Patient-Based Forecasting for Rare Disease BD&L Deals: A Practical Guide

Rare disease forecasting is where pharmaceutical analysts earn their credibility — or lose it.

The numbers are small. The data is scarce. The pricing is high. And the margin for error in your assumptions is enormous. A 10% change in your eligible patient pool estimate can move the revenue model by tens of millions of dollars. A misunderstanding of the diagnosis journey can make your forecast look credible on paper but completely disconnected from commercial reality.

Drawing on experience building rare disease revenue models across oncology, nephrology, and specialty care, here is the patient-based forecasting framework we use for rare disease BD&L deal evaluation.

Why Rare Disease Forecasting Is Different

Before getting into the methodology, it is worth understanding what makes rare disease BD&L modeling distinct from standard pharmaceutical forecasting.

1. Small patient populations amplify every assumption error.In a large primary care indication, being off by 5% on your diagnosis rate estimate is a rounding error. In a rare disease with a total US prevalence of 8,000 patients, that same 5% error changes your eligible patient pool by 400 patients — which at $400,000 annual drug cost per patient is $160 million in revenue difference. Small errors have large consequences.

2. Epidemiology data is often incomplete or unreliable.Rare diseases are, by definition, under-researched. Published prevalence estimates vary widely across studies, registries often undercount because many patients are never diagnosed, and some conditions lack any prospective epidemiology data at all.

3. Diagnosis rates are typically very low at baseline.Many rare diseases have average diagnosis delays of 5–7 years. A significant proportion of patients with the disease never receive a correct diagnosis in their lifetime. This means the "diagnosed prevalence" — the number of patients who are actually identified and potentially reachable by a treatment — can be a fraction of the true disease prevalence.

4. Pricing dynamics are fundamentally different.Orphan drugs command significantly higher prices than standard pharmaceutical products. Annual treatment costs of $100,000–$500,000+ are common. In ultra-rare conditions, seven-figure annual costs per patient exist. This pricing reality must be reflected in your model and defended to deal committees unfamiliar with rare disease commercial economics.

5. Market share dynamics are concentrated.In rare diseases, peak market share assumptions that would be considered unrealistic in a large indication are entirely credible. First-in-class drugs entering rare disease markets with no existing standard of care can achieve 70–90%+ market penetration. This is not optimism — it is the commercial reality of unmet need.

The Patient-Based Forecasting Framework for Rare Disease

A patient-based model works from the disease down to the treated patient population, step by step. Here is the funnel you need to build.

Step 1: Establish True Prevalence (or Incidence)

Start with the total number of people who have the disease, diagnosed or not.

For rare diseases, this data comes from multiple imperfect sources that you will need to triangulate:

• Published epidemiology studies — peer-reviewed literature on disease prevalence or incidence. Use the most recent, most geographically relevant studies. Check study design and sample size carefully; small registry-based studies can produce wide confidence intervals.
• Natural history studies — often the most rigorous source for rare disease epidemiology, particularly for conditions where genetic testing is the primary diagnostic pathway.
• Patient advocacy organizations — groups like the National Organization for Rare Disorders (NORD) often maintain patient registries and publish prevalence estimates. These are frequently cited in regulatory filings.
• Orphan drug designation documents — if the drug has received or applied for orphan designation from the FDA or EMA, the designation application typically includes a detailed patient population estimate with cited sources. These are publicly accessible and extremely useful.
• Disease registries — condition-specific registries (e.g., EURORDIS registries in Europe, disease-specific US registries) provide real-world patient counts, though they typically undercount due to incomplete enrolment.

When sources disagree — and they often do — present a range rather than a single point estimate. Document every source. Explain why you weighted certain sources more heavily than others.

Step 2: Apply the Diagnosis Rate

Of all patients who have the disease, what percentage are correctly diagnosed?

This is the single most underestimated variable in rare disease forecasting.

For many rare diseases, particularly those with nonspecific symptoms or that mimic more common conditions, the diagnosis rate at baseline is extremely low. It is not unusual for diagnosis rates to be 10–30% — meaning 70–90% of patients with the disease are living undiagnosed.

How to estimate the diagnosis rate:

• Published studies on diagnostic delay and misdiagnosis in the target condition
• Patient advocacy data on average time-to-diagnosis (a longer average delay implies a lower diagnosis rate)
• Comparison to analogous diseases where diagnosis rates have been better studied
• Post-marketing data from any existing treatments (prescription volumes divided by estimated prevalence can back-calculate an implied diagnosis rate)

For your BD&L model, also consider the diagnosis acceleration effect. The launch of a new effective therapy for a rare disease typically increases diagnosis rates significantly, as physician awareness grows and patients who were previously undiagnosed seek evaluation. Model this as an increasing diagnosis rate over time in your forecast.

Step 3: Apply the Treatment Rate

Of correctly diagnosed patients, what percentage receive pharmacological treatment?

For rare diseases with high unmet need and no existing therapies, treatment rates are often very high once a drug is available — because there was previously no treatment option and physicians and patients are highly motivated. In these cases, treatment rates of 70–90% of diagnosed patients are reasonable.

For rare diseases where existing therapies are available (whether approved, off-label, or through compassionate use programs), treatment rates will be segmented:

• Currently treated with existing therapy (potential switchers)
• Previously treated and discontinued (potential re-starters)
• Treatment-naive (new starts)

Each of these segments has different commercial implications for a new drug launch and should be modeled separately if the data permits.

