Why Report Organization Matters
A PBPK submission report is easier to review when its structure makes the reasoning behind each decision claim visible. A reviewer should be able to identify the question the analysis addresses, understand how the analysis was conducted, locate the evidence supporting its use, and see what the simulations do and do not establish. A collection of technically correct sections can still be difficult to assess if those sections are disconnected or presented in an order that obscures this chain.
The report should therefore be organized around the analysis’s context of use and the evidence required to review the resulting claim. There is no single section order that is mandatory for every submission. The sequence below is a practical framework: adapt it to the analysis, while preserving clear links among the decision question, methods, supporting evidence, predictions, and limitations.
A Reviewer-Oriented Sequence
A useful starting point is to state the decision question before describing the model or presenting results. Then organize the material so that each section answers a question a reviewer would reasonably ask next. The table gives one adaptable sequence. Detailed content within several of these sections is the subject of specialized reporting guidance, but the organizing principle is to make the evidence findable and connected.
| Report section | Reviewer’s question | Evidence to make findable |
|---|---|---|
| Purpose and context of use | What decision is the analysis intended to inform? | Target question, population, use of the prediction, and scope of the claim |
| Methods and analysis plan | How was the question translated into an analysis? | Analysis approach, planned comparisons or scenarios, and decision-relevant criteria |
| Inputs and supporting evidence | What information entered the model, and where did it come from? | Input definitions, evidence sources, and links to the relevant methods and assumptions |
| Model evaluation | What evidence supports applying the model to this use? | Evaluation evidence, its relevance to the intended use, and any unresolved discrepancies |
| Simulations | What was simulated, under what conditions, and why? | Scenario definitions, populations or conditions, and the connection to the decision question |
| Results | What did the specified simulations produce? | Decision-relevant outputs, units, comparisons, and clear links to the methods and scenarios |
| Limitations and interpretation | What could weaken or restrict the conclusion? | Assumptions, evidence gaps, applicability limits, and their implications for the claim |
| Supporting material | Can a reviewer inspect details without losing the main line of reasoning? | Appendices, tables, figures, and referenced records, with precise pointers from the main text |
This sequence separates the description of what was planned and done from the presentation of what was found. It also places evaluation evidence before the simulation results it is intended to support. The report can include a brief orientation to the decision question at the beginning, but it should not rely on an opening statement alone: the body must show the evidence and reasoning that make the claim assessable.
Justifying the Sequence Through Traceability
The organizing logic can be derived from the questions needed to examine a prediction. First, a prediction has meaning only in relation to a specified decision question. Second, the reviewer needs to know how the analysis addresses that question and what information it uses. Third, evidence must be presented to support the use of the model for that purpose. Finally, the reported outputs need to be linked to the simulated conditions and interpreted in light of limitations. Omitting a link makes it harder to determine what a result supports.
This chain is a reporting aid, not a mathematical model or a claim that every submission must use identical headings. Its value is that it exposes missing links. For example, a prediction with no identifiable scenario cannot be checked against the intended decision, while a limitations section that does not point back to affected results may not help the reviewer judge the claim. Cross-references should make those connections explicit rather than requiring the reviewer to infer them.
Describe the intended use and the scope of the claim before presenting supporting results.
Locate the methods, inputs, and planned comparisons that generate the relevant predictions.
Place evaluation evidence before the results that depend on it, and identify its relevance to the intended use.
Connect each result to its simulation conditions, then state limitations that could restrict its interpretation.
Worked Examples of Report Planning
Worked Example: Mapping a Drug-Interaction Claim
Hypothetical case. A submission uses PBPK simulations to inform whether a fictional investigational drug, AX-17, may affect exposure to a fictional substrate, BZ-4, when the drugs are coadministered. The report team has five required planning items: the decision question, the interaction scenario, the model inputs, evaluation evidence, and the predicted outcome. The task is to place each item and check that the claim is traceable.
The decision question and scope belong in the purpose and context-of-use section. The interaction scenario belongs in the simulations section, with a pointer to the methods describing how it was specified. The inputs belong in the inputs section, with their evidence sources identified. Evaluation evidence belongs in the model-evaluation section, before the result is interpreted. The predicted outcome belongs in the results section and should refer back to the scenario that generated it.
For a simple planning audit, define the mapped-item fraction as the number of required items assigned to a report location divided by the total number of required items. This is an internal completeness check, not a regulatory acceptance criterion and not a measure of evidence quality. Here, all five items have locations:
The arithmetic shows that all five planning items are located. It does not show that the inputs are adequate, that the model is suitable, or that the predicted outcome is supported. Those questions require reviewing the evidence itself. The planning check is useful because it distinguishes a missing report link from a substantive scientific weakness.
Worked Example: Keeping Discordant Evidence Visible
Hypothetical case. A team plans to report three clinical evaluation comparisons for a fictional compound. The observed-to-predicted exposure ratios are 0.94, 1.06, and 1.68. The first two comparisons are close to one another, while the third is visibly higher. The purpose here is to organize the report, not to decide whether the model is qualified.
