The Business Case for Psoriasis Research

This chapter is different from the rest of the book. It’s not about understanding psoriasis or managing it. It’s about making the case that investing in psoriasis research is one of the best bets in medicine, and then sketching out how someone might actually do it.

If you’re a patient, this chapter explains where your treatment money goes and why better options aren’t here yet. If you’re a researcher, it maps the white spaces. If you’re an investor or entrepreneur, it lays out the numbers.

31.1 The Market Opportunity

Start with the scale. The Global Burden of Disease Study 2021 estimates 42.98 million people are living with psoriasis worldwide (Section 3.1). The WHO puts the lifetime prevalence at 125 million. And 76% of countries lack even basic epidemiological data, so the true figures are almost certainly higher.

The global psoriasis therapeutics market was valued at approximately $21.1 billion in 2024 and is projected to reach $39.1 billion by 2030, growing at a compound annual growth rate (CAGR) of 10.9% (Grand View Research, 2025). Some forecasts are more aggressive: $57-74 billion by 2034. Either way, this is a large, growing market with durable demand. Psoriasis doesn’t go away.

The total economic burden is far larger than the therapeutics market. In the US alone, the annual cost of psoriasis has been estimated at $112-135 billion when you include direct medical costs ($52-63 billion), lost productivity ($24-35 billion), and the costs of treating comorbidities attributable to psoriatic inflammation ($36 billion) (Brezinski et al., 2015). Patients with moderate-to-severe disease lose 19-29% of their working time (Section 16.5). That’s real money, not just suffering.

Here’s the uncomfortable part. The vast majority of this spending goes to managing the disease, not curing it. Current biologic therapies cost $21,000-47,000 per patient per year in the US (Section 24.5). Multiply that by decades of treatment, and you start to see what economists call the “psoriasis tax”: the cumulative lifetime cost of a disease that the healthcare system has learned to treat but not to solve. Every dollar spent on a cure is a dollar that eventually eliminates decades of management costs.

And then there’s the access gap. Approximately 80% of global biologic sales occur in Western countries. In Central and Eastern Europe, on average only 0.25% of psoriasis patients receive biologics (Nast et al., 2015). In sub-Saharan Africa, the figure is even lower. The WHO recognised this in 2025 by adding adalimumab and ustekinumab to the Essential Medicines List for psoriasis, the first biologics for the condition to receive that designation. But getting on the list and getting into patients’ hands are two very different things.

31.2 The Drug Economics: What It Costs and What It Returns

Developing a new drug is expensive. A widely cited study estimated the mean capitalised cost at $1.3 billion per approved compound, with a median of $985 million (Wouters, McKee & Luyten, 2020). For immunology drugs specifically, the median rises to around $2.8 billion. These figures include the cost of failed candidates, since roughly 85% of drugs entering clinical trials never make it to market (Wong et al., 2019).

So what do you get for that investment in psoriasis?

Adalimumab (Humira), AbbVie’s anti-TNF antibody, is the highest-grossing pharmaceutical product in history. It peaked at $21.2 billion in global revenue in 2022 and has generated over $242 billion in cumulative sales since its 2003 launch. Psoriasis is one of several indications (alongside rheumatoid arthritis, Crohn’s disease, and others), but it’s a major revenue driver. Even in decline due to biosimilar competition, Humira still brought in $4.5 billion in 2025.

AbbVie’s successor, risankizumab (Skyrizi), is already bigger. It generated $17.6 billion in 2025, up from $11.7 billion in 2024, and AbbVie projects it will exceed $20 billion by 2027. Skyrizi treats psoriasis and Crohn’s disease, but psoriasis was the launch indication and remains a major share.

Ustekinumab (Stelara), Johnson & Johnson’s anti-IL-12/23 antibody, peaked at approximately $10.4 billion in 2024 before biosimilar entry cut sales roughly in half. Secukinumab (Cosentyx), Novartis’s anti-IL-17A antibody and the first psoriasis-first biologic, reached $6.1 billion in 2024. Guselkumab (Tremfya) crossed $5 billion in 2025, growing 65% year-on-year, with J&J projecting $10 billion in peak sales. Ixekizumab (Taltz) generates around $3.3 billion annually for Eli Lilly. And bimekizumab (Bimzelx), UCB’s dual IL-17A/F inhibitor, went from essentially zero to EUR 2.2 billion (~$2.6 billion) in 2025, with 53% of sales from psoriasis and a target of EUR 4 billion at peak.

Add it up. These seven drugs alone generated over $40 billion in 2025 revenue. Against a development cost of $1-3 billion each, the return on investment is extraordinary. Even accounting for the roughly 85% of drug candidates that fail during development, psoriasis biologics have been among the most commercially successful therapeutic investments in pharmaceutical history.

The patent cliffs are coming. Stelara’s biosimilars launched in January 2025, with at least seven FDA-approved alternatives already available. Taltz’s composition patent expires in 2027. Cosentyx faces biosimilar competition from around 2029, Tremfya from 2031, and Skyrizi from 2033. Each patent cliff creates disruption but also opportunity: biosimilar developers can enter with lower development costs, and the resulting price competition makes treatment more accessible. Adalimumab biosimilars have already saved the US health system over $11 billion in their first 18 months of competition (Biosimilars Review & Report, 2024).

31.3 How Psoriasis Compares to Other Disease Areas

Is psoriasis underfunded? Yes.

The NIH allocated approximately $14 million per year to psoriasis research in recent years. For context, diabetes receives roughly $1 billion, cancer roughly $6.3 billion, and even atopic dermatitis (eczema) receives about $28 million, double the psoriasis allocation despite a comparable disease burden. On a per-patient basis, psoriasis gets about $1.85 per affected American per year in federal research funding, compared to $28 per patient for diabetes.

A 2026 cross-sectional analysis found that psoriasis received only 0.82% of the funding expected by its disease burden (JMIR Dermatology, 2026). The National Academies independently identified psoriasis as one of three diseases with “substantially lower funding relative to disease burden” across both 2008 and 2019. And while the overall NIH budget increased 148% over the past decade, psoriasis funding at NIAMS (the National Institute of Arthritis and Musculoskeletal and Skin Diseases) actually decreased by 13% in real terms.

This isn’t because psoriasis is a solved problem. It’s because psoriasis lacks the advocacy infrastructure and political salience of cancer or diabetes. The commercial success of biologics may have paradoxically reduced urgency: if patients can achieve clear skin with existing drugs, why invest in something better? The answer, of course, is that “something better” means a cure, not just lifelong management, and that the majority of the world’s psoriasis patients can’t access existing treatments at all.