Step 4: Define Eligible Patients for Your Asset

Apply any additional eligibility filters specific to your drug:

• Biomarker selection — if the drug targets a specific genetic mutation or biomarker, what proportion of diagnosed patients carry that marker?
• Line of therapy — is this a first-line, second-line, or later-line treatment? Apply the appropriate line-of-therapy filter.
• Age or weight criteria — some rare disease drugs are approved in specific age ranges (pediatric, adult, or both)
• Contraindications — any patient subgroups that would be excluded based on the anticipated label

After applying all filters, you have your eligible patient pool — the denominator for your market share calculation.

Step 5: Model Market Share and Competitive Dynamics

In rare diseases, competitive dynamics are unlike anything in standard pharmaceutical forecasting.

For first-in-class therapies with no existing standard of care:

Market share is less about competition and more about penetration of the eligible patient pool. The key question is not "what share of the market will this drug capture?" but "what percentage of eligible patients will ultimately be treated with this drug?"

For the first approved therapy in a rare disease with high unmet need, lifetime penetration rates of 60–80%+ of the eligible pool are achievable and defensible. Your market share ramp reflects the speed of diagnosis acceleration, physician adoption, and access decisions — not traditional competitive share-splitting.

For rare diseases with existing therapies:

Apply standard competitive analysis — identify current therapies, their real-world penetration, their clinical limitations, and the differentiation your asset provides. Model a share transition from existing to new therapy based on the strength of comparative clinical data.

Step 6: Apply Orphan Drug Pricing

Rare disease drug pricing follows different logic than standard pharmaceutical pricing.

Cost-effectiveness and unmet need drive price, not competitive benchmarking.

For an orphan drug in the US, the typical pricing framework starts with the cost per quality-adjusted life year (QALY) — regulatory agencies and payers increasingly use this metric even when they do not formally acknowledge it.

Key pricing benchmarks by disease type:

• Rare enzyme deficiencies and metabolic disorders: $100,000–$500,000 annual per-patient cost (Enzyme Replacement Therapies as comparators)
• Rare oncology/hematology: $150,000–$400,000 annual per-patient cost (comparable to other specialty oncology drugs)
• Ultra-rare conditions (prevalence <1,000 patients in the US): $500,000–$3,000,000+ annual per-patient cost (gene therapies, one-time curative treatments)

Gross-to-net adjustments in rare diseases are typically smaller than in large primary care markets — often 10–25% — because payer leverage is lower when patient volumes are small and the drug addresses a clear unmet need. Document your gross-to-net assumption with a rationale.

Step 7: Build Your Scenarios

For rare disease BD&L deals, your three scenarios should specifically stress-test the highest-uncertainty variables:

• Conservative: Lower end of prevalence range, lower diagnosis rate, slower diagnosis acceleration, higher gross-to-net discount
• Base: Best-estimate across all variables
• Optimistic: Upper end of prevalence, faster diagnosis acceleration (e.g., a high-profile registry study or newborn screening program in development), higher treatment penetration

Additionally, run a specific sensitivity on patient pool size — because in rare diseases, a new epidemiology study or registry enrollment data can significantly revise the prevalence estimate after deal close. Show your deal committee the revenue and NPV impact of a 30–50% change in the patient population estimate. This is not a remote possibility in rare disease; it happens regularly.

A Note on Orphan Drug Designation

If the asset you are evaluating has received, or is a candidate for, orphan drug designation from the FDA or EMA, factor the regulatory benefits into your commercial model:

• 7 years of market exclusivity from the date of approval (US), regardless of patent status — this extends the effective commercial window and should be incorporated into your LoE assumptions
• 50% tax credit on qualified clinical trial expenses (US) — a development cost consideration for deal valuation
• Priority review and fee waivers — relevant to development timeline assumptions

Orphan designation is a material commercial asset and should be explicitly addressed in your forecast assumptions.

Common Mistakes in Rare Disease BD&L Forecasting

Applying standard gross-to-net discounts from primary care markets. Rare disease payer dynamics are different. A 50% gross-to-net discount that is reasonable for a primary care drug is almost certainly too aggressive for an orphan disease with limited treatment alternatives.

Ignoring the diagnosis acceleration effect. A new drug launch in a rare disease significantly changes physician awareness and diagnostic behavior. Models that hold diagnosis rates flat from year 1 to year 10 are missing one of the most important revenue drivers in rare disease commercial forecasting.

Using total prevalence as your patient pool. Always work down from total prevalence through diagnosis rate, treatment rate, and eligibility filters. Using total prevalence as your denominator — a surprisingly common shortcut — produces patient pool estimates that are 3–10x the realistic treatable population.

Failing to account for off-label use in the base case. In some rare diseases, physicians prescribe approved drugs off-label before a new therapy receives formal approval in that specific indication. This affects your new-drug launch dynamics and should be modeled explicitly.

Rare disease BD&L forecasting rewards analysts who are methodical about building the patient funnel, transparent about data limitations, and rigorous about documenting every assumption. The models that survive deal committee scrutiny are the ones built on a clear chain of reasoning — not optimistic assumptions.

PharmaceuticalForecasting.com supports epidemiology-based patient-funnel modeling with built-in scenario planning, designed specifically for the BD&L workflow. If your team evaluates rare disease licensing opportunities and the forecasting process is creating bottlenecks, request a demo and we will show you how the platform works.