Modeling patient lifetime value across episodes of care

A woman delivers her first baby at your hospital in 2026. Over the next decade she returns for a second delivery, brings both children in for pediatric and urgent visits, sends her sister to your OB-GYN group, and moves the family's imaging and lab work into your network. A man arrives in urgent care the same week with a sprained ankle, pays, and never comes back.

Your record system counts two new patients. Your media plan should not treat them as equal. Putting a defensible number on that gap is the whole job of patient lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition →, and the number it produces is the ceiling on the per-service-line acquisition costacquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → the sibling lesson shows you how to calculate.

Why PLV, not just a single visit

Patient Lifetime Value (PLV) is the marketing-relevant contribution a patient, and often their household, is expected to generate across the whole relationship, net of what it costs to win and keep them.

Note the "marketing-relevant" framing. This is not hospital finance. It answers one question: how much is it rational to spend to acquire and retain this patient?

The classic mistake is valuing a patient at the margin of visit one. Four things push true value well past that first episode, and the fourth is the one that gets left out of most spreadsheets.

The drivers most hospitals leave out

Recurring conditions. Diabetes, heart failure and COPD (chronic obstructive pulmonary disease) generate predictable encounters for years. One diabetes patient is a decade of endocrinology visits, labs and pharmacy touchpoints.

Downstream referrals. Trust is high stakes in care, so word of mouth and physician referral patterns compound harder than in retail.

Household value. The parent who trusts your maternity unit routes pediatric, urgent and often spousal care to the same network.

Payer trajectory. In the US, commercial reimbursement for the same episode sits far above Medicare; RAND's hospital price work has repeatedly put average commercial prices above 200 percent of Medicare rates. A 62-year-old commercially insured orthopedic patient crosses into Medicare at 65, so the margin per episode drops partway through your horizon. Model a ten-year relationship for a joint-replacement line whose patients are mostly in their early sixties at commercial rates and you will overstate value by a wide margin.

Building the model

Here is a structure you can defend to a CFO without wandering into capital ratios.

PLV = (Direct episode value)
    + (Recurring condition value)
    + (Referral value)
    + (Household value)
    - (Cost to acquire + cost to retain)

Direct episode value = contribution margin per episode x expected episodes, with the persistent continuity rate from the retention lesson applied year over year.

Recurring condition value = annual episodes x margin x expected years of relationship.

Referral value = (referrals x conversion x average new-patient PLV) x attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → weight. You should not claim full credit for a referral the patient would have made anyway; a weight of 0.3 to 0.5 is common practice.

Household value = additional household members captured x their PLV x capture probability.

Discounting. Apply a modest rate (a marketing simplification, not a WACCWACCThe blended rate a company pays to finance itself through debt and equity. It sets the minimum return an investment must clear to create value.View full definition → exercise). Eight percent is a defensible round figure; flag it as an assumption.

When the sign flips: capitation

The formula above assumes you get paid per episode. Where the system also carries the insurance risk, more episodes are cost, not revenue. Kaiser Permanente covers over 12 million members through its own health plan and delivery arms, and Geisinger, which ran its own health plan for decades before becoming the first system to join Kaiser's Risant Health venture in 2024, sits in the same position. In that world PLV becomes membership years multiplied by per-member-per-month margin, and the marketing objective changes: enrol and keep members, and value the ones whose conditions are managed cheaply and well. Bupa, as an insurer that also owns clinics and dental centres, models the same way: value is policy renewal years, and a claims-heavy member can carry negative modelled value while still being exactly the person you should serve properly.

Get this wrong and you build campaigns that pay per booked episode inside a business that loses money on every avoidable one.

Worked example: maternity vs urgent care

All figures below are illustrative teaching estimates, not benchmarks. Contribution margins vary enormously by payer mix and region.