The pipeline is active, though. Seventy-five to 100 drugs are currently in clinical development for psoriasis, comparable to rheumatoid arthritis. The most exciting near-term prospect is icotrokinra (Johnson & Johnson), an oral IL-23 inhibitor that filed for FDA approval in July 2025 after Phase 3 data showed 65% of patients achieving IGA 0/1 and 50% achieving PASI 90 at week 16. If approved, it would be the first oral therapy to match biologic-level efficacy. Takeda’s zasocitinib, an oral TYK2 inhibitor, reported similarly strong Phase 3 data in December 2025. These represent exactly the kind of innovation the field needs: biologic-level results in a pill.

31.4 Failed Drugs and What They Teach Us

Not every investment pays off. Understanding the failures is as important as celebrating the successes, because each one narrowed the search space for what works.

Efalizumab (Raptiva) was an anti-CD11a monoclonal antibody that blocked T-cell activation and migration. The FDA approved it for moderate-to-severe psoriasis in October 2003. It worked, but at a terrible cost. Four patients developed progressive multifocal leukoencephalopathy (PML), a devastating brain infection caused by reactivation of the JC virus under immunosuppression. Three of them died. All four cases occurred after more than three years of continuous treatment, in a total exposed population of around 46,000. Genentech voluntarily withdrew the drug in April 2009. The lesson: targeting broad immune adhesion molecules suppresses too much. The immune system needs those pathways to surveil for opportunistic infections, and shutting them down indefinitely creates unacceptable risk.

Briakinumab (ABT-874) was Abbott’s anti-IL-12/23 antibody, targeting the same p40 subunit as ustekinumab (Stelara). In Phase 3 trials it showed strong efficacy, outperforming etanercept. But pooled safety data revealed 27 major adverse cardiovascular events (MACE) in briakinumab-treated patients, including four cardiovascular deaths, compared to zero MACE events in the placebo arms (Langley et al., 2013). Abbott withdrew regulatory applications in both the US and EU in January 2011 and never resubmitted. The drug was never approved anywhere. The lesson: even “me-too” drugs in the same mechanistic class can fail on safety. Ustekinumab, targeting the same protein, didn’t show the same cardiovascular signal. Small molecular differences matter enormously.

The industry estimates that roughly 86% of drugs entering clinical trials ultimately fail. For autoimmune and inflammatory diseases specifically, only about 15% of drugs that enter Phase 1 ever reach approval. The successful ones have to cover the cost of all the failures. This is the fundamental economics of drug development: high risk, high reward, and the rewards in psoriasis have been substantial enough to keep attracting investment despite the attrition.

31.5 The Cure Question

Here’s the question that changes everything: what would a cure for psoriasis actually be worth?

Consider the maths. If a patient starts biologic therapy at age 30 and continues for 40 years at $30,000 per year, that’s $1.2 million in lifetime treatment costs for a single person. Multiply by the roughly 43 million people currently living with the disease (even assuming only the fraction with moderate-to-severe disease would be candidates), and the aggregate lifetime treatment cost runs into the trillions. A one-time cure, even at a high price point, would represent a fraction of that.

There’s a precedent. Sofosbuvir (Sovaldi), Gilead Sciences’ cure for hepatitis C, launched at $84,000 for a 12-week course in 2013. The price sparked outrage. But it eliminated a disease that previously required lifelong management, prevented liver transplants, and averted liver cancer. Gilead’s hepatitis C franchise generated approximately $58.6 billion in cumulative revenue between 2013 and 2018, and the company’s market cap rose from $24 billion to a peak of $173 billion, roughly a sevenfold increase driven largely by a single curative therapy.

Could psoriasis follow the same playbook? The closest candidate is TRM cell targeting (Section 28.8). Tissue-resident memory T cells persist in previously affected skin even after complete clinical clearance, and they’re what drives relapse when therapy is stopped. If you could selectively deplete or reprogram these cells, you might achieve durable remission without ongoing immunosuppression. The science is real. The preclinical data are compelling. But the field is early-stage, mostly academic, and massively underfunded relative to the prize.

The hepatitis C analogy isn’t perfect. Sofosbuvir targets a virus with a defined genome; psoriasis is an autoimmune process without an external pathogen to eliminate. But the economic logic holds: a one-time intervention that eliminates decades of treatment costs can command a premium price and still save the system money. The first company to deliver a durable, drug-free remission for psoriasis will have one of the most valuable assets in immunology.

31.6 Underfunded Research Areas

Where are the biggest gaps between scientific promise and actual investment?

Oral therapies with biologic-level efficacy. Deucravacitinib (Section 24.4) was the first TYK2 inhibitor approved for psoriasis, but its efficacy falls short of the best biologics. The next generation, icotrokinra and zasocitinib, may close that gap. An oral drug that matches injectable biologic efficacy would transform access, particularly in LMICs where cold-chain logistics and injection training are barriers.

TRM cell biology. As discussed above, this is the most plausible cure pathway. Most of the work is happening in a handful of academic labs. Commercial interest is growing but hasn’t yet translated into large clinical programmes.

Microbiome therapeutics. We know the gut and skin microbiomes are altered in psoriasis (Section 6.5). We know that certain bacterial signatures correlate with disease severity and treatment response. What we don’t have is a single completed interventional trial showing that modifying the microbiome improves psoriasis. The correlational data are abundant. The causal data are almost nonexistent.

Affordable biologics for low- and middle-income countries. The WHO’s 2025 addition of adalimumab and ustekinumab to the Essential Medicines List was a symbolic step. The practical step requires biosimilar manufacturing at scale, cold-chain infrastructure, and training for healthcare providers who’ve never prescribed these drugs. This isn’t a scientific problem. It’s a logistics and economics problem, and it’s barely being addressed.

Site-specific psoriasis. Palmoplantar and nail psoriasis cause disproportionate disability relative to affected body surface area (Section 11.2, 11.3). But because BSA is small, these patients often don’t meet PASI-based eligibility criteria for biologic trials. The result: the subtypes that are hardest to treat are also the least studied.

Paediatric psoriasis. Children have fewer approved treatment options than adults, partly because of ethical barriers to enrolling minors in trials and partly because the commercial incentive is smaller. This is a gap that regulatory incentives (orphan designation, paediatric exclusivity extensions) could help close.

Predictive biomarkers. We can sequence a patient’s genome, profile their cytokines, and image their skin at cellular resolution. We still can’t reliably predict which biologic will work best for a given individual before prescribing it (Section 13). The data exist in registries like BADBIR and Corrona. Someone needs to build the models.

