WHY ENGLISH?
Write prediction-model decisions before results arrive, then make every change visible and explainable
A prediction model protocol is a promise made before the results are known. English matters because a vague promise can be rewritten after the data arrive, while a precise protocol makes objectives, populations, predictors, outcomes, analysis choices and validation plans open to later comparison.
TRIPOD-P was registered with EQUATOR on 5 September 2026 as a reporting guideline under development for protocols of clinical prediction-model and predictive-AI studies. The proposal describes it as an extension of TRIPOD+AI for diagnostic and prognostic studies that develop or validate models using statistical or machine-learning methods.
The registered aim is a minimal set of essential protocol-reporting items across clinical specialties. As checked on 6 November 2026, the EQUATOR page did not publish a final checklist. This guide therefore explains the protocol logic documented by the proposal and established prediction-study principles without inventing final TRIPOD-P items.
Article contents
Open all chapters
- 1. State the Protocol Status and Version
- 2. Write the Prediction Objective Precisely
- 3. Separate Development, Updating and Validation
- 4. Describe the Intended Use Before the Algorithm
- 5. Define Data Sources, Sites and Dates
- 6. Set Participant Eligibility and Sampling
- 7. Define Predictors and Measurement Timing
- 8. Define the Outcome and Assessment
- 9. Plan Sample Size and Information Needs
- 10. Prespecify Missing-Data Handling
- 11. Prespecify Model Development
- 12. Protect Internal and External Validation
- 13. Choose Performance Measures and Uncertainty
- 14. Plan Comparisons and Sensitivity Analyses
- 15. Record Deviations Without Rewriting History
- 16. Plan Access, Reproducibility and a Bounded Claim
ORIENTATION
What this guide owns
Current fact check
EQUATOR lists TRIPOD-P as under development and registered 5 September 2026. The proposal says it will extend TRIPOD+AI, cover protocols for diagnostic and prognostic prediction-model development and validation using statistical or machine-learning methods, and seek a minimal essential item set across specialties.
Canonical boundary
This article owns protocol-stage planning for prediction models under the registered TRIPOD-P proposal. TRIPOD+AI owns completed prediction-model reports, TRIPOD-LLM owns LLM studies, and DEAL owns LLM development and application reporting.
Worked example
Every chapter returns to a fictional multicentre protocol for developing and externally validating a model that predicts which adults with asthma may need urgent review within seven days. It is fictional, low-stakes and educational; it does not report a real school, clinical decision or guaranteed outcome.
Did You Know? For tripod_p, complete reporting makes appraisal possible but does not perform the appraisal, remove bias or guarantee transfer.
Warm up with eduKate’s guides to explaining scientific evidence and asking better questions in class.
CHAPTER 1 OF 16
1. State the Protocol Status and Version
Why this move matters
Give registration, version, date, amendment history and where readers can find the current protocol.
Did You Know? Protocols become auditable when versions and changes are visible.
Write the protocol before seeing outcomes
Give registration, version, date, amendment history and where readers can find the current protocol. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The team timestamps version 1.0 before data extraction and records every amendment. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A registration date does not prove that analysis followed the plan. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “State the Protocol Status and Version”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 2 OF 16
2. Write the Prediction Objective Precisely
Why this move matters
Name diagnostic or prognostic purpose, target population, outcome, prediction time, intended user and decision context.
Did You Know? TRIPOD-P is intended for diagnostic and prognostic prediction studies.
Write the protocol before seeing outcomes
Name diagnostic or prognostic purpose, target population, outcome, prediction time, intended user and decision context. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The model predicts urgent review within seven days at the point of initial assessment. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A clear objective can still address an unhelpful or harmful decision. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Write the Prediction Objective Precisely”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 3 OF 16
3. Separate Development, Updating and Validation
Why this move matters
State whether the study creates a model, updates an existing model, tests transportability or combines these aims.
Did You Know? Different study purposes require different data boundaries and claims.
Write the protocol before seeing outcomes
State whether the study creates a model, updates an existing model, tests transportability or combines these aims. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: Development uses three hospitals; a fourth is reserved for external validation. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Calling a split sample external validation does not make it independent across settings. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Separate Development, Updating and Validation”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 4 OF 16
4. Describe the Intended Use Before the Algorithm
Why this move matters
Explain how predictions might enter workflow, what action could follow and what comparison represents current practice.
Did You Know? Clinical usefulness begins with a decision, not a model family.
Write the protocol before seeing outcomes
Explain how predictions might enter workflow, what action could follow and what comparison represents current practice. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The score supports clinician review rather than automatic referral. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A proposed workflow may change before deployment. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Describe the Intended Use Before the Algorithm”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 5 OF 16
5. Define Data Sources, Sites and Dates
Why this move matters
Name records, collection purpose, locations, time windows, linkage, access and whether data are prospective or retrospective.
