Why Product-Specific Guidance Matters
Regulatory bioequivalence requirements are not always identical from one product to another. A product-specific guidance (PSG) can tailor recommendations to a particular drug product, including which study conditions to use, which exposure measures to evaluate, and what acceptance criteria apply. The practical task is to turn a regulatory document into a clear set of requirements without treating every sentence as interchangeable or assuming that a recommendation for one product applies to another.
A PSG is guidance, not a statute or a guarantee that a submission will be accepted. Its recommendations should be read in the context of the relevant agency's current regulations and other applicable guidance. An alternative approach may be possible, but it should not be treated as permitted merely because it seems scientifically reasonable. The guidance may specify conditions for an alternative, or the sponsor may need to justify the approach through the applicable regulatory process.
Locate the Right Document Before Extracting Requirements
Begin at the official regulatory agency source. For a U.S. generic-drug development question, the relevant collection is the FDA Product-Specific Guidances for Generic Drug Development. Other jurisdictions may publish their own product-specific documents. A third-party summary, search result snippet, or archived copy can help locate a document, but it should not be the final basis for protocol decisions.
Search using identifying details such as the active ingredient and dosage form, then verify that the document covers the product of interest. A guidance for an immediate-release tablet should not be assumed to apply to a modified-release capsule simply because the active ingredient matches. Check the document's title, scope, publication or revision date, and status. Determine whether it is draft or final, and whether the agency has posted a later version or a notice that changes its status. Read any referenced general guidance when the PSG directs the reader to it.
This check is substantive, not administrative. A seemingly small difference in formulation or route can change the relevant study conditions or endpoints. Likewise, a document may describe recommendations for a specific product while leaving some details to general guidance or scientific justification. Record what the document actually says, what it incorporates by reference, and what remains unspecified.
Match the active ingredient, dosage form, route, and other scope details stated in the official document to the product being developed.
Check the official source for the current document, its date, and whether it is draft, final, revised, or superseded.
Record study conditions, endpoints, acceptance criteria, and any alternatives in separate fields rather than compressing them into a single summary sentence.
Preserve conditional language, cross-references, and unanswered questions. Do not convert silence or an unfulfilled condition into permission.
Build a Requirement Map
A useful extraction table distinguishes four categories. Study conditions describe how the comparison is to be conducted: for example, fasting or fed conditions, dose, population, or the specified study setting. Endpoints identify what is evaluated, such as AUC, Cmax, a specified partial AUC, or a supportive measure. Acceptance criteria state the statistical or other decision limits attached to those endpoints. Alternatives describe other approaches and the conditions under which they may be used.
For endpoints, copy the definition precisely. If a guidance specifies a partial AUC, preserve its time interval and any stated calculation details; “partial AUC” alone is not a complete endpoint description. Distinguish primary endpoints from supportive measures. For acceptance criteria, record the endpoint to which each limit applies, the scale on which the comparison is made, and the confidence interval or other decision rule specified. Do not assume that criteria for one endpoint automatically apply to another.
The extraction should preserve the guidance’s qualifiers. Words such as “if,” “unless,” “may,” and “should” can determine whether a statement is a general recommendation or a conditional branch. A permitted alternative is not necessarily the default approach, and a branch that depends on an eligibility condition cannot be selected unless that condition is met.
| Category | What to record | Question for review |
|---|---|---|
| Study conditions | Meal state, dose, population, and other specified conditions | Does the proposed study match each stated condition? |
| Endpoints | Metric, time interval, and whether primary or supportive | Is each endpoint defined exactly as the guidance states? |
| Acceptance criteria | Limits, scale, interval, and endpoint association | Are all required criteria met for every primary endpoint? |
| Alternatives | Alternative method and its prerequisites or qualifications | Is the alternative explicitly available, and are its conditions satisfied? |
Worked Examples With Illustrative Guidance
The following examples use invented guidance language and hypothetical results. They illustrate how to read and organize requirements; they do not describe actual agency guidance or an actual product. In each example, assume that the stated document is current, applies to the product described, and is the relevant agency document. Statistical calculations are taken as already performed using the specified analysis; the examples focus on extracting and applying the guidance.
Worked Example: Extracting Conditions and Joint Criteria
A hypothetical final guidance for an immediate-release tablet recommends a single-dose study under fasting conditions. It names AUC and Cmax as primary endpoints and states that the 90% confidence interval for each test-to-reference geometric mean ratio must lie within 80.00% to 125.00%. A hypothetical study reports a 90% confidence interval of 92.0% to 104.0% for AUC and 79.0% to 113.0% for Cmax.
Extract the study condition. The study is to be single-dose and fasting. If the proposed study were fed, it would not match the stated condition without a separately justified basis.
Extract the endpoints and criteria. There are two primary endpoints, AUC and Cmax. Each has the same stated interval requirement, but each must be evaluated separately. The guidance’s rule can be written as:
Here, \(L\) and \(U\) denote the lower and upper confidence limits for the named endpoint. For AUC, \(92.0\%\geq80.00\%\) and \(104.0\%\leq125.00\%\), so the AUC interval is within the limits. For Cmax, \(79.0\%<80.00\%\), so its interval is not within the limits. Because both endpoints are primary criteria, the result does not satisfy the full set of criteria.
The conclusion follows from the all-criteria rule: passing one primary endpoint cannot compensate for failing another. It would be incorrect to report that the study meets this hypothetical guidance merely because the AUC interval is within range.
Worked Example: Preserving a Partial-AUC Definition
A hypothetical guidance for a modified-release product names AUC and Cmax as primary endpoints and additionally requires a partial AUC from 0 to 4 hours. It specifies that the 90% confidence interval for the test-to-reference ratio for each named endpoint must be within 80.00% to 125.00%. A hypothetical analysis reports intervals of 94.0% to 108.0% for AUC, 101.0% to 119.0% for Cmax, and 83.0% to 121.0% for AUC from 0 to 4 hours.