Repurposing candidates hiding in plain sight. We talked about mining the literature for free hypotheses. Here are some that already have published data:

Metformin (approved for type 2 diabetes) has the strongest signal. A 2023 meta-analysis of three small RCTs (148 patients total) found that metformin significantly increased the likelihood of achieving PASI 75 compared to placebo (OR 22.02, though with wide confidence intervals reflecting the small sample sizes) (Xu & Yin, 2023). The benefit appears strongest in patients with coexisting metabolic syndrome. Metformin is dirt cheap, off-patent, and has 60 years of safety data. A properly powered trial in psoriasis patients with metabolic comorbidities is the obvious next step. Nobody has run one.

Pioglitazone (another diabetes drug, a PPAR-gamma agonist) has even more trial data. Multiple small RCTs and a meta-analysis show significant PASI improvement versus placebo (Hafez et al., 2019). It works by promoting keratinocyte differentiation and reducing inflammation through a completely different pathway to existing biologics.

GLP-1 receptor agonists (semaglutide, liraglutide) are the most interesting emerging signal. A 2025 open-label RCT of semaglutide in obese type 2 diabetes patients with psoriasis found that median PASI dropped from 21 to 10 after 12 weeks, with significant reductions in IL-6 and CRP (Medic et al., 2025). But here’s the catch: a placebo-controlled trial of liraglutide in glucose-tolerant obese patients showed no benefit (Faurschou et al., 2015). The drug appears to help psoriasis only when metabolic dysfunction is present. With tens of millions of people now taking GLP-1 agonists for obesity, the epidemiological data to confirm or refute this signal is accumulating fast.

Simvastatin (a statin) showed significant PASI improvement in a meta-analysis, but the effect appears specific to simvastatin and not atorvastatin (Chua et al., 2020). Small trials, plausible mechanism (pleiotropic anti-inflammatory effects beyond lipid lowering), but needs replication.

SSRIs (antidepressants) have a population-level signal: a Swedish cohort study of nearly 70,000 psoriasis patients found that SSRI users had significantly lower odds of escalating to systemic psoriasis therapy (Thorslund et al., 2013). Is this the drug, or is it that treating depression reduces stress, which reduces flares? No RCT has tried to answer that question.

Each of these is a testable hypothesis with published supporting data, an off-patent drug, and a clear trial design. The total cost of running a properly powered proof-of-concept for any one of them, using the lean approach described in Section 31.7, would be $2-4 million. The fact that none of these trials has been run tells you everything about where the funding priorities lie.

31.7 The Startup Scenario

So you want to build a psoriasis company. What does that actually look like?

At $1 million, you’re not developing a drug. You’re developing a tool. A clinical decision support platform that predicts biologic response from baseline features. A digital therapeutic for adherence. A dermoscopy AI for primary care screening. These require software engineering more than wet-lab science, and the regulatory pathway for software-as-a-medical-device (SaMD) is faster and cheaper than for therapeutics. The FDA’s Digital Health Center of Excellence and the EU’s MDR Class IIa pathway are your friends here.

At $10 million, you can run a small clinical trial. The most capital-efficient approach is drug repurposing: taking a compound already approved for another indication and testing it in a psoriasis subtype. The 505(b)(2) regulatory pathway (FDA) or hybrid application (EMA) lets you reference the originator’s safety data, cutting years off the development timeline.

The orphan drug pathway is worth understanding in detail, because it changes the economics dramatically. In the US, any disease affecting fewer than 200,000 people qualifies for orphan drug designation (in the EU, it’s fewer than 5 in 10,000). Plaque psoriasis is too common to qualify, but rare subtypes can. GPP has orphan designation, which is how spesolimab (Spevigo) got approved with relatively small trials. Paediatric psoriasis subtypes and rare variants like erythrodermic psoriasis could also qualify. What do you get? Seven years of market exclusivity in the US (10 in the EU), during which no competitor can market the same drug for the same indication, even if your patents have expired. The FDA waives its $4.3 million filing fee. You get tax credits of up to 25% on qualifying clinical trial costs. And regulators accept smaller trials: 30-50 patients instead of the hundreds required for a common indication. The strategic play: get orphan approval for a rare subtype first, then use that revenue, safety data, and regulatory relationship to fund a larger trial in common plaque psoriasis.

What does a trial actually cost? The major line items: per-patient clinical costs (screening, drug supply, lab work, monitoring visits) run $30,000-80,000 each in dermatology trials. A contract research organisation (CRO) to manage the trial costs $500,000-1 million. GMP-grade drug supply, even for a repurposed compound, runs $500,000-1 million. Regulatory filings, ethics committees, and site agreements add $200-500,000. Statistics, data management, and medical writing cost another $200-300,000. For a conventional 60-100 patient trial, that adds up to $5-8 million. But with an orphan indication and a repurposed drug, you can shrink this considerably. A 30-patient, single-arm, open-label proof-of-concept study using a known compound can realistically be done for $2-4 million. That’s within reach of a seed-funded startup or a well-written grant.

At $50 million, you’re in the game properly. You can fund a novel mechanism through Phase 2, develop a biosimilar, or build a companion diagnostic paired with a treatment algorithm. Partnership models matter at this scale: academic spin-outs bring novel targets but need commercial expertise; licensing deals with pharma provide milestone payments to fund the next stage; CRO partnerships reduce fixed overhead.

But what if you don’t have $10 million? The budget tiers above assume the traditional model: hire a CRO, rent clinical sites, run everything yourself. There’s a leaner way.

Start with what’s free. Decades of real-world outcomes data already exist in registries like BADBIR (over 20,000 biologic-treated patients), Corrona, and PsoBest. Academic collaborations can access these. Published clinical trial data, including patient-level summaries in supplementary materials and FDA review documents, are public. Open-source genomic and transcriptomic datasets (GEO, UK Biobank) contain thousands of psoriasis samples. You can identify a drug target, validate it against existing data, and generate a repurposing hypothesis without spending a penny on wet-lab work.

Next, look for free hypotheses. Scattered across the medical literature are case reports where drugs approved for unrelated conditions unexpectedly improved someone’s psoriasis. A statin, an antihypertensive, an antidepressant. Each one is a testable signal. Mining these systematically can surface repurposing candidates that already have decades of safety data. Here’s how to actually do it.

