and smartphone-based cognitive assessments are all being evaluated as digital biomarkers that could inform precision dosing and treatment selection. These approaches have the potential to make precision medicine research more accessible to participants in low- and middle-income countries, where travel to centralised trial sites can be a significant barrier to enrollment.
Expanding Pharmacogenomic Implementation
Despite strong evidence and published guidelines, pharmacogenomic testing remains underutilised in routine clinical practice. Fewer than five percent of patients prescribed a drug with a CPIC guideline receive pre-emptive pharmacogenomic testing. Implementation science initiatives, including the Electronic Medical Records and Genomics (eMERGE) Network and the Implementing GeNomics In PracTicE (IGNITE) Pragmatic Trials Network, are testing strategies to embed pharmacogenomic decision support into EHR systems. As these programmes demonstrate clinical utility and cost-effectiveness, adoption is expected to accelerate.
Population-Scale Biobanks
Large-scale biobanks continue to expand. The UK Biobank, with approximately 500,000 participants and deep phenotyping data, has become one of the most productive research resources in history, supporting thousands of publications. The NIH's All of Us programme is building a comparably diverse resource in the United States, with a deliberate focus on enrolling communities historically underrepresented in biomedical research, including African American, Latino, and rural populations. Emerging biobanks in Africa, the Middle East, and Southeast Asia will be necessary to ensure that the next generation of precision medicine tools works equitably across the global population.
Getting Started with Precision Medicine: A Practical Summary
For clinicians, researchers, and industry professionals looking to deepen their engagement with precision medicine, the field can be distilled to a few core principles:
Stratification over standardisation. The goal is to identify the biological subgroups within a disease and match each to the most effective intervention at the subgroup level.
Evidence at the biomarker level. Precision medicine requires robust evidence linking biomarkers to clinical outcomes. This evidence comes from clinical trials, genomic studies, and increasingly from real-world data.
Diversity as a prerequisite. Biomarkers validated in one population cannot be assumed to work in another. Global precision medicine depends on research conducted across genetically diverse populations, and that requires data infrastructure in regions that have historically been excluded from genomic research.
Infrastructure matters. The clinical impact of precision medicine depends on the systems that connect genomic insights to clinical decisions: EHR integration, clinical decision support, pharmacogenomic alerts, and the data pipelines that feed them. Without that infrastructure, even the best biomarker discoveries stay in academic journals.
Precision medicine is already a clinical reality in oncology, cardiology, rare diseases, and a growing number of therapeutic areas. Its trajectory depends on whether the field can build the diverse, high-quality data foundations necessary to ensure that its benefits reach all patients, including those whose biology remains underrepresented in existing databases.
Kapsule provides access to structured, de-identified health records covering over 75 million patients across 9 African countries. Contact our team to discuss how diverse African patient data can support your precision medicine research and development.
This article is intended for informational purposes only and does not constitute legal, medical, or regulatory advice. Readers should obtain independent professional counsel for their specific circumstances.