It is nine in the evening and a hospital oncology pharmacist is looking at a genotype report she has seen maybe twice before in her career.
The patient is starting a combination regimen. One of the drugs is metabolized through a pathway where this specific CYP2D6 variant matters, and the standard dosing table does not quite cover the case in front of her. Lexicomp gives her the general mechanism. CPIC's guideline, which is genuinely excellent, gives her the population-level recommendation. Neither tells her what actually happens in a real patient who looks like this one.
She checks whether anyone else in her department has handled this specific combination before. Nobody has. She posts to an ACCP Practice and Research Network listserv, the closest thing her profession has to a peer channel for exactly this kind of question, and settles in to wait, because a listserv is asynchronous by design and this patient's dosing decision is not.
She could call the manufacturer's medical information line, and she knows, as every pharmacist in this position knows, that whatever she hears back will be filtered through a promotional relationship she did not ask for and cannot fully trust.
What she wants, and does not have, is a specific human being: a verified oncology pharmacist with pharmacogenomics training who has actually dosed this drug-gene combination before, reachable tonight.
She has excellent static references and no live, verified peer to ask, so the decision in front of her tonight rests on her own best judgment, built from resources that were never designed to answer a case-specific question in real time.
The gap between what pharmacists know and what they were taught
Start with the number that frames everything else in this article, because it is unusually precise about exactly where the problem sits.
A 2026 international survey found 72.5 percent of pharmacists reported familiarity with pharmacogenomics, but only 8.8 percent had received formal education in it. That is a 63.7 point gap between what pharmacists say they understand and what they were actually, formally trained to do.
The same survey asked pharmacists what stood between them and confident pharmacogenomics-informed practice. The leading barrier, cited by 47.9 percent, was insufficient knowledge. Cost followed at 26.2 percent, and limited trained personnel at 14.3 percent.
Here is the detail that reframes the whole problem: 92.5 percent of the same pharmacists said they wanted more training. This is not a population that has checked out. It is a population that has correctly identified its own gap and is asking, in large numbers, for someone to close it.
One caveat has to sit alongside these numbers honestly: this survey (Balaban et al., Scientific Reports, 2026) is international, not US-specific. This article has not located a directly comparable US-only dataset with the same precision. The direction of the finding, a wide gap between self-reported familiarity and formal training, is consistent with what US pharmacy workforce and PGx-implementation commentary generally describes, but the exact percentages should not be quoted as a US statistic.
The static reference is not the missing piece
It would be easy to read the training gap above and conclude the fix is more continuing education modules on pharmacogenomics. That is not quite what the evidence supports, and getting this distinction right matters for what gets built next.
The contrarian read: pharmacists already report high willingness to train, at 92.5 percent. The barrier that shows up most in their own accounting is not motivation, and arguably not even raw factual knowledge in the abstract. It is the absence of a live, verified peer who has actually managed the specific rare case in front of them, right now.
CPIC guidelines are a genuinely strong resource: peer-reviewed, systematically maintained, and specific about drug-gene pairs. UpToDate, Lexicomp, and Micromedex are comprehensive static references that most pharmacists reach for reflexively. None of these tools, however excellent, can answer the question a pharmacist actually has at nine at night: has anyone dosed this exact combination in a real patient, and what happened when they did.
A static guideline tells you what the evidence supports in general. It cannot tell you what a rare, ambiguous, real-world case looked like when someone else handled it. That distinction, between population-level guidance and case-specific peer judgment, is the same one that recurs across clinical specialties whenever a rare situation exceeds what any single practitioner has personally seen.
The current workflow, laid out in full
Walk through what a pharmacist actually does today, in order, when a rare-drug or pharmacogenomics question arises that exceeds their own department's in-house expertise.
First, ask whether any specialist exists inside the pharmacy department itself. In many hospitals, especially outside major academic centers, the answer is no.
Second, post to an ACCP Practice and Research Network listserv. PRNs are genuinely the closest thing this profession has to organized peer infrastructure, organized by topic area, with members who opted in because they share a clinical interest. They are also, structurally, asynchronous, member-gated, and not built for a same-day question with any accountability for response time.
Third, call the drug manufacturer's medical information line. This channel exists, is often responsive, and carries an unavoidable structural conflict: the person answering works for the company that makes the drug, and their responses are constrained by what they are legally and promotionally permitted to say.
Fourth, search static references (UpToDate, Lexicomp, Micromedex) for whatever general guidance exists, and check CPIC's guidelines (cpicpgx.org) directly for the specific drug-gene pair in question, if one has been published.
Notice what is absent from every step of this sequence: at no point does the pharmacist have a way to say "verified oncology pharmacist with a PGx credential, currently reachable" and get routed to exactly that person.
Why nobody owns it
Look at the organizations that could plausibly build this, and the structural reason each one does not.
