As a regulatory professional I often find myself being the bad guy and asking our clients to scale back the claims in their marketing. This can be frustrating, as many companies are eager to share data developed by their R&D department that supports an expanded claim for the product.
Typically, the marketing team feels it's appropriate to make sure the sales team has this new information so they can share it with doctors. My position is that this kind of data shouldn't be used until those results can be repeated within the CLIA lab, by CLIA licensed personnel, with a statistically significant set of samples and analyzed based on a written protocol.
As it turns out, FDA shares my position. At recent meetings, FDA has been highlighting a the recent settlement of the Quest/Nichols Institute case which alleged misbranding of a IVD product. The $302 million settlement is one of the largest-ever settlements involving a medical device. The settlement included pleading guilty to felony charges.
The case involved a kit Nichols produced and sold to other laboratories. The case apparently originated from claims Nichols made that their test performed nearly identical to a competitor despite knowing that this was not true. I assume they believed that these claims provided them an important marketing advantage and that the risk was worth the benefit. I have trouble imagining they would weigh the risks and benefits the same way now.
The best policy for claims is to make sure your claims are appropriate based on data that has either been sent to FDA or is contained in their CLIA validation. Although I understand the desire to talk about recent developments and use incredible data from R&D, these sorts of claims must be vetted thoroughly before they can be made to the outside world. It can be frustrating to 'just know' your product can do more and not be allowed to say it, but making claims that are not truly validated is simply unwise.
Your competitors are watching. If you cross the line, they will make sure regulators take notice.