Hannah Mamuszka and Lena Chaihorsky take on the drug pricing debate from the patient's side, where the question is not what a drug costs but whether it works for the person taking it. Chaihorsky frames it with a metaphor that runs through the episode: we argue endlessly over the price of milk while ignoring that the whole family is lactose intolerant. Mamuszka tells a story that shaped her career. As a young scientist at a small pharma company, she spent years developing a biomarker for a drug at the FDA's suggestion, only for the agency to drop the requirement and for her CEO, at the launch party, to explain that they would treat every patient the label allowed because there were investors to repay. The drug worked in roughly 38% of patients and caused serious side effects in about 40%, and the biomarker could tell those groups apart. They trace how that logic became structural, including the FDA's early-2000s move toward companion diagnostics for all targeted therapies, which collapsed under industry pushback. Chaihorsky then walks through the economics that decide which drugs patients can get when no biomarker stands in the way: PBM formularies rank-ordered by rebate rather than by mechanism of action, step therapy, and a pharmacy-versus-medical budget split that leaves no one owning the cost of being wrong. Their closing asks are simple. Patients should ask how their doctor knows a drug will work for them. Employers heading into benefits season should ask who their PBM is and what its contract rewards.

Key takeaways

• The price of the drug is only half the conversation, whether it works is the other.
• Response rates are far lower than most people assume. Schork's 2015 Nature analysis found the ten highest-grossing US drugs help between 1 in 25 and 1 in 4 of the people who take them.
• Nothing rewards a higher response rate. A drug's price is the same whether it works in 15% of patients or 80%.
• PBMs are paid on rebates rather than net cost, so formularies are rank-ordered by what pays best rather than by who is likely to respond. Using biomarkers to predict response would break that model, which is a reason to expect resistance rather than a reason it doesn't work.
• Pharmacy and medical spend sit in separate budgets, so a wasted prescription and the hospitalization it causes are never added together, and nobody is accountable for the total.
• The question for patients: how do you know this drug will work for me? The question for employers: who is our PBM, and is our formulary built on rebates or on evidence?

Relevant links

• Schork NJ, "Personalized medicine: Time for one-person trials," Nature 2015;520(7549):609–611 — the source of the "1 in 25 to 1 in 4" figure and the imprecision-medicine graphic Lena describes: https://www.nature.com/articles/520609a
• FTC interim staff report on pharmacy benefit managers (July 2024) — the top three PBMs processed nearly 80% of the ~6.6 billion US prescriptions dispensed in 2023: https://www.ftc.gov/news-events/news/press-releases/2024/07/ftc-releases-interim-staff-report-prescription-drug-middlemen
• Substack, "The Price of Drugs Is Only Half the Problem": https://hannahmamuszka.substack.com/p/the-price-of-drugs-is-only-half-the?r=4n5pb&utm_campaign=post-expanded-share&utm_medium=web

Hannah Mamuszka and Lena Chaihorsky interview Kristine Ashcraft, a molecular biologist who has worked in pharmacogenomics (PGx) since 2000, first at Genelex, one of the earliest labs to offer the testing, and later as founder of the clinical-decision-support tool YouScript. Ashcraft traces the field's slow adoption despite strong evidence, explaining how most people carry gene variants affecting how they metabolize common drugs, using a "liver highway" analogy: some people have fewer or more "lanes" (enzyme activity), and drugs, foods, and other medications can shut lanes down. YouScript's clinical studies showed large reductions in hospitalizations, ER visits, and deaths, and the European PREPARE trial's 30% drop in adverse drug reactions. A recurring theme is misaligned incentives — she describes a hospital CFO explaining they make money from the very hospitalizations that PGx would prevent. The conversation covers the Right Drug Dose Now Act, DPYD testing in oncology (where the wrong genotype plus a common chemo drug can be fatal), and practical advice: get tested once, log dangerous drugs as "allergies" so they follow you, and advocate for yourself.

Key takeaways

• Pharmacogenomic variation is nearly universal; most people carry at least one clinically actionable variant affecting how they process common drugs, and yet testing still isn't routine, largely because of misaligned financial incentives and gaps in clinician education.
• The evidence base is strong: PGx-guided prescribing has produced large reductions in hospitalizations, ER visits, and adverse drug reactions in both U.S. studies and the multi-country European PREPARE trial.
• Some drug-gene interactions are life-or-death: DPYD testing before 5-FU/capecitabine chemotherapy, now backed by FDA boxed warnings and NCCN guidelines, can prevent fatal toxicity.
• Practical self-advocacy: pharmacogenomic testing is a one-time, low-cost test; logging a dangerous drug-gene result as an "allergy" helps it travel across fragmented medical records.