ACCP and ASHP are membership, continuing-education, and advocacy organizations. Their revenue model and organizational mission do not include building or maintaining live consult-routing software with response-time accountability. PRNs are a genuinely valuable community structure built inside an organization that was never designed to operate as a real-time dispatch system.
Manufacturer medical information lines carry a structural promotional conflict that cannot be engineered away. A pharmacist asking a manufacturer's own line for guidance on that manufacturer's own drug is, by definition, not getting an independent second opinion.
EHR-embedded clinical decision support, from vendors like Lexicomp and First Databank, is static reference content delivered inside a workflow. It is valuable and widely used, and it is categorically not a peer network; it cannot tell a pharmacist whether a human being has actually managed this exact case.
CPIC is, by design, a guideline body. It produces excellent, rigorously maintained population-level dosing recommendations for specific drug-gene pairs. It was never structured to be, and does not claim to be, a live consult mechanism for the ambiguous cases that fall outside or between its published guidelines.
The pattern across all four: every existing institution touching this space does its actual job well. None of those jobs is "verified, specialty-tagged, response-time-accountable peer consult," so nobody has built it, not because the need is unclear but because it falls in the gap between four organizations' clearly defined mandates.
Why now
Two trends are converging faster than pharmacist training capacity can currently absorb them.
Pharmacogenomics-guided prescribing is scaling: FDA labels increasingly carry PGx information, and CPIC guideline adoption continues to spread into routine practice. That 63.7 point gap between self-reported familiarity (72.5 percent) and formal education (8.8 percent) is the direct, measurable signature of training capacity falling behind clinical demand.
At the same time, specialty drug spending continues to rise as a share of total drug spending, which raises the stakes attached to every dosing decision in this category. A wrong or delayed call on a narrow-therapeutic-index or pharmacogenomically sensitive drug carries direct patient-safety consequences, and specialty and rare drugs are disproportionately expensive, so an avoidable adverse event or wasted dose is disproportionately costly as well.
What would actually work
A verified, specialty-and-credential-tagged consult layer, not another listserv. The routing key has to be specific: not "pharmacist," but "oncology pharmacist, PGx-credentialed, currently reachable," matched to the actual clinical question rather than a general topic area.
A response-time expectation attached to the request. A system with no expectation of a timely answer will not get used for a same-day dosing decision, no matter how good its member base is. The six-hour or same-shift window that matters in acute clinical decisions elsewhere in medicine applies here too.
De-identified, case-based consultation. No patient identifiers cross institutional lines. The conversation is about the drug-gene pair, the clinical pattern, and the management approach, structured the same way de-identified peer consultation works in other high-stakes clinical domains.
Independent of manufacturer influence. For this to be trusted the way the current manufacturer medical-information line is not, it has to be structurally separate from any pharmaceutical company's promotional interest, funded through pharmacist dues, health-system sponsorship, or a comparable independent model rather than industry sponsorship of the answer itself.
Built to complement CPIC and static references, not compete with them. The value proposition is not "we know better than the guideline." It is "here is what to do when the guideline does not cover your specific case, or when you need to know whether anyone has actually seen this combination before."
Same-profession, not physician-gated. Pharmacists carry their own dosing liability and their own clinical judgment. A consult layer that routes pharmacist questions through a physician's authority, rather than directly to a peer pharmacist, filters out exactly the kind of practitioner-to-practitioner exchange this problem requires.
Growing its own reference asset over time. Every de-identified case that moves through a structured consult layer, if captured with informed contributor consent and no PHI, becomes a searchable case library organized by drug-gene pair and clinical scenario, a resource that gets more valuable as it accumulates rather than staying static like a guideline document.
What you can do now
If you are a clinical pharmacist
Name your own network today, before you need it. Identify two or three colleagues, inside or outside your institution, whose specialty or PGx training would make them a genuine resource for a rare-drug question, and confirm directly with them that they are willing to be asked.
Use ACCP PRNs deliberately, not just reactively. Join the PRN closest to your actual practice area now, rather than discovering it exists the night you need an answer. Community built in advance responds faster than a cold post from an unfamiliar name.
Document your own rare cases as you go. The rare drug-gene combination you dose correctly tonight, and the reasoning behind it, is exactly the kind of case-specific knowledge that has no institutional home right now. Keeping your own record of it is the only way that knowledge survives past this one patient.
If you lead a pharmacy department
Map your department's actual PGx and rare-drug coverage honestly. Know which of your pharmacists, if any, hold specific pharmacogenomics training, and make that internal expertise visible and reachable to the rest of the department rather than assumed.
Budget for external consult access, not just internal training. The 92.5 percent of pharmacists who want more training in the 2026 survey are asking for both education and access; a department that only invests in the former still leaves its staff without anyone to call on a genuinely novel case.
Treat a near-miss on a rare-drug dosing decision as a system signal, not an individual failure. If a pharmacist had to make a difficult call without adequate peer input, that is information about a structural gap in your department's access to expertise, not evidence that the individual pharmacist did anything wrong.
