Why Define a Context of Use?
A model is not simply “useful” or “validated” in the abstract. Its usefulness depends on what it is intended to do, for which population, in which setting, and for whom. A model may be informative for one purpose but unsuitable for another. Defining its context of use makes the intended role explicit before people rely on its results.
The earlier material on moving from a development question to a modelable question helps make a question answerable using a model. A context-of-use statement adds a different kind of precision: it describes the circumstances in which the model is intended to be used and what it is expected to provide there. It does not, by itself, show that the model is adequate for that use or specify exactly how a result will determine an action.
These elements work together. “Estimate drug exposure” names an output, but does not say for whom, at what stage, or for what decision setting. “Support dose selection” names a purpose, but remains unclear without the population and the quantity the model will provide. A useful statement gives readers enough information to recognize both its intended scope and what lies outside that scope.
Six Elements of a Useful Statement
A practical context-of-use statement addresses six elements. They can be written in one or two sentences, or organized as a short structured description when the situation needs more detail.
- Purpose: What is the model intended to contribute? State its role in plain language, such as estimating a quantity or supporting a defined development activity.
- Target population: For which people or other units is the use intended? Specify relevant features such as age group, disease, treatment, or development stage. Avoid treating “patients” as a sufficiently precise population when the intended group is narrower.
- Decision setting: What kind of development decision or assessment will the model inform, and at what stage or time? Name the setting without implying that the model alone makes the decision.
- Intended users: Who is expected to interpret or use the model outputs? For example, this may be a clinical pharmacology team preparing an internal development assessment.
- Outputs: What quantities or summaries will be provided, at what time scale, and in what units where applicable? An output should be specific enough that a reader can tell what the model will report.
- Boundaries: What important populations, settings, inputs, or uses are outside the stated scope? Boundaries prevent a statement from being read more broadly than intended.
The six elements are not a checklist that must produce six separate sentences. They are prompts for checking completeness. A concise statement can combine elements, but it should not hide a consequential omission behind broad words such as “support,” “patients,” or “results.”
Worked Example: Exposure Assessment in an Adult Trial
Hypothetical setting. A development team plans an internal assessment of drug exposure during a 12-week trial of an investigational medicine in adults with a specified inflammatory disease. The intended regimen is given once daily, and the model will summarize exposure at steady state.
Purpose and decision setting. The purpose is to support the team’s assessment of the proposed regimen during planning for the next adult trial. This names the contribution and setting without claiming the model will decide whether the regimen proceeds.
Population and users. The target population is adults with the specified disease receiving the once-daily regimen. The intended users are the clinical pharmacology and clinical development teams. These details keep the statement from being mistaken as applying to children, other diseases, or other regimens.
Outputs and boundaries. The output is the model-estimated steady-state area under the concentration-time curve over 24 hours, reported in ng·h/mL, for the stated adult population. The scope excludes pediatric use, other diseases, and regimens other than the once-daily regimen. The final statement is:
Context of use: The model is intended to provide estimates of steady-state 24-hour area under the concentration-time curve, in ng·h/mL, for adults with the specified inflammatory disease receiving the once-daily regimen, to support the clinical pharmacology and clinical development teams’ assessment of that regimen during planning for the next adult trial. It is not intended for pediatric populations, other diseases, or other regimens.
This statement is specific about what is estimated, for whom, by whom, and in which setting. It does not state that the estimate is accurate, that a particular value is acceptable, or that any action must follow. Those claims would require additional specification and supporting evidence.
Building the Statement Without Overclaiming
A reliable way to draft a context of use is to begin with the intended use, then test each part for missing scope. The logic is straightforward: a statement can guide use only if readers can identify what task is intended, where and for whom it applies, what information it provides, and what it does not cover. If one of these is unclear, readers may silently fill in different assumptions.
Write what the model is expected to estimate, summarize, or otherwise contribute. Use a concrete verb rather than “be useful.”
Identify the relevant population and the development stage or decision setting. Include details that distinguish the intended use from plausible broader uses.
State the quantities or summaries and their units or time frame where relevant, and identify who will interpret them.
Name important uses or groups that are not covered. Then read the draft as a person unfamiliar with the project and check whether it could reasonably be interpreted more broadly.
This construction assumes that the intended use can be described at the level needed for the current development discussion. If the population or setting is still undecided, record that uncertainty rather than disguising it with a broad label. A context-of-use statement is a scope description, not a substitute for resolving an unresolved development question.
Worked Example: Pediatric Exposure Projection
Hypothetical setting. A team is considering model-based projections for a future study of children aged 6 to under 12 years with a chronic condition. The medicine is given twice daily. The planned output is predicted concentration at the end of the dosing interval, in ng/mL, for the proposed regimen.
The purpose is to inform planning for the pediatric study, not to claim that the model establishes a pediatric regimen. The population should specify the age range and condition rather than simply say “children.” The setting is planning for a future study, and the intended users are the pediatric clinical pharmacology and development teams. The boundary excludes children outside the stated age range and use in adults.
Context of use: The model is intended to provide projections of concentration at the end of the dosing interval, in ng/mL, for children aged 6 to under 12 years with the specified chronic condition receiving the proposed twice-daily regimen, to inform planning by the pediatric clinical pharmacology and development teams for a future study. The statement does not cover children outside this age range or adult use.
