IRS

The research and development tax credit has existed for decades. It is legislated, it is available, and it is audited more aggressively than most business owners realize. The IRS has designated research credit claims as a Tier I audit issue, noting that claims are often filed using high-level estimates, invalid assumptions, lack of means between qualified research expenses and the business component without contemporaneous documentation to support the claim.

The documentation standard is specific. The IRS requires claimants to identify all business components, identify all research activities performed for each component, and provide total qualified wage, supply, and contract research expenses . A general statement that a practice is innovative does not satisfy that requirement. A percentage estimate of time spent on research does not satisfy it either. What satisfies it is a record, built as the work happens, that connects specific activities to specific costs.

That standard is where many claims fail. The Tax Court reinforced this in a 2026 decision, with Judge Greaves framing the issue as which came first, the research or the research credit study, and finding one claimed activity was a clear example of the chicken (research credit study) coming before the egg (research) because contemporaneous records contradicted the claim, as reported by Current Federal Tax Developments. The lesson is straightforward. The documentation has to exist before the study does.

The four-part test itself is not a secret. It requires a permitted purpose, a technological basis in the physical or biological sciences, technical uncertainty at the outset, and a process of experimentation. What is difficult is applying that test to the actual work of a business, especially a business whose work does not look like traditional laboratory research. The gap between the statutory language and the daily record is where claims are lost.

Benjamin T. Dyches has watched that gap from both sides. He practiced dentistry for about eleven years before going to law school, and he now runs TaxRx Group, a firm that prepares R&D tax credit studies for healthcare practices. His view is that the credit is real and the complexity is real, but the difference between a defensible claim and a vulnerable one is the quality of the analysis behind it.

"Most of the industry approaches this with a percentage estimate," Dyches says. "They ask a doctor what proportion of their time was spent on research, and they write that number down. That is not a study. That is a guess with a cover page." His firm's stated approach is to break clinical procedures into their constituent components, identify where technical uncertainty and experimentation occurred, and build documentation that can answer questions if the IRS examines the claim.

The clinical work that can qualify is not routine patient care. A practice refining a surgical protocol, testing a new material, or building a custom software integration may be doing work that meets the four-part test. an industry mandated protocol with no uncertainty is not. The distinction is technical, and it is exactly the kind of distinction a generalist adviser is not equipped to make.

That is the gap Dyches is arguing against. The IRS has warned about firms marketing research credit studies as a product, often on contingency, and the audit consequences fall on the taxpayer who signed the claim. The businesses most exposed are not the ones deliberately cheating. They are the ones who trusted an estimate because it sounded professional and never asked what evidence sat behind it.

The practical implication is not that every practice should claim the credit. It is that any practice considering a claim should ask a specific question. What documentation exists, created at the time the work happened, that connects this activity to this cost? If the answer is a study assembled last month from an interview and a percentage, the claim is likely not defensible. Qualification is much more likely when contemporaneous care records contribute directly to the development of a clinical practice guideline. The credit rewards evidence. The question is whether the evidence exists.