Did You Know? Data provenance constrains who and what the model can represent.
Write the protocol before seeing outcomes
Name records, collection purpose, locations, time windows, linkage, access and whether data are prospective or retrospective. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The protocol lists each hospital system and the exact years contributing data. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Routinely collected data can be systematically incomplete. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define Data Sources, Sites and Dates”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 6 OF 16
6. Set Participant Eligibility and Sampling
Why this move matters
Prespecify inclusion, exclusions, repeated episodes, sampling, case definition and handling of people with incomplete follow-up.
Did You Know? Population language determines the denominator behind every later metric.
Write the protocol before seeing outcomes
Prespecify inclusion, exclusions, repeated episodes, sampling, case definition and handling of people with incomplete follow-up. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: Only first eligible episodes are used for development and repeat visits follow a stated rule. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Eligibility rules can create a cleaner population than practice. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Set Participant Eligibility and Sampling”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 7 OF 16
7. Define Predictors and Measurement Timing
Why this move matters
List candidate predictors, definitions, units, availability at prediction time, measurement reliability and planned transformations.
Did You Know? Timing protects a protocol from information leakage.
Write the protocol before seeing outcomes
List candidate predictors, definitions, units, availability at prediction time, measurement reliability and planned transformations. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: Only information available at initial assessment can enter the model. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A predictor can be available in records but unavailable at the real decision time. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define Predictors and Measurement Timing”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 8 OF 16
8. Define the Outcome and Assessment
Why this move matters
Specify outcome, horizon, ascertainment source, blinding, competing events, adjudication and uncertain cases.
Did You Know? Outcome definitions create the target the model learns.
Write the protocol before seeing outcomes
Specify outcome, horizon, ascertainment source, blinding, competing events, adjudication and uncertain cases. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: Urgent review is confirmed through a blinded outcome process within seven days. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
An operational outcome may imperfectly represent clinical need. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define the Outcome and Assessment”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 9 OF 16
9. Plan Sample Size and Information Needs
Why this move matters
Explain the rationale for participants, events, predictors, precision and expected missingness rather than using a simple cases-per-variable slogan.
Did You Know? Sample-size planning is about information for the intended analysis.
Write the protocol before seeing outcomes
Explain the rationale for participants, events, predictors, precision and expected missingness rather than using a simple cases-per-variable slogan. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The protocol justifies anticipated events and precision for calibration estimates. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Assumptions may be wrong when prevalence shifts. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Plan Sample Size and Information Needs”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 10 OF 16
10. Prespecify Missing-Data Handling
Why this move matters
Describe missingness exploration, imputation, variables in the imputation model, number of datasets and sensitivity analyses.
Did You Know? A missing-data plan reduces result-driven choices.
Write the protocol before seeing outcomes
Describe missingness exploration, imputation, variables in the imputation model, number of datasets and sensitivity analyses. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The team avoids complete-case analysis as an unexplained default. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Imputation cannot recover information that was never measured. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Prespecify Missing-Data Handling”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 11 OF 16
11. Prespecify Model Development
Why this move matters
State model family, functional forms, interactions, shrinkage, variable handling, tuning, optimisation and software environment.
Did You Know? Model development is a sequence of decisions, not a single command.
Write the protocol before seeing outcomes
State model family, functional forms, interactions, shrinkage, variable handling, tuning, optimisation and software environment. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The protocol names the modelling pipeline without promising one algorithm will win. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A detailed pipeline can still overfit if information is limited. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Prespecify Model Development”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 12 OF 16
12. Protect Internal and External Validation
Why this move matters
Describe resampling, nested tuning, held-out sites or times, access controls and how performance will be estimated without leakage.
Did You Know? Validation evidence depends on an honest information boundary.
Write the protocol before seeing outcomes
Describe resampling, nested tuning, held-out sites or times, access controls and how performance will be estimated without leakage. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: Hyperparameters are selected inside development resampling; the external hospital remains sealed. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
One external site cannot establish universal transportability. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Protect Internal and External Validation”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 13 OF 16
13. Choose Performance Measures and Uncertainty
Why this move matters
Prespecify discrimination, calibration, clinical-utility measures, thresholds, confidence intervals and subgroup analyses with rationale.
Did You Know? Measures answer different questions and should not be collapsed into one score.
Write the protocol before seeing outcomes
Prespecify discrimination, calibration, clinical-utility measures, thresholds, confidence intervals and subgroup analyses with rationale. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The protocol includes calibration plots and decision-curve analysis, not accuracy alone. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
A prespecified metric can still be clinically irrelevant. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Choose Performance Measures and Uncertainty”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 14 OF 16
14. Plan Comparisons and Sensitivity Analyses
Why this move matters
Name current practice, simple baselines, alternative definitions, site or time checks and analyses for influential assumptions.