Where to look. The primary source is PubMed, the US National Library of Medicine’s index of over 37 million biomedical citations. It’s free, it’s searchable, and it covers virtually every peer-reviewed medical journal in the world. A search like psoriasis AND ("case report" OR "case series") AND ("unexpected improvement" OR "incidental finding" OR "off-label") will return hundreds of results. PubMed Central (PMC) is the subset where full-text articles are freely available, not just abstracts. Roughly 8 million articles are on PMC, and that number grows daily as funders increasingly mandate open access.

Beyond PubMed, there are preprint servers: medRxiv and bioRxiv host papers before peer review. The quality is variable, but they’re free and often contain findings 6-12 months before the journal version appears. Google Scholar casts a wider net, indexing conference abstracts, theses, and book chapters that PubMed misses. ClinicalTrials.gov lists every registered clinical trial, including results of completed trials, some of which never make it into a journal publication.

How to get access to paywalled papers. Here’s the reality. Most medical journals charge $30-50 per article, and a serious literature review might need 200-500 papers. That’s $6,000-25,000 if you pay retail, which nobody does. Your options:

PMC and open access. Check PMC first. Many papers are freely available there even if the journal’s own website shows a paywall. Append site:pmc.ncbi.nlm.nih.gov to a Google search, or use the “Free full text” filter on PubMed.
Institutional access. If you have any university affiliation (visiting researcher, adjunct, alumni), you likely have access to the library’s journal subscriptions. Some universities offer community borrower cards for a small annual fee.
Author copies. Email the corresponding author and ask for a PDF. This works more often than you’d expect. Researchers want their work read. Many also post preprints or accepted manuscripts on their personal websites or institutional repositories.
Interlibrary loan. Most public libraries can request papers from university libraries for free. It takes a few days.
Unpaywall (unpaywall.org) is a browser extension that automatically finds free legal versions of paywalled papers. It checks PMC, institutional repositories, and author websites.

Let’s not pretend Sci-Hub doesn’t exist. It does, it hosts virtually every paywalled paper ever published, and researchers worldwide use it daily. It’s also illegal in most jurisdictions. Make your own decision.

What does a medical paper actually look like? If you’ve never read one, the format can feel alien, but it’s highly standardised. A typical research paper follows the IMRAD structure:

Abstract. A 250-word summary. Read this first. If it’s not relevant, move on.
Introduction. Why the study was done. Background, knowledge gap, hypothesis.
Methods. How the study was designed. Patient population, interventions, outcome measures, statistical analysis. This is where you assess quality: how many patients? Was there a control group? Was it randomised? Blinded?
Results. What they found. Tables and figures. The numbers.
Discussion. What the authors think the results mean. Limitations. Comparison to other studies.

Case reports are shorter and simpler. They describe something unusual that happened to one or a few patients. A typical structure: patient background, what happened, what was done, what the outcome was, why the authors think it’s noteworthy. A case report titled “Resolution of chronic plaque psoriasis during treatment with venlafaxine for depression” is exactly the kind of signal you’re looking for. One case proves nothing, but five independent case reports of the same drug improving psoriasis in different patients is a pattern worth investigating.

What signals matter for repurposing? You’re looking for:

Case reports or small case series where a drug approved for condition X unexpectedly improved psoriasis
Post-hoc analyses of large trials where psoriasis outcomes were measured as secondary endpoints
Epidemiological studies showing lower psoriasis rates in patients taking a specific medication for another condition (e.g., “metformin users have lower psoriasis incidence”)
Mechanistic papers showing a drug affects pathways relevant to psoriasis (IL-23, IL-17, TNF-alpha, Th17 differentiation) even if it was never tested in psoriasis

How to feed this to AI. You’ve got the papers. Now you want to process hundreds of them systematically. A few approaches:

Copy and paste. The simplest. Paste an abstract or full text into Claude or ChatGPT and ask it to extract the drug name, the observed effect on psoriasis, the number of patients, and the proposed mechanism. Slow for hundreds of papers, but fine for an initial pass.
PubMed API (E-utilities). The NCBI provides a free API that lets you programmatically search PubMed, retrieve abstracts in XML or JSON, and download full texts from PMC. A simple Python script can pull 10,000 abstracts in minutes. Feed those to an LLM in batch, ask it to flag any mention of psoriasis improvement, and you’ve got a shortlist in hours instead of weeks. The API documentation is at ncbi.nlm.nih.gov/books/NBK25501/.
Bulk PMC download. PMC offers bulk download of its entire open-access subset in XML format. It’s several terabytes, but if you’re serious about comprehensive mining, this is the dataset. Parse the XML, extract the text, run it through an LLM or a simpler NLP pipeline for entity extraction (drug names, disease names, outcome descriptors).
Semantic Scholar API. Semantic Scholar (by the Allen Institute for AI) provides a free API with abstracts, citation graphs, and extracted entities. Useful for finding papers that cite a key reference, which helps you trace a finding forward through the literature.
Claude with tool use. If you’re using Claude’s API, you can set up a pipeline that takes a list of PMIDs, fetches each abstract via the PubMed API, passes it to Claude with a structured extraction prompt, and collects the results in a spreadsheet. Cost: a few dollars for thousands of abstracts. Time: an afternoon of coding, then overnight to run.

The point isn’t that AI replaces reading. You still need to read the key papers carefully. But AI turns a six-month literature review into a two-week sprint, and it doesn’t miss things because it got tired on page 300.

For the actual trial, consider an investigator-initiated trial (IIT) at an academic medical centre. In this model, a university dermatology department runs the trial. They provide the investigators, the ethics approval, the clinical infrastructure, and the overhead. You provide the drug (which, if it’s a repurposed generic, might cost a few thousand dollars) and a modest research grant. Total cost to the startup: $500,000-1 million, sometimes less. The academic partner gets a publication. You get clinical data.

Non-dilutive funding can cover most or all of that. In the US, the SBIR (Small Business Innovation Research) and STTR (Small Business Technology Transfer) programmes provide up to $2 million for small companies doing early-stage research. You don’t give up equity. In the UK, Innovate UK offers similar grants. The EU’s Horizon Europe programme funds collaborative research. These programmes exist specifically for this scenario: a small company with a promising idea and limited capital.

Decentralised trial design cuts the most expensive line item: site visits. Instead of patients travelling to a hospital for every assessment, they do telemedicine visits from home. Smartphone-based photography with AI-assisted PASI scoring replaces in-clinic evaluation by a trained dermatologist. Drug is delivered to the patient’s door. The FDA’s 2023 decentralised clinical trial guidance explicitly supports this approach. It doesn’t just save money; it improves recruitment, because patients who can’t take time off work or travel long distances can now participate.

