Key Takeaways
- A meta-analysis of 1,800 Phase III studies spanning 2004 to 2024 finds that biomarker-selected programmes are 2.4x more likely to reach approval than unselected broad-market trials, but the ROI gap is narrowing as companion diagnostic costs rise.
- Precision medicine programmes now average $312 million more in development costs per approved asset than broad-market equivalents, a 38% increase since 2018, driven largely by companion diagnostic development and smaller eligible patient populations.
- Broad-market trials in mid-sized therapeutic areas are outperforming precision approaches on 10-year net present value in 4 out of 9 disease categories studied, including certain cardiovascular and metabolic indications.
- Portfolio modelling by 47 leading biopharma organisations indicates that a blended 60/40 split favouring precision programmes delivers superior risk-adjusted returns compared to pure-play precision or pure-play broad-market strategies.
For the better part of a decade, the industry consensus has been settled: biomarker-selected precision programmes win, full stop. Higher approval rates, cleaner trial data, faster regulatory review. The financial logic seemed airtight. A landmark meta-analysis published this month by the Centre for Medicines Outcomes Research, covering 1,800 Phase III studies conducted between 2004 and 2024, now introduces a more uncomfortable set of questions. The precision medicine premium is real, but it is compressing. And for a meaningful slice of therapeutic categories, broad-market approaches are quietly outperforming on the metric that matters most to investors: long-term net present value.
The Approval Rate Advantage Is Real, but the Cost Structure Has Changed
The meta-analysis confirms that biomarker-selected programmes retain a substantial approval rate advantage: 2.4x versus unselected broad-market trials across the full 20-year dataset. That finding is not in dispute. What has shifted is the economics underpinning it. Average companion diagnostic development costs have risen 61% since 2019, reaching a median of $84 million per programme. Combined with smaller eligible patient populations, the per-approved-asset cost for precision medicine programmes now averages $312 million more than broad-market equivalents, a 38% increase over the past six years.
The revenue side of the equation is also tightening. Payer pressure on ultra-targeted therapies has intensified across the UK, Germany, and the United States, with health technology assessment bodies increasingly scrutinising cost-per-QALY thresholds for treatments serving populations under 50,000 eligible patients. The result is that the premium pricing that once made small addressable markets financially viable is under sustained institutional pressure.
Broad-Market Trials Are Outperforming in More Categories Than Sponsors Expected
Perhaps the most provocative finding in the report is this: in 4 of the 9 disease categories studied, including type 2 diabetes management, heart failure with preserved ejection fraction, and two classes of chronic respiratory disease, broad-market trials delivered superior 10-year net present value compared to precision approaches. The margin was not marginal. In cardiovascular indications, broad-market programmes outperformed on NPV by an average of 22% over the study horizon, largely because large addressable populations and established commercial infrastructure offset lower per-patient pricing power.
Researchers caution against reading this as a blanket endorsement of the unselected approach. In oncology, CNS, and rare disease categories, precision programmes retained a commanding NPV advantage, averaging 34% better returns than broad-market comparators. The nuance is therapeutic category specificity: the financial case for precision medicine is not uniformly strong across the portfolio, and assuming it is has cost sponsors dearly.
"We ran our portfolio assumptions through the meta-analysis framework and found three programmes where we had been systematically overvaluing the precision premium. The companion diagnostic cost assumptions alone were off by a factor we would have considered implausible two years ago. This changes how we model Phase II go/no-go decisions materially."
Chief Development Officer at a top-20 global pharmaceutical company (survey respondent)
What Portfolio Construction Looks Like When the Assumptions Shift
The report surveyed 47 biopharma organisations on how they currently allocate development spend between precision and broad-market programmes. The modal response was a 70/30 split favouring precision, broadly consistent with industry rhetoric over the past five years. But when researchers modelled risk-adjusted returns across 10,000 simulated portfolio scenarios, the optimal split shifted to approximately 60/40 in favour of precision, with the remaining 40% weighted towards broad-market programmes in cardiovascular, metabolic, and select infectious disease categories.
Three portfolio principles emerged from the modelling as statistically robust across scenario assumptions:
- Category-first allocation: Precision medicine delivers superior risk-adjusted ROI in oncology, rare disease, and CNS. Broad-market approaches consistently outperform in cardiovascular, metabolic, and chronic respiratory indications where population size and established infrastructure reduce commercialisation risk.
- Companion diagnostic cost as a first-order variable: Programmes where companion diagnostic development is projected to exceed $100 million require a minimum addressable population of 80,000 to 120,000 eligible patients to sustain positive NPV under current payer pricing environments in the US and EU.
- Phase II biomarker flexibility: Programmes that retain the option to expand eligibility criteria at Phase III, pending interim data, outperform locked precision designs by an average of 18% on risk-adjusted NPV, because they preserve commercial optionality without sacrificing early-phase signal quality.
The broader strategic implication is not that precision medicine has lost its edge. It remains the single most reliable lever for improving approval probability in the hardest-to-treat indications, and the regulatory environment in oncology and rare disease continues to favour biomarker-selected programmes. But the industry's default assumption, that precision always means better ROI, is no longer supportable as a blanket portfolio rule. Sponsors heading into the second half of 2026 face a more demanding modelling task: matching therapeutic category characteristics to development strategy with far greater granularity than the precision-versus-broad binary has historically required. Those who do that work rigorously will allocate capital more effectively. Those who don't will keep learning it the expensive way.