If you build systems
Design for credential-tagged routing from day one. The value of this product lives entirely in matching a specific question to a specific, verified expertise profile; a generic peer-messaging tool without that structure recreates the listserv problem in a new interface.
Partner with ACCP rather than compete with it. ACCP holds the professional community and credibility this needs; it lacks the software product infrastructure and funding model to build live consult routing itself. The realistic path is partnership, not displacement.
Frequently asked questions
How do pharmacists get help with rare drug dosing questions? Today, mainly by asking colleagues inside their own department if any exist, posting to an ACCP Practice and Research Network listserv and waiting for a reply, calling the manufacturer's medical information line, or consulting static references like Lexicomp, Micromedex, and CPIC guidelines. None of these is a verified, response-time-accountable live peer consult mechanism.
What is a Practice and Research Network (PRN) in pharmacy? PRNs are topic-based member groups run by the American College of Clinical Pharmacy (ACCP), organized around shared clinical interests such as oncology or pharmacogenomics. They function largely as asynchronous listservs and meeting communities, not as a same-day, verified consult routing system.
How do pharmacists implement pharmacogenomics guidance? Largely through CPIC's published, peer-reviewed drug-gene guidelines, alongside EHR-embedded decision support tools like Lexicomp. A 2026 international survey found 72.5 percent of pharmacists reported familiarity with pharmacogenomics, but only 8.8 percent had received formal education in it, indicating a substantial gap between awareness and structured implementation capability (Balaban et al., Scientific Reports, 2026).
Where can a pharmacist ask a peer about an unusual drug interaction? The closest existing options are ACCP PRN listservs, informal colleague networks, or, indirectly, a manufacturer's medical information line for that specific drug. No verified, specialty-tagged, response-time-accountable peer network currently exists to route an unusual drug interaction question to a specific qualified pharmacist on demand.
Why don't manufacturer medical information lines solve this problem? Because the person answering works for the company that makes the drug, creating a structural promotional conflict that cannot be resolved through good intentions alone. Pharmacists consistently treat manufacturer lines as one input among several rather than an independent source of peer judgment.
Is the pharmacogenomics training gap a US-specific problem? The clearest published figures (72.5 percent familiarity versus 8.8 percent formal education, from a 2026 international survey) are not US-specific, and this article did not locate an equally precise US-only dataset. The general direction, awareness outpacing formal training, is consistent with broader commentary on US pharmacogenomics implementation, but the exact percentages should be read as international rather than domestic.
The bottom line
The oncology pharmacist looking at an unfamiliar genotype report tonight has, by any reasonable measure, excellent tools. CPIC's guidelines are rigorous. Lexicomp and Micromedex are comprehensive. Her own training, whatever its formal gaps, is real and current enough to get her most of the way there.
What she does not have is the one thing that would resolve her actual uncertainty: a verified person who has dosed this specific combination before, reachable tonight, independent of the manufacturer's promotional interest and faster than an asynchronous listserv post.
The 72.5 percent to 8.8 percent gap between reported familiarity and formal pharmacogenomics education describes a profession that knows roughly where its blind spots are and has said, at 92.5 percent, that it wants help closing them. That is not an engagement problem. It is an infrastructure problem, and infrastructure is exactly what nobody with the incentive to fix training has built.
ACCP's PRNs, CPIC's guidelines, and manufacturer information lines each do their actual job well. None of their jobs is live, verified, credential-matched consultation with an expectation of a timely response, which means the gap between what a rare-drug case demands and what current channels can deliver sits precisely in the space between four institutions each doing something else correctly.
Until that gap closes, the decision at nine in the evening still rests on one pharmacist's best judgment, built from resources that were never designed to answer a case-specific question in real time, for a patient who cannot wait for the listserv to catch up.
Part of a series on the missing professional infrastructure of healthcare. Previously: Decision Orphaning
Evidence note: the pharmacogenomics familiarity/education gap (72.5 percent familiar, 8.8 percent formally trained) and the associated barrier and training-interest figures (47.9 percent insufficient knowledge, 26.2 percent cost, 14.3 percent limited trained personnel, 92.5 percent wanting more training) are from Balaban et al., Scientific Reports, 2026 (PMID 42135435), an international survey that is not US-specific; this article did not locate a directly comparable US-only dataset with the same precision and flags the figures accordingly. ACCP's Practice and Research Network structure is confirmed via ACCP's own published materials (accp.com); this article did not independently verify live PRN membership counts or actual response times, and describes PRNs structurally rather than quantitatively. The cited context on healthcare workforce strain (RN turnover and shortage figures from NSI Nursing Solutions, 2026) is used only as broader supporting context for declining in-house specialist depth and is not specific to pharmacy. Nothing in this article is clinical dosing guidance for any specific patient or drug-gene combination.