Did You Know? A protocol should say which analysis is primary and why.
Write the protocol before seeing outcomes
Name current practice, simple baselines, alternative definitions, site or time checks and analyses for influential assumptions. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The model is compared with clinician judgement and a transparent baseline under the same outcome. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Multiple analyses can multiply opportunities for selective emphasis. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Plan Comparisons and Sensitivity Analyses”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 15 OF 16
15. Record Deviations Without Rewriting History
Why this move matters
Define how amendments, unexpected data problems and analysis deviations will be dated, justified and reported.
Did You Know? Protocols are useful because later reports can be compared against them.
Write the protocol before seeing outcomes
Define how amendments, unexpected data problems and analysis deviations will be dated, justified and reported. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: A coding change after data inspection appears in a deviation table with its effect. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Disclosure does not erase bias, but concealment makes appraisal impossible. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Record Deviations Without Rewriting History”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
CHAPTER 16 OF 16
16. Plan Access, Reproducibility and a Bounded Claim
Why this move matters
State availability of code, data dictionaries, model specification and restrictions, then limit conclusions to the planned evidence.
Did You Know? As TRIPOD-P develops, readers should check the live EQUATOR record for the final item set.
Write the protocol before seeing outcomes
State availability of code, data dictionaries, model specification and restrictions, then limit conclusions to the planned evidence. Fix the objective, data source, participants, predictors, outcome, modelling strategy and validation plan before results can steer the language.
Trace one prediction study from plan to audit
Worked example: The team plans a model card, code release and controlled validation dataset. Connect each planned decision to a timestamp, rationale, responsible person and deviation rule.
Separate development from validation
State whether the study develops, updates or validates a model and keep training, tuning and final evaluation information boundaries visible.
Respect the proposal boundary
Open code cannot compensate for inaccessible data or weak design. TRIPOD-P is an under-development extension of TRIPOD+AI. It is intended to define a minimal protocol set, but a final checklist was not published on the checked EQUATOR record.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Plan Access, Reproducibility and a Bounded Claim”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports clinical epidemiology, biostatistics, data science, medicine, health research, peer review, research governance and technical writing without promising a grade, admission, funding decision or job.
PRACTICE LAB
Build a tripod_p decision trail
Student route
- Choose one original report inside this guide’s scope.
- Copy its objective and underline actor, action, context, evidence and intended decision.
- Draw the specific reasoning chain used in this article.
- Mark each step reported, inferred, contested or unknown.
- Rewrite the conclusion so it preserves the narrowest important limitation.
Parent or teacher route
Ask the learner to explain write prediction-model decisions before results arrive, then make every change visible and explainable twice: once to a classmate and once to a method specialist. Praise a precise correction and a well-stated unknown; both show disciplined attention to evidence.
Early-career route
Create a one-page tripod_p audit with source link, date checked, scope, method logic, decisive wording, contrary evidence, limit and next action. A colleague should be able to continue without guessing.
Decision rule
If a shorter explanation removes the condition that made the original claim true, it is not clearer yet. Restore the condition.
Continue with eduKate’s guides to reading course and career information and giving useful feedback in group projects.
USEFUL NEXT READING
Connect this narrow owner to its neighbour
Read the ELEVATE-GenAI health-economics reporting guide for a related but different reader job. Return here when the question is specifically how to write prediction-model decisions before results arrive, then make every change visible and explainable. This boundary keeps one canonical owner per intent.
For English practice across research, school and career communication, visit the English at work and in public life guide.
FREQUENTLY ASKED QUESTIONS
Questions students and parents often ask
What is TRIPOD-P?
A proposed TRIPOD+AI extension for protocols of diagnostic and prognostic prediction-model studies.
Is it final?
No. The checked EQUATOR record lists it as under development.
Does it cover machine learning?
The proposal explicitly includes statistical and machine-learning methods.
Why write a protocol before analysis?
It separates planned decisions from choices influenced by observed results.
Is internal validation enough?
It helps estimate optimism but does not by itself establish transportability to new settings.
Can a protocol change?
Yes, but amendments and deviations should be dated, justified and reported.
PRIMARY SOURCES
Read the original tripod_p guidance
Facts were checked against original publications and official records on 5 November 2026. Follow the live source for later corrections, completed updates or companion material.
A cheerful final note
Precise English makes write prediction-model decisions before results arrive, then make every change visible and explainable more generous, not less. People can see how their evidence and values entered the work, where disagreement remains and what a responsible next step looks like.