De-identification: The two lawful methods and their limits

Auditors ask about de-identification early, because the answer tells them how the rest of the program is likely to look. The requirement is 45 CFR 164.514(a).

What the rule says

Standard: De-identification of protected health information. Health information that does not identify an individual and with respect to which there is no reasonable basis to believe that the information can be used to identify an individual is not individually identifiable health information.

That is the text of 45 CFR 164.514(a). It is a definition, not a duty; what it requires is in the paragraphs that apply it.

There are exactly two lawful methods, and 45 CFR 164.514(b) sets them out. Expert determination under 164.514(b)(1) requires a person with appropriate knowledge of and experience with generally accepted statistical and scientific principles to determine that the risk is "very small" that the information could be used to identify an individual, and, under 164.514(b)(1)(ii), to document "the methods and results of the analysis that justify such determination". Safe Harbor under 164.514(b)(2) requires removing the 18 enumerated identifiers (paragraphs (b)(2)(i)(A) through (R)) of the individual and of relatives, employers and household members, including names, all geographic subdivisions smaller than a State, all elements of dates except the year, and any other unique identifying number, characteristic or code, and having no actual knowledge that the remaining information could be used to identify the individual. Neither method is the default: pick one, and record which you used.

Two Safe Harbor identifiers carry their own exceptions: the initial three digits of a ZIP code may stay where the geographic unit they form contains more than 20,000 people (and are changed to 000 where it does not), and ages over 89 are aggregated into a single category of 90 or older (164.514(b)(2)(i)(B), (C)).

What this looks like in practice

In an assessment, 45 CFR 164.514(a) is not one question. It resolves into several, and each one is really asking for different evidence. These are among the questions asked about it:

• Has the organization assessed whether its de-identification practices adequately address the elevated re-identification risk of genetic information?
• Does the organization have a policy for de-identifying protected information when individually identifiable information is not required for the intended purpose?

Ownership usually sits with the privacy official, though the evidence is often produced by someone else, which is where the trail breaks. A control that works but has no owner tends to stop working the month the person who quietly maintained it changes roles.

Who this applies to

Covered entities. A business associate is bound directly by the provisions that name it (for example 45 CFR 164.502(a)(3), the minimum necessary standard at 164.502(b), and 164.410) and otherwise through its business associate agreement, so read the provision for whom it names. The Privacy Rule has no counterpart to the Security Rule's flexibility-of-approach clause (164.306(b)). Instead, 45 CFR 164.530(i)(1) requires policies and procedures reasonably designed for the size and type of your activities, and any deadline a provision sets applies at every size.

Building it, in order

The sequence below is the order that produces evidence as a by-product rather than as a separate documentation exercise:

1. For genetic information shared for secondary purposes, decide which §164.514(b) method applies: expert determination (§164.514(b)(1)), or safe harbor (§164.514(b)(2)), which requires removing every listed identifier, including any other unique identifying number, characteristic, or code, and having no actual knowledge that what remains could identify the individual. The Privacy Rule names no genetic-specific assessment, so record how the method chosen meets §164.514(a).
2. Account for the risk that genomic sequences and familial patterns may remain re-identifiable after standard identifier removal.
3. Apply additional controls or constraints where standard de-identification is insufficient.
4. Document each determination.
5. Develop a de-identification policy specifying when de-identification is used instead of authorization or full disclosure.
6. Define the de-identification method used, whether removal of specified identifiers or expert statistical determination, and the conditions for each.

Evidence an assessor asks for

For this requirement the evidence types that satisfy it are policy and risk assessment. The distinction matters more than it looks: a policy states what you intend to do, a procedure states how, and a record proves it happened on a date. Auditors ask for all three, and a practice that has written the first two often has nothing for the third.

Date every document and keep the superseded versions. Under 45 CFR 164.530(j)(2) the documentation 164.530(j)(1) requires is kept six years from the date of its creation or the date when it last was in effect, whichever is later. It is the same period the Security Rule sets at 164.316(b)(2)(i), under a different provision, so cite the one that governs the document you are holding.

Where this usually goes wrong

The usual gap is a control that exists in practice and nowhere in writing. Someone does the work, reliably, and it has never been written down, so the organization cannot demonstrate it and cannot notice when it stops. The fix is not more control, it is a dated record and a named owner.

What to do about it

• Produce the evidence, not the intention. For 45 CFR 164.514(a) that means a signed, dated policy with a named owner and a review date and a completed assessment showing the method, the findings and the date. If the only evidence is that someone knows how this works, you have a practice and not a record.
• Put a name on it. Your privacy official should be recorded as the owner. A control with no owner stops working the month the person who quietly maintained it changes roles, and nobody notices.
• Set a review cadence and keep the evidence of each review. An annual date with a dated output is a common floor; what fails an assessment is a control implemented once and never re-checked against a changed environment.

The recurring lesson with de-identification is that documentation is not the paperwork after the control. It is the part an assessor can read, and therefore the part that has to exist.

Related requirements

• Serious Threat to Health or Safety: The narrow safety exception
• Limited Data Set: When a limited data set is the right tool
• Law Enforcement Purposes: What you may disclose, and what you must not

References

• 45 CFR 164.514(a), in 45 CFR 164.514 on eCFR. Quoted text verified against the eCFR text current as of 2026-10-06.
• This post describes what a requirement asks for and what evidence documents it. It is not legal advice, and your counsel owns the conclusion. State law may impose stricter obligations (45 CFR 160.203).