The methods section should define how each comparison was made and identify the three comparisons before results are shown. The evaluation section should present all three results together, with units or a clear definition of the ratio, and identify which evidence corresponds to each comparison. The discussion should point to the third comparison and explain what follow-up or limitation is relevant to the intended use. Placing only the first two comparisons in the main text and leaving the third unexplained in an appendix would make the evidence harder to assess.
As a report-planning check, count the comparisons identified in the methods and the comparisons presented in the evaluation. Here, the methods identify 3 comparisons and the evaluation presents 3, so the count difference is \(3-3=0\). That agreement checks coverage, not scientific acceptability. The third ratio, 1.68, remains part of the evidence and should not be hidden by a complete count or by a summary that emphasizes only the closer values.
A reviewer can then follow the cross-reference from the evaluation result to the corresponding method, and from the result to the limitation or interpretation that addresses its relevance. This makes the discrepancy visible without implying that its presence alone determines the decision.
Worked Example: Reporting a Scenario-Specific Prediction
Hypothetical case. A PBPK analysis reports a BZ-4 exposure ratio of 1.60 for one simulated coadministration scenario. The ratio is defined as exposure in the simulated coadministration condition divided by exposure in the corresponding reference condition. The report also considers whether that scenario represents the proposed use.
The result can be described as a predicted exposure 1.60 times the reference exposure in that specified scenario. Expressed as a percentage increase, the calculation is:
Thus, the simulated exposure is 60% higher than the reference exposure under the stated scenario. The report should place the scenario definition in the simulations section, state the ratio definition and comparison in the methods, and present the 1.60 result with the relevant units or exposure measure. The interpretation should then identify whether assumptions or applicability limits affect the relevance of this scenario to the decision question.
The arithmetic does not establish that the scenario represents every patient or every possible use. It only translates the reported ratio into a percentage increase for this defined comparison. A reviewer needs the scenario, the reference condition, and any relevant limitations to interpret that number.
Diagnostics, Alternatives, and Limitations
A useful report review checks not only whether headings exist, but whether the evidence can be followed across them. Common warning signs include a result that cannot be tied to a named simulation; methods that describe a different comparison from the one shown in results; evaluation evidence that is mentioned without explaining its relevance; and limitations that are generic rather than connected to particular claims. Missing or ambiguous cross-references can produce the same problem even when the information exists somewhere in the submission.
One alternative is to organize the report by model component or by the order in which the analysis was performed. That arrangement may help readers locate technical detail, but it can obscure the decision logic if it separates a claim from its supporting evidence. A decision-oriented sequence is generally easier to review for a specific regulatory use. These approaches can be combined: retain technical sections or appendices where useful, while adding a claim-to-evidence map and precise cross-references in the main report.
A cross-reference table is particularly useful when one item supports several claims or when a claim depends on evidence distributed across multiple sections. Keep it concise: identify the claim, the relevant section or figure, and the evidence needed to assess it. Do not treat a populated table as proof that the evidence is adequate. Organization supports review; it does not replace scientific evaluation.
The report architecture also has limits. A well-ordered report cannot resolve weak evidence, unsupported assumptions, or a mismatch between the analysis and its intended use. Nor does a complete section sequence guarantee that a reviewer will interpret the evidence as the authors intend. The report must state the scope of each conclusion, preserve relevant discordant findings, and make uncertainty or limitations visible where they affect the claim.
Common Mistakes
- Starting with technical detail before the question. A reviewer may not know why a model description or result matters. State the decision question and intended use first.
- Treating a contents page as traceability. Headings alone do not connect a result to its scenario, method, or supporting evidence. Add specific cross-references.
- Putting all supporting evidence in appendices without signposting. Appendices can hold detail, but the main text should identify the evidence and point to its location.
- Equating completeness with adequacy. A report may account for every planned item and still contain evidence gaps. Completeness is an organizational property, not a scientific conclusion.
- Reporting limitations apart from the claims they affect. Link limitations to the relevant result or scope so the reviewer can judge their implications.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions below.
- A result is presented without stating which simulated condition produced it. What reporting link is missing?
- A report maps 4 of 5 required planning items to sections. What is the mapped-item fraction, and what does it tell the reviewer?
- Why should a discordant evaluation result remain visible in the evaluation section rather than being omitted from the main report?
- A predicted exposure ratio is 1.25 relative to a stated reference condition. What percentage increase does this represent, and what should accompany the number in the report?
- Does a complete report sequence demonstrate that the PBPK model is adequate for the intended use? Explain.
Solutions. (1) The result needs a traceable link to its simulation conditions, including the relevant comparison and a pointer to the methods. (2) The fraction is \(4/5=0.80\), or 80%. It indicates that one required item has not yet been assigned a report location; it does not measure evidence quality or model adequacy. (3) The result is part of the evidence needed to assess the analysis. Omitting it could give an incomplete account and prevent the reviewer from judging its relevance; the report should explain its context and implications. (4) \((1.25-1)\times100\%=25\%\). State the exposure measure, the simulated condition, the reference condition, and any limitations relevant to interpretation. (5) No. A complete sequence improves traceability, but adequacy depends on the scientific evidence and its relevance to the intended use.