Synthetic control arms save you from randomising half your patients to placebo. By using historical placebo response rates from published trials (which are remarkably consistent for psoriasis), regulators can accept a single-arm trial design for proof-of-concept. This roughly halves your sample size and eliminates the ethical discomfort of withholding treatment.

Finally, patient advocacy partnerships with organisations like the National Psoriasis Foundation or the Psoriasis Association provide something money can’t easily buy: trust. They can help recruit patients, amplify your results, and lend credibility to a startup that doesn’t yet have a track record.

The leanest realistic path. A virtual company of two or three people identifies a repurposing candidate from registry mining and case report analysis. They secure orphan drug designation for a rare psoriasis subtype. An academic partner agrees to run a 30-patient investigator-initiated trial, funded by an SBIR grant. The trial uses decentralised design and AI-based outcome assessment. Total out-of-pocket for the startup: potentially under $500,000. That’s not theoretical. The regulatory and funding mechanisms all exist today. What’s missing is someone putting the pieces together.

Real examples. Arcutis Biotherapeutics was founded in 2016 with a focus on dermatology. They licensed roflumilast (a PDE4 inhibitor already approved orally for COPD) and reformulated it as a topical cream for psoriasis and atopic dermatitis. They IPO’d in January 2020 at a valuation of $782 million. The market was sceptical: their market cap fell to $305 million by late 2023. But the product worked, payer coverage expanded, and by December 2025 the company was worth $3.8 billion with $196.5 million in 2024 revenue and projected peak sales of $2.6-3.5 billion. The lesson: dermatology startups follow a brutal J-curve, but topical reformulation of known molecules is a viable commercial strategy.

UCB’s story with bimekizumab is different. The company invested heavily in a novel dual IL-17A/F inhibitor from first-in-human studies in 2012 through a massive Phase 3 programme (six trials across psoriasis, PsA, and axial spondyloarthritis). Nine years and hundreds of millions of euros later, they had a blockbuster: EUR 2.2 billion in 2025 revenue, growing over 200% year-on-year, with patent protection through 2037. That’s the big-bet model. It requires deep pockets and patience, but the payoff can be transformational.

How long does all this actually take? Here are two timelines. The first is the lean path: a small team repurposing an existing drug for a rare psoriasis subtype. The second is the well-funded path: a novel mechanism from scratch.

Scenario 1: Lean repurposing startup (optimistic / realistic)

Phase	Optimistic	Realistic
Literature mining, target identification, dataset analysis	2-3 months	4-6 months
Protocol development, academic partner agreements	2-3 months	4-8 months
Orphan drug designation application and decision	3-4 months	6-10 months
SBIR/grant application and award	4-6 months	6-12 months
Company incorporation, insurance, regulatory filing (IND/CTA)	1-2 months	2-4 months
Ethics approval, site setup	1-2 months	2-4 months
Patient recruitment and 16-week treatment period	4-6 months	8-14 months
Data analysis, publication, regulatory strategy for next phase	2-3 months	3-6 months
Total: first proof-of-concept result	~18-24 months	~30-48 months
Cumulative spend	$300-500K	$500K-1.5M

Many of these steps run in parallel. You can write the protocol while waiting for the grant decision, and file for orphan designation while negotiating with academic partners. The bottleneck is almost always patient recruitment, especially for rare subtypes.

Scenario 2: Novel mechanism, well-funded ($50M+)

Phase	Optimistic	Realistic
Target discovery and validation	6-12 months	12-24 months
Lead compound identification and optimisation	12-18 months	18-30 months
Preclinical studies (toxicology, pharmacokinetics)	12-18 months	18-24 months
IND/CTA filing and approval	2-4 months	3-6 months
Phase 1 (safety, healthy volunteers or patients)	6-12 months	12-18 months
Phase 2 (proof-of-concept, dose-finding)	12-18 months	18-24 months
Phase 3 (pivotal efficacy trials)	18-24 months	24-36 months
Regulatory review (NDA/BLA/MAA)	10-15 months	12-18 months
Total: from concept to market	~6-8 years	~9-13 years
Cumulative spend	$50-150M	$150-500M+

Bimekizumab took nine years from first-in-human to EU approval. Arcutis took about four years from founding to IPO, six years to meaningful revenue. These are real benchmarks, not theoretical estimates. The lean path can deliver a proof-of-concept result in under two years. Whether that result leads to a product depends on what the data show. But for under a million dollars, you’ll know if the idea has legs.

Building a team. You need fewer people than you think. The lean startup model works with four roles, none of which need to be full-time employees at the start:

A clinical dermatologist. Not as an employee. As an advisor on your scientific advisory board, paid in equity and a small retainer. They provide clinical credibility, help design the trial protocol, and connect you to the investigator network. Most academic dermatologists are happy to advise startups for 0.25-1% equity and a few hours a month.
A regulatory consultant. Don’t hire one full-time. Contract a specialist for the IND/CTA filing and any pre-submission meetings. This is project work, not ongoing. Budget $30,000-80,000 for the regulatory package for a repurposed drug.
A biostatistician. Your academic collaborator’s department almost certainly has one. If not, a contract biostatistician can design the trial’s statistical analysis plan and write the sample size justification. This is a few weeks of work, not a permanent role.
Someone who can code. Data pipelines, API integrations, literature mining scripts, AI model prototyping. If you’re reading this chapter and you’re a software engineer with psoriasis, you’re looking at your founding team.

What you don’t need: a lab (you’re repurposing, not synthesising), a manufacturing facility (you’re buying commercial drug product), a large headcount (every additional employee is burn rate), office space (work remotely). The virtual company model isn’t just cheaper. It’s faster, because you’re not spending six months hiring before you can start working.

IP strategy. “But the drug is off-patent. What’s the point?” This is the most common misconception about repurposing. The compound itself may be off-patent, but a new therapeutic use can absolutely be patented. Method-of-use patents cover the application of a known substance to a new indication, a new patient population, a new dosing regimen, or a new combination. If you discover that Drug X at a specific dose treats palmoplantar psoriasis, that method of use is patentable even if Drug X has been generic for twenty years.

Start with a provisional patent application. In the US, this costs $320 (small entity) at the USPTO. In the UK, it’s about £50 at the IPO. A provisional gives you 12 months of “patent pending” status, establishing your priority date, while you gather more data before filing the full application. File before you publish, present at a conference, or tell anyone outside your team. Public disclosure before filing destroys your patent rights in most jurisdictions.

