How to Handle HCP Specialty and Subspecialty Ambiguity
A confidence-tier method for building HCP audiences when specialty signals conflict across NPI taxonomy, claims, and partner data.
The short answer
HCP specialty is rarely a single clean field. A provider may carry one taxonomy in their NPI record, a different specialty in a licensed data set, and a third focus visible in their prescribing pattern. When those signals disagree, the audience design has to decide who is in, who is out, and how confident the team is in each decision.
A confidence-tier approach makes that decision explicit instead of leaving it buried in a vendor's default rules.
Where specialty signals come from
| Source | Strength | Weakness |
|---|---|---|
| NPPES primary taxonomy | Public, stable format | Self-reported, often stale |
| Secondary taxonomies | Captures dual focus | Rarely ranked by importance |
| Licensed specialty files | Often refreshed and curated | Methods vary by vendor |
| Prescribing or procedure data | Reflects actual practice | Requires licensing and privacy review |
| Board certification | Verifies training | Does not confirm current focus |
No single source is the truth. The goal is to combine them with rules you can explain.
Build confidence tiers
Assign each provider to a tier based on how many independent signals agree.
- Tier 1: confirmed. Two or more independent sources agree the provider practices in the relevant specialty, and behavior data (where available) supports it.
- Tier 2: probable. One strong source supports inclusion, with no conflicting evidence.
- Tier 3: possible. Signals conflict, or the only evidence is an old or secondary taxonomy.
- Excluded. Evidence points clearly to a different specialty or inactive status.
Then decide how each tier is used. Tier 1 might receive full frequency and all channels. Tier 2 might receive lighter frequency. Tier 3 might be held out of expensive channels or used only in a test cell to see whether it responds.
Handle subspecialties deliberately
Subspecialties create the hardest calls. A general cardiologist and an electrophysiologist share a specialty but may have very different relevance to a given therapy. Decide in advance:
- Does the message apply equally to the parent specialty and subspecialty?
- Is the subspecialty large enough to warrant separate creative or frequency?
- Is there a reliable signal to identify the subspecialty at all?
If the answer to the third question is no, do not pretend to target it. Target the parent specialty and say so.
Keep an exception review
Every rule set produces edge cases. Keep a short, owned list of exceptions reviewed each quarter: providers the field team flags as misclassified, groups where signals changed, and new data that contradicts old tiers. Assign an owner and a review date so the list does not grow forever. The NUCC taxonomy is updated periodically, which is another reason to revisit mappings.
Measure by tier
When reporting reach and outcomes, split results by confidence tier. If Tier 3 providers engage at similar rates to Tier 1, your rules may be too strict. If they barely respond, the rules are doing their job. Either way, you learn something that improves next year's list. This also helps explain match-rate differences across partners, since partners often apply their own specialty logic.
Practical takeaway
Replace the single "specialty" field on your target list with two fields: the specialty assignment and its confidence tier. Agree on how each tier is used before activation, and report results by tier afterward.
Sources
External guidance and platform documentation change. Links were current at publication; check them again before relying on them for a decision.
Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.
New pharma programmatic breakdowns, occasionally
One email when I publish something worth reading. Benchmarks, measurement teardowns, and case studies with the caveats attached. No cadence promises, no reselling your address.
Unsubscribe any time. See the privacy policy.
Working through this decision on a real plan?
I work on health and pharma data, identity, and activation, after five years running HCP and DTC programmatic agency-side. Happy to talk through how this applies to your situation.