The phrase “to inform planning” states a purpose without promising a particular study outcome. The defined output and population also make the scope more interpretable than a statement such as “the model will support pediatric dosing.”
Context of Use and Nearby Descriptions
A context of use is related to, but distinct from, a model description and a development question. Keeping these descriptions separate prevents a technical account from being mistaken for an intended-use statement.
| Description | What it answers | Example |
|---|---|---|
| Development question | What development uncertainty needs to be addressed? | What exposure is expected under the proposed regimen in the intended population? |
| Model description | What model, data, or methods are being used? | A document describing model structure, inputs, and estimation approach. |
| Context of use | For what purpose, population, setting, users, and outputs is the model intended, and what are its boundaries? | A statement specifying the population, exposure output, trial-planning setting, users, and exclusions. |
The context of use should be understandable without relying on a reader to infer the purpose from a model’s technical name. Conversely, it should not become a lengthy technical specification. The statement establishes scope; supporting documentation can describe how the model is built and what evidence is available.
It is also useful to distinguish a decision setting from a decision rule. A context of use can name the assessment or development activity that model outputs will inform. That does not require the context statement to define a threshold, prescribe an action, or assert that the model output alone determines the outcome. Keeping that distinction explicit avoids overstating the model’s role.
Worked Example: Disease-Progression Summaries for Trial Planning
Hypothetical setting. A team is planning a 16-week study in adults with a specified chronic disease. A model is intended to summarize the expected change in a disease-severity score from baseline to week 16 for the study population. The score is measured in points.
The purpose is to provide a summary for planning the study; the decision setting is planning a future trial. The target population is adults with the specified disease, and the intended users are the clinical development team. The output is the expected change in score, in points, over 16 weeks. The statement does not cover other diseases, pediatric populations, or follow-up periods beyond 16 weeks.
Context of use: The model is intended to summarize the expected change from baseline to week 16 in the disease-severity score, in points, for adults with the specified chronic disease, to support the clinical development team’s planning of a future study. It is not intended for pediatric populations, other diseases, or follow-up periods beyond 16 weeks.
This example shows why both an output and a time frame matter. “Summarize disease progression” leaves open which measure, which interval, and which population. Naming them makes the intended use easier to understand without claiming that the summary is appropriate for any other setting.
Reviewing for Ambiguity and Failure Modes
A draft can be polished and still fail to define its scope. Review it by asking whether a reader could identify each element without needing to guess. If two reasonable readers could infer different populations, outputs, or intended roles, revise the wording.
- Purpose stated too broadly: “The model supports development” does not say what contribution it makes or which activity it supports.
- Population left generic: “Patients” may conceal important distinctions in age, disease, treatment, or setting.
- Output left unnamed: “Model results” does not identify a quantity, summary, time point, or unit.
- Users omitted: Without intended users, it may be unclear who is expected to interpret the output and in what organizational setting.
- Boundaries absent: Readers may assume the statement covers populations or uses that were never considered.
- Capability confused with intended use: A model may be technically capable of producing additional outputs. That does not make those outputs part of the stated context of use.
A common mistake is to write, “The model is validated for patients and can support dose selection,” then treat that as a complete context of use. The reasoning is flawed because “patients” does not define a target population, “validated” does not identify the intended purpose or setting, and “support dose selection” does not specify users, outputs, or boundaries. The correction is to state those elements directly and reserve claims about evidence for the appropriate supporting assessment.
There are limits to what a context-of-use statement can accomplish. A clear statement cannot guarantee that a model is adequate, repair weak supporting data, or remove uncertainty about its outputs. Nor should the statement promise more precision than the development question and available information justify. Its value is that it makes the intended scope visible, so model use and subsequent assessment can be discussed against the same stated purpose.
Check Your Understanding
For each item, identify what needs clarification or draft a more complete statement. Suggested solutions follow the questions.
- A statement says, “The model supports treatment planning in patients.” Name three missing or ambiguous elements.
- A model will provide predicted concentration in mg/L for adults with a specified disease receiving a once-daily regimen, for an internal assessment before a planned trial. Name one useful boundary and one possible intended user group.
- Explain why “The model is validated for pediatric use” is not, by itself, a complete context of use.
- Draft a one-sentence context of use for a hypothetical model that estimates a 24-hour exposure summary in adults with a specified disease to inform planning for a later adult study. State one exclusion.
- Does a context-of-use statement need to prescribe the threshold at which a model output triggers an action? Explain.
Solutions: (1) The purpose is vague; “patients” does not define the target population; and the decision setting, users, outputs, and boundaries are not stated. Any three of these identify genuine gaps. (2) A boundary could exclude pediatric use or regimens other than once daily; possible users include the clinical pharmacology and clinical development teams. (3) It does not identify the pediatric age range or condition, the model’s purpose, the decision setting, intended users, outputs, or boundaries. “Validated” is not a description of intended use. (4) One possible answer is: “The model is intended to estimate the 24-hour exposure summary for adults with the specified disease to inform the clinical development team’s planning of a later adult study; it is not intended for pediatric use.” Other clear statements are acceptable if they specify the requested elements. (5) No. The statement should identify the decision setting the outputs will inform, but prescribing an action threshold is a separate specification and is not necessary to define the context of use.