Beyond patents, there’s data exclusivity. The FDA grants 3 years of market exclusivity for a new indication of an already-approved drug (submitted via 505(b)(2)). The EMA grants up to 1 year. This isn’t as long as a patent, but it’s automatic: you get it just by generating new clinical data. Combined with a method-of-use patent and orphan drug exclusivity (7 years US, 10 years EU), you can build a layered IP position that protects your market for a decade or more.

One more thing: do a freedom-to-operate search early. Make sure nobody else has already patented the use you’re planning to study. Patent databases (Google Patents, Espacenet, USPTO PAIR) are free to search. A patent attorney can do a formal FTO analysis for $5,000-15,000. Cheaper than finding out after you’ve run the trial.

Competitive moats. Ideas aren’t moats. Anyone can read the same case reports you did. What stops a larger company from simply copying your approach once you’ve shown it works? Here’s what actually creates defensible advantages:

Orphan drug designation. Seven years of market exclusivity in the US, ten in the EU. During that period, no other company can market the same drug for the same orphan indication, even if they run their own trial. This is the single most powerful moat available to a small company, and it’s free to apply for.
Published clinical data. Once you’ve run a trial and published positive results, replicating that data costs someone else the same time and money it cost you. Your data is a fait accompli. A competitor would need 2-4 years and millions of dollars to generate their own dataset, by which time you’ve moved to Phase 3 or been acquired.
Proprietary datasets. Exclusive access agreements with registries, hospitals, or biobanks. If you’ve negotiated sole commercial rights to mine a specific dataset for drug repurposing candidates, that’s a moat. The data existed before you, but the agreement is yours.
Validated AI models. A model trained on real-world psoriasis outcomes data that predicts treatment response is valuable, but only if it’s been validated prospectively. Anyone can train a model. Validation requires clinical data, which takes time and money. The validated model is the moat, not the architecture.
Regulatory relationships and know-how. Having been through a pre-IND meeting, knowing the specific reviewer’s concerns, understanding exactly what data package the agency expects for your indication. This tacit knowledge is hard to transfer and expensive to replicate.

What’s NOT a moat: a business plan, a pitch deck, a literature review, an untested hypothesis, or an idea you haven’t acted on. Execution is the moat.

Exit strategies. How does this end? For most successful dermatology startups, one of four ways:

Acquisition by pharma. This is the most common outcome and often the most lucrative. Large pharmaceutical companies have massive commercial infrastructure (sales forces, payer relationships, global distribution) but chronically depleted pipelines. They buy clinical-stage assets to fill gaps. In dermatology, acquisition multiples of 5-15x trailing revenue are typical for commercial-stage companies. For pre-revenue companies with strong Phase 2 or Phase 3 data, pharma pays large upfront premiums plus milestone payments. AbbVie acquired Allergan for $63 billion partly for its dermatology portfolio. UCB built bimekizumab internally, but many companies at their stage would have been acquired.

Licensing deals. You don’t have to sell the whole company. Out-license your clinical data and IP to a pharma partner for an upfront payment, development milestones, and royalties on sales. This lets you retain ownership while accessing pharma’s development and commercial muscle. Typical structures: $5-50 million upfront, $100-500 million in milestones, 5-15% royalties. You can license for specific geographies (e.g., give a partner Asia-Pacific rights, keep US/EU) or specific indications.

IPO. Arcutis went public at $782 million. This is viable but requires scale, a commercial product (or late-stage pipeline), and favourable market conditions. The biotech IPO window opens and closes unpredictably. Not a strategy you can control, but one you should be positioned for if the window opens.

Stay private and grow. If your product is a digital therapeutic, a SaaS clinical decision tool, or a direct-to-patient service, you may not need pharma at all. Build revenue, stay lean, and grow. Not every company needs an exit. Some are just businesses.

What do acquirers look for? Clean IP (no freedom-to-operate issues), reproducible clinical data (peer-reviewed publications), a clear regulatory pathway (ideally with some regulatory interaction already completed), and a differentiated mechanism (something they can’t easily replicate internally). If you’ve followed the playbook in this chapter, you’ll have all four.

The reimbursement problem. Getting a drug approved is only half the battle. Getting it paid for is the other half, and it kills more products than failed trials do. Approval means a regulator says your drug is safe and effective. Reimbursement means a payer agrees to cover it. These are different decisions made by different people using different criteria.

In the UK, NICE evaluates cost-effectiveness using a threshold of roughly £20,000-30,000 per quality-adjusted life year (QALY). If your drug costs more per QALY gained, NICE won’t recommend it, and the NHS won’t fund it. In the US, there’s no single cost-effectiveness threshold, but CMS (Medicare/Medicaid), private insurers, and pharmacy benefit managers all make independent formulary decisions. A drug can be FDA-approved and still sit on a shelf because no payer will cover it at the manufacturer’s price.

What does this mean for a startup? You need to think about health technology assessment (HTA) from day one, not after approval. Design your trial to capture the outcomes that payers care about: not just PASI improvement, but quality-of-life gains (DLQI), work productivity, reduced hospitalisations, and comorbidity prevention. If your drug costs $5,000 per year and prevents $15,000 per year in cardiovascular events and lost productivity, that’s the value story. Build the health economics model while you’re running the trial, not after.

For a repurposed generic, this problem is actually easier. If metformin improves psoriasis, the drug costs pennies. The cost-effectiveness argument writes itself. The challenge is different: how do you make money selling a $0.10 pill? The answer is usually not the drug itself but the indication, the data, and the method-of-use patent.

Common mistakes that kill dermatology startups. Having watched the space, here are the patterns that recur:

Spending on infrastructure before proof-of-concept. Hiring 20 people, leasing lab space, and buying equipment before you know if your idea works. The lean path exists for a reason. Prove the concept first, scale second.
Targeting too broad an indication. “We’re developing a treatment for psoriasis” is a $500 million trial programme. “We’re testing simvastatin as adjunctive therapy for psoriasis patients with metabolic syndrome” is a $3 million proof-of-concept. Start narrow, expand later.
Ignoring reimbursement until after approval. See above. If NICE won’t pay for it, the NHS won’t use it. If no US payer covers it, patients can’t access it. Reimbursement strategy is not a post-approval activity.
Not involving patients early enough. Patient-focused drug development isn’t just an FDA buzzword. Patients tell you what outcomes matter (itch reduction? hand clearance? being able to wear short sleeves?), what trial design they’ll actually complete (monthly hospital visits vs home monitoring), and what price they’d consider reasonable. The FDA now has formal Patient-Focused Drug Development guidance, and patient input in trial design can accelerate recruitment and strengthen regulatory submissions.
Underestimating the time between approval and first revenue. Even after approval, you need to negotiate payer contracts, build distribution, train prescribers, and generate real-world evidence. For Arcutis, it took two years after FDA approval to reach meaningful revenue. Budget for this gap.
Falling in love with the science and ignoring the business. A beautiful mechanism of action means nothing if you can’t manufacture the drug affordably, get it reimbursed, and deliver it to patients. The best drug in the world is the one patients can actually get.