Relevant links

• PREPARE study — 12-gene panel, 30% reduction in adverse drug reactions (The Lancet, 2023): https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(22)01841-4/abstract
• YouScript home-health RCT (52% fewer readmissions, 85% mortality reduction), PLOS ONE: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5289536/
• CPIC guideline for CYP2D6/CYP2C19 and SSRIs (citalopram/escitalopram dosing): https://pmc.ncbi.nlm.nih.gov/articles/PMC4512908/
• FDA codeine boxed warning (ultrarapid CYP2D6 metabolism): https://www.ncbi.nlm.nih.gov/books/NBK100662/
• NCCN/FDA DPYD testing before fluoropyrimidine chemotherapy: https://www.fda.gov/drugs/resources-information-approved-drugs/safety-labeling-update-capecitabine-and-fluorouracil-5-fu-risks-associated-dihydropyrimidine
• Right Drug Dose Now Act (H.R. 2471, 119th Congress): https://www.congress.gov/bill/119th-congress/house-bill/2471/text

Hannah Mamuszka and Lena Chaihorsky make the case that diagnostics — not just drugs — are the key to controlling healthcare spending, and that the U.S. systematically underuses them. They lay the groundwork for the series: primary diagnosis (what's actually causing your symptoms), the "diagnosis of exclusion" trap, and two kinds of tests that determine whether a prescribed drug will work for you — pharmacogenomic tests (how your genes affect drug metabolism) and response-predicting diagnostics (whether you'll respond at all). The recurring villain is economics: insurers balk at a $200 test to protect a $10 prescription, ignoring the clinical and downstream costs; physicians skip testing to spare patients surprise bills; and because most drugs work in only a minority of patients, the resulting churn is profitable. They argue this is why response-predicting tests in areas like rheumatoid arthritis exist but aren't covered — using them would upend rebate-driven formularies. The through-line: we'll pay almost anything for a drug, but won't pay to find out if it will work. Their closing prompt for listeners: ask your doctor how they know this drug is right for you.

Key takeaways

• Diagnostics are the "GPS" of care: without an accurate diagnosis and the right tests, treatment becomes expensive trial-and-error — yet the U.S. spends a small fraction on diagnostics relative to drugs.
• Two kinds of tests can tell you whether a drug will work before you take it: pharmacogenomic tests (how your body metabolizes a drug) and response-predicting tests (whether you'll respond at all).
• The economics are backwards: a cheap test that prevents a wrong prescription is often refused because the savings and harms are downstream, while the churn of failed prescriptions is profitable.
• The patient's question — for any new prescription — should be: "How do you know this drug is right for me, and is there a test that would tell us?"

Relevant links

• Check to see if your drug has a pharmacogenomic test associated with prescription: https://www.clinpgx.org/
• Diagnostic errors in the U.S. — overall error rate and harms (NIH/National Academies review): https://www.ncbi.nlm.nih.gov/books/NBK588113/
• How effective are common medications — realistic drug-efficacy meta-analyses (BMC Medicine): https://link.springer.com/article/10.1186/s12916-015-0494-1
• Genetic determinants of warfarin dose (VKORC1, CYP2C9) — PLOS Genetics: https://journals.plos.org/plosgenetics/article?id=10.1371%2Fjournal.pgen.1000433
• Inadequate response to first-line anti-TNF therapy in RA (~30–40%): https://journals.sagepub.com/doi/10.1177/1759720X221114101
• FDA approval standards and the absence of comparative-effectiveness requirements (NEJM): https://www.nejm.org/doi/full/10.1056/NEJMp0906490

In the debut episode of Dirty Little Secrets, Hannah Mamuszka and Lena Chaihorsky introduce themselves and the podcast's central thesis: that the U.S. healthcare system is built on financial incentives that reward treatment over prevention and volume over value, systematically blocking patients from the best data-driven care. Mamuszka, a molecular biologist who moved from cancer drug development into diagnostics, and Chaihorsky, who came to the field through math and finance, each recount early-career moments when validated science was undervalued by the payment system — a high-risk-pregnancy test reimbursed at $11, and a urine test that could spare men unnecessary prostate biopsies but was resisted by urologists whose income depended on the procedure. They argue that patients meet the system at their most vulnerable, unable to shop as empowered consumers, and that misaligned incentives — including the ACA's medical loss ratio, which they contend removes insurers' incentive to cut costs — keep prices climbing. Citing research that rising premiums have eroded decades of middle-class wage gains, they make the case for a more consumer-driven, data-personalized system and invite listeners to start asking harder questions.