The extraction must retain the partial-AUC window: the endpoint is AUC from 0 to 4 hours, not an unspecified partial AUC. Each interval is then checked against both limits. For AUC, \(94.0\%\geq80.00\%\) and \(108.0\%\leq125.00\%\). For Cmax, \(101.0\%\geq80.00\%\) and \(119.0\%\leq125.00\%\). For the partial AUC, \(83.0\%\geq80.00\%\) and \(121.0\%\leq125.00\%\).
All three intervals lie within the stated limits, so all three hypothetical endpoint criteria are met. This conclusion assumes the endpoints were calculated as defined in the guidance and the specified analysis was appropriate. A result labeled only “partial AUC” would not be enough to verify compliance: the time window must match the required 0-to-4-hour interval.
Worked Example: Interpreting a Conditional Alternative
A hypothetical guidance recommends a standard study and states that a second approach may be considered if a specified formulation characteristic is present and the sponsor provides supporting justification. The document does not say that the second approach is an unrestricted substitute. The development team confirms that the characteristic is present and proposes the alternative, but its protocol contains no supporting justification.
The requirement map should list the standard study as the recommendation and record the alternative as conditional. The two stated conditions are: the specified characteristic is present, and supporting justification is provided. The first is met; the second is not. Therefore, the proposed plan has not yet documented that it satisfies the guidance’s stated basis for the alternative.
The correction is not to treat “may be considered” as automatic approval. The team should provide the required justification and follow the applicable regulatory process, or use the recommended approach. This is a document-interpretation conclusion, not a claim that an agency must accept or reject a future submission.
Why the All-Criteria Logic Follows
When a guidance requires several primary criteria, the decision rule is a conjunction: every required condition must hold. Let \(C_1,\ldots,C_k\) denote the individual criteria, with each \(C_j\) equal to true when its specified endpoint meets its specified rule. The combined requirement is:
If even one required criterion is false, the conjunction is false. This follows directly from the meaning of “and”: \(C_{\mathrm{overall}}\) is true only when every \(C_j\) is true. For interval criteria, each \(C_j\) is true only if the full interval meets the applicable bounds. This formalization is a useful check against accidentally substituting an average across endpoints or treating a supportive endpoint as though it replaced a primary one.
A condition that determines whether an alternative is available works similarly. The alternative is supported by the stated guidance only when the required conditions for that branch are met. Keep that eligibility check distinct from the endpoint acceptance decision: meeting an endpoint criterion does not, by itself, establish eligibility for a different study approach.
Diagnostics, Comparisons, and Limitations
Several warning signs suggest that the extraction needs review:
- The document was found only on a third-party site, or its date and status have not been checked against the official source.
- The product name matches, but the dosage form, route, or other stated scope does not.
- An endpoint is recorded without its required time window, analysis scale, or primary/supportive status.
- A criterion is copied without identifying which endpoint it governs, or the proposed plan applies one endpoint’s limit to all endpoints without support.
- A conditional alternative is summarized as generally available, or silence in the guidance is interpreted as permission.
The main alternative to consulting product-specific guidance is to plan from applicable general regulatory requirements and scientific justification. That may be necessary when no applicable PSG exists or when a proposed approach differs from its recommendation. A PSG offers product-focused direction; general requirements provide broader context. Neither should be silently substituted for the other. If the document is ambiguous, appears inconsistent with another applicable source, or does not address a material design choice, identify the uncertainty and resolve it through appropriate regulatory consultation rather than inventing a rule.
A PSG also has limits. It cannot remove the need to demonstrate that the study actually matches its stated conditions, that endpoints were defined and measured appropriately, or that the analysis is valid. Nor does a checklist eliminate scientific judgment: the guidance may be revised, may refer to other documents, or may not address an unusual formulation or development circumstance. Preserve the source and version used, the extracted text or precise section, the interpretation, and any unresolved questions so another reviewer can trace the protocol decisions.
Common Mistake
Wrong reasoning: “The guidance lists a standard study and says an alternative may be considered, so either approach is acceptable without further explanation.”
Correction: Read the conditions attached to the alternative. Record whether the guidance makes the option conditional, whether the stated prerequisites are met, and whether supporting justification or further regulatory engagement is needed. A conditional option is not an unconditional choice.
Check Your Understanding
Answer each question using the requirement-extraction principles in this tutorial. Brief solutions follow.
- A guidance is found on a commercial database, but the official agency site shows a newer revision. Which version should anchor the extraction, and what should be checked next?
- A hypothetical guidance requires AUC and Cmax intervals to fall within 80.00% to 125.00%. The intervals are 90.0% to 110.0% and 80.0% to 126.0%, respectively. Are all primary criteria met?
- A guidance specifies partial AUC from 0 to 3 hours. Why is recording only “partial AUC” inadequate?
- A conditional alternative requires a stated product characteristic and supporting justification. The characteristic is present, but no justification is provided. Is the alternative fully supported by the extracted conditions?
- What is the difference between a study condition and an endpoint in a requirement map?
Solutions. (1) Use the current version posted by the official agency, after confirming that it applies to the product and checking its status and any cross-references. (2) No. The Cmax upper limit is \(126.0\%>125.00\%\), so that criterion fails even though the AUC interval is within range. (3) The time window defines which exposure is measured; a different window would be a different endpoint and may not satisfy the guidance. (4) No. One stated condition is unmet, so the alternative has not been supported on the terms given. (5) A study condition describes circumstances such as fasting or fed administration; an endpoint is the measure evaluated, such as AUC or Cmax.