Your first 90 days: a checklist. If you’ve read this far and you’re ready to start, here’s what to do.

Week 1-2: Orientation

Pick your therapeutic hypothesis. One drug, one psoriasis subtype, one patient population. Be specific.
Run a PubMed search. Is there existing evidence? Case reports, epidemiological signals, mechanistic data? If yes, you have something to build on. If no, you’re doing basic research, which is a different (harder, more expensive) game.
Check the patent landscape. Is the drug off-patent? Has anyone patented the method of use you’re considering? Google Patents, Espacenet, USPTO PAIR. Free to search.
Identify 2-3 academic dermatology departments that might collaborate. Look for investigators who have published on your target subtype.

Month 1-2: Foundation

Write a one-page concept document: the hypothesis, the existing evidence, the proposed trial design, the target patient population, the expected outcome. This is your pitch to academic partners, grant bodies, and advisors.
Recruit a clinical advisor. Email a dermatologist who’s published on your subtype. Offer advisory board membership. Most will take a call.
Request a pre-IND meeting with the FDA (or scientific advice from the EMA/MHRA). This is free, and the agency’s feedback tells you exactly what data they’ll need. It’s also a signal of seriousness to future partners and investors.
File a provisional patent application if your hypothesis is novel. $320 (US). Do this before talking to anyone outside your team.

Month 2-3: Funding and partnerships

Apply for an SBIR/STTR grant (US), Innovate UK grant (UK), or equivalent. Write the application around your concept document and the regulatory feedback.
Approach your target academic partner formally. Propose an investigator-initiated trial. Bring the protocol outline and the funding plan.
Incorporate the company. You need the legal entity for the grant application and the partnership agreement. A standard limited company or LLC is fine. Don’t over-engineer the corporate structure.
Engage a regulatory consultant (contract, not hire) to prepare the IND/CTA package.

By the end of 90 days, you should have: a testable hypothesis with supporting data, a provisional patent filing, a clinical advisor, a regulatory strategy informed by agency feedback, a grant application in process, and an academic partner in discussion. Total spend: under $5,000 (patent filing, incorporation fees, and coffee). Everything else is your time.

31.8 What’s Regulated and What Isn’t

If you’ve read this far and you’re thinking “can I actually do any of this?”, the answer is: more than you’d expect. The regulatory landscape for drug development has a reputation for being impenetrable, but the boundaries between what requires authorisation and what doesn’t are surprisingly clear.

Anyone can do these things without regulatory approval:

Computational research. Bioinformatics, AI/ML modelling, target identification, molecular docking, literature mining. No regulatory body in any jurisdiction regulates what you do on a computer with publicly available data. Open-source datasets like GEO (Gene Expression Omnibus) and the UK Biobank contain thousands of psoriasis-relevant samples. You can build a drug repurposing model on a laptop.
Analysis of existing de-identified data. Secondary analysis of registry data, electronic health records, or published trial data doesn’t require clinical trial authorisation. You’ll need a data use agreement and often ethics committee approval (or an exemption determination), but not an IND or equivalent.
Observational studies. If you’re observing what happens in routine clinical practice without assigning an intervention, you don’t need clinical trial authorisation in any of the major jurisdictions. Ethics approval, yes. Trial authorisation, no.
Protocol development. You can write a full clinical trial protocol, statistical analysis plan, and investigator brochure before you’ve formed a company, let alone filed with a regulator.

You need trial authorisation (IND, CTA, or equivalent) when you:

Assign patients to a specific intervention as part of a research protocol
Test an unapproved drug in humans
Study an approved drug outside its licensed indications in a formal trial setting

Note the nuance: a physician prescribing an approved drug off-label in clinical practice is not conducting a clinical trial. That’s legal everywhere. But if you’re systematically collecting data according to a protocol with the intent to generate generalisable knowledge, you’ve crossed into regulated territory.

Who can sponsor a clinical trial? In every major jurisdiction, the answer is: almost anyone. The FDA, EMA, MHRA, HSA (Singapore), and Israeli Ministry of Health all allow individuals, limited companies, academic institutions, and corporations to act as trial sponsors. You don’t need to be a pharmaceutical company. In the US, a single physician can serve as both sponsor and investigator under the “sponsor-investigator” designation (21 CFR 312.3). In the EU, you need to be established in the EU or appoint a legal representative there. Same principle in the UK, Singapore, and Israel for foreign sponsors.

What does the regulatory pathway actually look like?

	US (FDA)	EU (EMA/CTIS)	UK (MHRA)	Singapore (HSA)	Israel (MoH)
Application	IND	CTA via CTIS	CTA	CTC or CTN	MoH application
Review time	30 days	~60 days	30-60 days	30-45 days	30-45 days
Ethics	IRB (separate)	Integrated in CTA	REC via IRAS	IRB (separate)	Helsinki Committee
Insurance required	Practically yes	Legally yes	Legally yes	Yes	Yes
Time to first patient	2-4 months	3-6 months	3-5 months	2-4 months	3-5 months

Two to six months from application to dosing your first patient. That’s faster than most people assume.

Do you need your own GMP manufacturing? No. If you’re repurposing an already-approved drug, you can use the commercially available product directly. Buy it from a pharmacy or wholesaler. The drug is already manufactured to GMP standards. This is one of the biggest cost advantages of repurposing: you skip the entire manufacturing infrastructure. The only exception is if you need to modify the product (repackaging for blinding, creating matched placebos), in which case the repackaging operation needs to be GMP-compliant, but clinical packaging companies handle this cheaply.

Data protection. You’ll need to comply with GDPR (EU/UK), HIPAA (US), or equivalent local laws. This sounds daunting for a small company, but in practice you use established HIPAA/GDPR-compliant cloud infrastructure and an off-the-shelf electronic data capture system like REDCap (free for academic use) or Castor (commercial). In the EU, you’ll likely need a Data Protection Officer, but this can be outsourced. The startup itself doesn’t need a special certification.