Key takeaways

• The podcast's throughline: U.S. healthcare pays generously for treatment (surgery, drugs, hospitalizations) but poorly for the information — diagnostics, screening, risk data — that would let patients be treated as individuals rather than averages.
• Fee-for-service can create perverse incentives: the hosts describe a validated test that could reduce unnecessary prostate biopsies going unused because biopsies are a revenue source — a systems problem, not a critique of individual doctors.
• The ACA's medical loss ratio (insurers must spend 80–85% of premiums on care) is framed as a misincentive: since profit is a slice of a bigger pie, there's little reward for lowering total costs.
• Rising premiums have quietly consumed middle-class wage growth, which the hosts argue should reframe healthcare from an abstract cost into a personal, paycheck-level issue.

Relevant links

• ACA Medical Loss Ratio and rebates — CMS: https://www.cms.gov/marketplace/private-health-insurance/medical-loss-ratio
• "A Decade of Health Care Cost Growth Has Wiped Out Real Income Gains for an Average US Family" (Auerbach & Kellermann, Health Affairs 2011): https://www.healthaffairs.org/doi/10.1377/hlthaff.2011.0585
• Employer premium growth and wage stagnation/earnings inequality (JAMA, 2024): https://pmc.ncbi.nlm.nih.gov/articles/PMC10792464/
• Exo mdx urine test to help men avoid prostate biopsy (Harvard Health): https://mdxhealth.com/exo-mdx-for-physicians/
• "Unintended Consequences of the ACA's Medical Loss Ratio Requirement" (Health Affairs Forefront): https://www.healthaffairs.org/content/forefront/unintended-consequences-aca-s-medical-loss-ratio-requirement

Hannah Mamuszka and Lena Chaihorsky talk with Jane Cheshire Gilbert, a CPA who spent two decades running the health plan for the Teachers' Retirement System of Kentucky (TRS) and became an unlikely pioneer in pharmacogenomics. Serving tens of thousands of retired teachers — many in their 80s and 90s, on an average of 15 medications — Gilbert treated genetic testing not as a clinical curiosity but as a fiduciary tool: if a plan is paying full price for a blood thinner or antidepressant a member's body can't use, everyone loses. She recounts building a program with Coriell Life Sciences and the Know Your Rx Coalition, testing roughly a third of the Medicare-eligible population, and a published study showing meaningful per-member savings and fewer hospitalizations. The conversation is candid about the rebate machine (which she used dollar-for-dollar to hold premiums down), why insurers wouldn't pay for testing that saved them money, and her prescription for the next generation of purchasers: end direct-to-consumer drug advertising, require comparative-effectiveness research, and make real prices visible through reference-based pricing. Her closing line summarizes the whole show: "Mad is not a strategy."

Key takeaways

• A fixed-budget purchaser has nowhere to hide costs, which is exactly why pharmacogenomics made sense here: paying full price for a drug a member's genetics won't let them use is pure waste, clinically and financially.
• The program has real evidence behind it — a published study of the TRS/Coriell program documented significant per-member savings and reduced hospitalizations across thousands of Medicare-eligible retirees.
• The incentives are backwards: the insurer that captured most of the savings wouldn't pay for the testing that produced them, because under a fully insured arrangement those savings were its margin.
• Gilbert's three systemic fixes — ending direct-to-consumer drug ads, requiring drug-vs-drug (not drug-vs-placebo) comparative-effectiveness research, and reference-based pricing tied to Medicare — are areas where purchasers across the political spectrum tend to agree.

Relevant links

• Published study of the TRS Kentucky PGx + medication-management program (Journal of Personalized Medicine, 2022): https://www.mdpi.com/2075-4426/12/3/421
• Coriell Life Sciences' TRS Kentucky case study: https://www.coriell.com/resources/trs/
• Clopidogrel (Plavix) and CYP2C19 genotype — NIH Medical Genetics Summaries: https://www.ncbi.nlm.nih.gov/books/NBK84114/
• CPIC clopidogrel/CYP2C19 guideline (2022 update): https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.2526
• Direct-to-consumer drug advertising — U.S. and New Zealand only (USC Schaeffer): https://schaeffer.usc.edu/research/should-the-government-restrict-direct-to-consumer-prescription-drug-advertising-six-takeaways-from-research-on-the-effects-of-prescription-drug-advertising/