Patient A: maternity, first delivery at age 30

• Delivery episode margin: 3,000 dollars
• Second delivery in year 3: 3,000 dollars
• Two children, pediatric plus urgent care: 400 dollars/year x 2 kids x 8 years = 6,400 dollars
• Her own routine care, 10 years: 300 dollars/year x 10 = 3,000 dollars

Direct plus recurring subtotal (undiscounted): 15,400 dollars

Referrals: 2 referrals, 50 percent convert at ~5,000 dollars PLV each, attribution weight 0.4 = 2,000 dollars

Spouse capture: 60 percent probability, spouse PLV ~4,000 dollars = 2,400 dollars

Gross PLV before discounting: 19,800 dollars. A simplified midpoint discount trims roughly 25 to 30 percent off value spread over a decade, landing near 14,500 dollars.

Patient B: urgent care sprained ankle, age 30

• Single episode margin: 200 dollars
• 20 percent chance of another urgent visit in ten years x 200 dollars = 40 dollars
• Referrals: minimal from a low-trust event, call it 50 dollars weighted
• Household: negligible

Gross PLV: roughly 290 dollars, effectively unchanged after discounting because it is front loaded.

The punchline for your ad budget

Patient A is worth roughly 50 times Patient B in modelled terms. At a 3-to-1 LTV-to-CAC target, that is about 4,800 dollars of justifiable acquisition spend for the maternity patient and about 95 dollars for the ankle. Same hospital, same team, two unrelated playbooks: prenatal classes and unit tours on one side, local search and signage on the other.

Benchmarks and where to anchor

Public PLV benchmarks for hospitals barely exist, because contribution margin depends on payer mix. Treat any single "average patient value" figure with suspicion, and check what share of a service line's volume is Medicaid before you extrapolate a commercial margin across ten years.

What you can anchor on:

• LTV:CAC of 3:1 as a healthy marketing target, common across service industries.
• Service lines with recurring or high-margin episodes first: maternity, cardiology, orthopedics, oncology.
• Payer-mix drift over the horizon, modelled explicitly rather than assumed flat.

One edge case worth naming: oncology looks like the richest line in any PLV model and is the one where a flat ten-year horizon is least honest, because expected years must reflect survival. Model it any other way and you overstate value, and the model reads badly if it ever leaves the marketing department.

For methodology grounding, the Agency for Healthcare Research and Quality (AHRQ) publishes utilization and patient experience data useful for estimating repeat-visit rates. In systems where care runs through public funding, the acquisition side compresses: patients are often allocated rather than competed for, so PLV modelling shifts almost entirely to household capture and continuity inside the network.

Putting the model to work

Segment acquisition spend by PLV, not by volume. A campaign delivering 100 urgent care visits can look excellent on a dashboard and be worth less than one delivering 10 maternity patients.

Extend the funnel past discharge. Referral and household value carry a large share of maternity PLV, so postpartum follow-up, pediatric onboarding and family portal enrolment are the marketing steps that actually realise the modelled number. If they do not exist, delete that value from the model.

Do not double count. The referred sister's PLV appears in her own record and inside the referrer's referral term. Sum both across a whole panel and you can inflate portfolio value by a third or more. Pick one place to book referral value and be consistent, which is why the attribution weight exists.

Watch record linkage. Household value assumes you can see the household. Duplicate medical records inside a single system commonly run in the mid single digits as a percentage of the master index, and worse after a merger; address-based family linking breaks on divorce, moving and adult children. Epic Systems, which sells the record platform most large US systems run on, gives you the longitudinal patient view, not a verified household view. Overstated capture probability is usually a data problem wearing a marketing costume.

Stress test. Referral weight and household capture are the sensitive inputs. Run conservative and optimistic cases; if the budget decision survives the conservative one, it is safe.

PLV is a marketing planning tool, not a clinical prioritization tool. It informs outreach spend, nothing more.

Key Takeaways

• PLV extends far past the first visit. Recurring conditions, referrals and household value often dwarf the initial episode in maternity, chronic care and oncology.
• A maternity patient can model at 50x a one-time urgent care visit, which justifies completely different acquisition budgets under a 3:1 LTV:CAC target.
• Payer mix moves over the horizon. Commercial-to-Medicare crossover and capitated contracts (Kaiser Permanente, Geisinger, Bupa) can shrink or invert episode value, and under capitation the model becomes membership years x PMPM margin.
• Attribution weight, household capture and record linkage are where the model breaks. Model both cases, book referral value once, and never claim full credit for organic referrals.
• PLV guides marketing spend, never clinical priority. Keep the two firmly separate.