What can you do before forming a company? Quite a lot. All the computational research, literature analysis, protocol development, dataset analysis, and academic collaboration discussions can happen pre-incorporation. You can even request a pre-IND meeting with the FDA or scientific advice from the EMA as a prospective sponsor. The company becomes necessary when you start entering contracts with clinical sites, obtaining trial insurance, and submitting regulatory applications. For the lean startup path described in Section 31.7, that means you can do months of foundational work as an individual researcher or academic affiliate, incorporate only when you’re ready to pull the trigger on the actual trial, and keep your burn rate at zero until then.

31.9 How AI Can Accelerate Psoriasis Research

AI won’t replace the wet lab, but it can make every stage of drug development faster and cheaper.

Drug discovery. The traditional approach to finding new drug targets takes years of painstaking biology. AI models trained on multi-omics data (genomics, transcriptomics, proteomics, metabolomics) can identify candidate targets in months. Molecular simulation tools like AlphaFold predict protein structures, enabling in silico screening of potential binding compounds before synthesising a single molecule. Several AI-native drug discovery companies (Recursion, Insilico Medicine, Isomorphic Labs) are already applying these approaches to autoimmune diseases.

Clinical trial design. AI can improve trials in three ways. First, synthetic control arms: using historical data from registries and previous trials to simulate a placebo group, reducing the number of patients who receive no treatment. Second, adaptive trial designs: AI-powered interim analyses that adjust dosing, endpoints, or sample sizes in real time. Third, patient matching: NLP algorithms that screen electronic health records to identify eligible patients and predict their likelihood of completing the trial.

Diagnostics. Automated PASI scoring using computer vision is already being validated (Section 29). But the bigger opportunity is earlier and more accurate diagnosis. Dermoscopy AI could flag psoriasis in primary care settings where GPs see it infrequently. Imaging-based screening could detect subclinical psoriatic arthritis before joint damage occurs, since early treatment of PsA dramatically improves outcomes (Section 15.5).

Real-world evidence. BADBIR has data on over 20,000 biologic-treated psoriasis patients. PsoBest, Corrona, and other registries hold similarly rich datasets. NLP tools can extract structured outcomes from unstructured clinical notes in electronic health records. Linking these datasets and applying machine learning to identify predictors of treatment response, adverse events, and long-term outcomes is possibly the highest-impact near-term application of AI in psoriasis.

Personalised treatment selection. Can you predict which biologic will work best for a specific patient? The answer is: probably, but nobody has built the model yet. Baseline clinical features (PASI, age, weight, previous treatments), genetic markers (HLA-C*06:02 status, IL-23R variants), serum biomarkers (IL-17A, IL-22, beta-defensin levels), and microbiome profiles all carry predictive signal. Integrating these into a decision-support tool would reduce the current trial-and-error approach to biologic selection, which costs both time and money.

31.10 Where to Start: Countries and Regions

Location matters. The regulatory environment, market size, talent pool, and funding ecosystem vary enormously across countries. Here’s how the major options compare.

United States. The largest market, the highest drug prices, and the deepest talent pool. FDA approval is the gold standard: expensive ($4.3 million filing fee alone for a clinical data application) and slow (10-12 months standard review), but it unlocks the most valuable market. The NIH funds $14 million in psoriasis-specific research annually, and the VC ecosystem in Boston/Cambridge and San Francisco is the most active in biotech globally. Downside: the cost of running clinical trials in the US is the highest in the world, and the regulatory pathway is the most demanding. Best for: well-funded companies targeting the premium end of the market.

European Union. The EMA’s centralised procedure grants market access across all 27 member states in a single application (210-day assessment timeline, typically 12-15 months total). Drug prices are lower than the US but markets are large. EU Horizon Europe provides non-dilutive research grants. Basel (Switzerland), London/Oxford (UK), and Munich/Berlin (Germany) are major biotech hubs. Downside: pricing and reimbursement decisions are made country-by-country after EMA approval, which can delay actual market access. Best for: companies with a broad geographic strategy and willingness to navigate fragmented payer systems.

United Kingdom. Post-Brexit, the MHRA has launched the Innovative Licensing and Access Pathway (ILAP), which integrates regulatory approval with NICE health technology assessment and NHS commercial negotiations into a single coordinated process. The UK has world-class dermatology research (King’s College London, University of Manchester), the NIHR funds clinical trials, and Innovate UK provides startup grants. BADBIR, the largest biologic psoriasis registry, is UK-based. Downside: the UK market alone is small (~67 million people), and NICE’s cost-effectiveness threshold limits pricing. Best for: companies that want an efficient regulatory-to-reimbursement pathway and access to rich clinical data.

Israel. A 2025 regulatory reform introduced three reliance-based registration tracks, including a fast-track pathway offering 70-day registration for drugs already approved by the FDA or EMA. Israel has a thriving biotech sector (Tel Aviv), mandatory universal health coverage that facilitates clinical trials, and excellent academic institutions. Downside: small domestic market (~9.5 million people), though it serves as a gateway to broader international registration. Best for: companies seeking fast, low-cost regulatory approval as a stepping stone to larger markets.

Singapore. The Health Sciences Authority (HSA) offers a 180-day standard review and expedited pathways for breakthrough therapies. Singapore is a biotech hub for the Asia-Pacific region with strong IP protection, generous government R&D incentives (A*STAR), and proximity to the massive Chinese, Indian, and Southeast Asian markets. Downside: small domestic market, high operating costs. Best for: companies targeting Asia-Pacific markets who want a well-regulated base.

China and India. Together, these countries have several hundred million potential psoriasis patients and rapidly growing pharmaceutical markets. Regulatory barriers have been decreasing: China’s NMPA has shortened approval timelines significantly since 2017, and India’s CDSCO has expedited pathways. Manufacturing costs are a fraction of Western equivalents. Downside: IP protection remains weaker, pricing power is limited, and navigating the regulatory systems requires local expertise. Best for: biosimilar developers and companies targeting affordable access.

The bottom line. There’s no single best country. A psoriasis startup’s optimal strategy depends on what it’s building. For a digital health product, start in the US or UK where the regulatory pathway for SaMD is clearest. For a novel therapeutic, consider first-in-human studies in a country with fast ethics approval (UK, Israel, Singapore) and a pivotal trial in the US or EU for commercial impact. For a biosimilar, India or China offers the lowest manufacturing costs. For a repurposed drug targeting a rare subtype, the FDA’s orphan drug pathway provides the best incentive package. The smart play is usually a multi-jurisdiction strategy, starting where approval is fastest and cheapest, then expanding to where the money is.