This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for A Safety Study of Epoetin Alfa in Patients With Cancer Who Have Chemotherapy-Related Anemia (NCT01394991), a randomized phase 4 study. The primary outcome is Number of Participants With at Least 1 Clinically Relevant and Objectively Confirmed Thrombovascular Event From Randomization Through Week 16.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT01394991 |
| Phase | Phase 4 |
| Condition | Anemia, Neoplasms |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 504 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Johnson & Johnson Pharmaceutical Research & Development, L.L.C. |
| Results posted | Yes |
2. Clinical Question
Population
Participants had cancer and chemotherapy-related anemia. The primary analysis population was the modified intent-to-treat population: randomized participants who received at least 1 dose of study drug. The registry record reports enrollment of 504 participants.
Intervention
Epoetin alfa 450 IU/kg once a week, identified as QW. It was given by subcutaneous injection, preferably in the abdomen, for up to 4 weeks after the last chemotherapy dose, for a maximum of 26 weeks.
Comparator
Epoetin alfa 150 IU/kg 3 times a week, identified as TIW. It was also given by subcutaneous injection, preferably in the abdomen, for up to 4 weeks after the last chemotherapy dose, for a maximum of 26 weeks.
Primary question
From randomization through Week 16, how many participants in each group had at least 1 clinically relevant and objectively confirmed thrombovascular event? The registry also reports comparisons of the event counts between the randomized groups.
The trial’s stated purpose was to evaluate the safety of epoetin alfa in patients with cancer receiving chemotherapy. Its primary endpoint focuses on specified thrombovascular events rather than on anemia response. The registry lists additional outcomes, including suspected and adjudicated events over a longer timeframe, time to first event, mortality, hemoglobin response, and red blood cell transfusions.
3. Trial Design
This was a randomized, open-label, multicenter, parallel-group phase 4 study. Participants were assigned to one of two epoetin alfa schedules: 450 IU/kg once a week or 150 IU/kg 3 times a week. The registry reports enrollment of 504. The study started in 2006-01, primary completion was 2009-09, and its status is completed.
Because the study was open-label, participants and study personnel were not masked to treatment assignment. The registry record does not report further details about the randomization procedure, allocation concealment, or the number and locations of participating centers. These design details therefore cannot be assessed from the record.
The primary endpoint counted participants with at least 1 clinically relevant and objectively confirmed thrombovascular event from randomization through Week 16. The registry identifies the analysis population as modified intent-to-treat: all randomized participants who received at least 1 dose of study drug. The reported group sizes for this endpoint were 242 in the QW group and 262 in the TIW group.
The intervention descriptions specify subcutaneous administration, preferably in the abdomen, continuing for up to 4 weeks after the last dose of chemotherapy and for a maximum of 26 weeks. The registry lists 7 secondary outcomes, including outcomes covering the study period through Week 26. It does not report the protocol’s sample-size justification, planned interim analyses, or a complete statistical analysis plan.
4. Participant Flow and Accounting
The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.
Overall Study
| Stage | Epoetin Alfa QW | Epoetin Alfa TIW |
|---|---|---|
| STARTED | 242 | 262 |
| COMPLETED | 183 | 190 |
| NOT COMPLETED | 59 | 72 |
Reasons for not completing this period:
| Reason | Epoetin Alfa QW | Epoetin Alfa TIW |
|---|---|---|
| Adverse Event | 0 | 4 |
| Lack of Efficacy | 5 | 7 |
| Lost to Follow-up | 6 | 8 |
| Physician Decision | 1 | 3 |
| Protocol Violation | 1 | 1 |
| Withdrawal by Subject | 20 | 17 |
| Chemotherapy Termination | 10 | 15 |
| Disease Progression | 7 | 7 |
| Prematurely Termination of Trial | 9 | 10 |
Analysis population for the primary outcome, as defined in the registry: The modified intent-to-treat (mITT) population was defined to include all randomized participants who received at least 1 dose of study drug.
5. Baseline Characteristics
The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.
| Characteristic | Unit | Epoetin Alfa QW | Epoetin Alfa TIW | Total |
|---|---|---|---|---|
| Participants analysed | 242 | 262 | 504 | |
| Age, Categorical: <=18 years | Participants | 1 | 1 | 2 |
| Age, Categorical: Between 18 and 65 years | Participants | 160 | 173 | 333 |
| Age, Categorical: >=65 years | Participants | 81 | 88 | 169 |
| Age Continuous | years | 58.8 (SD 13.35) | 57.5 (SD 13.53) | 58.1 (SD 13.44) |
| Sex: Female, Male: Female | Participants | 156 | 181 | 337 |
| Sex: Female, Male: Male | Participants | 86 | 81 | 167 |
| Region of Enrollment: Bulgaria | participants | 6 | 7 | 13 |
| Region of Enrollment: France | participants | 7 | 6 | 13 |
| Region of Enrollment: Germany | participants | 51 | 55 | 106 |
| Region of Enrollment: Great Britain | participants | 1 | 4 | 5 |
| Region of Enrollment: Greece | participants | 8 | 12 | 20 |
| Region of Enrollment: Italy | participants | 13 | 13 | 26 |
| Region of Enrollment: Poland | participants | 7 | 6 | 13 |
| Region of Enrollment: Romania | participants | 21 | 20 | 41 |
| Region of Enrollment: Russia | participants | 88 | 93 | 181 |
| Region of Enrollment: Slovakia | participants | 9 | 11 | 20 |
6. Primary Endpoint
The registry lists the primary outcome as Number of Participants With at Least 1 Clinically Relevant and Objectively Confirmed Thrombovascular Event From Randomization Through Week 16, measured over from randomization through Week 16.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | particpants |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint was the number of participants with at least 1 clinically relevant and objectively confirmed thrombovascular event from randomization through Week 16. It is a participant-level count of people experiencing an event, not a count of every event experienced by each person. A participant with more than 1 event would still be counted once for this endpoint.
The registry defines clinically relevant events to include deep vein thrombosis of the limbs; thromboses of other major veins; pulmonary embolism; acute coronary syndrome; ischemic stroke of arterial or cardiac origin; cerebral venous thrombosis; and arterial thrombosis. An event was considered objectively confirmed when the clinical diagnosis was confirmed by appropriate medical imaging studies and laboratory tests. An Adjudication Committee determined whether events met the criteria.
Focusing on clinically relevant, objectively confirmed events provides a defined safety outcome rather than relying only on suspected events or unverified clinical impressions. The adjudication process is intended to apply the event definitions consistently. As a general statistical point, a participant-level endpoint can be compared between groups as the proportion experiencing at least 1 event, while retaining the clinically interpretable question of how many people were affected. The registry’s stated primary endpoint and reported analyses should be distinguished from the separate secondary outcome that counts positively adjudicated events through Week 26.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
For the primary endpoint, 5 participants in the QW group and 10 in the TIW group had at least 1 qualifying event through Week 16. The registry reports a chi-squared analysis of the risk difference, with an estimate of -0.018 and a 95% confidence interval from -0.051 to 0.016; the reported p-value is 0.248. The negative estimate reflects the reported comparison’s direction, QW relative to TIW. The interval includes zero, so the posted result is compatible with no difference as well as with differences in either direction under this measure.
The registry also reports logistic regression, with an odds ratio of 0.53 and a 95% confidence interval from 0.18 to 1.58; the p-value is 0.254. The interval includes 1, the value representing equal odds between groups. Both reported analyses therefore leave substantial uncertainty about the relative event frequency. The point estimates favor fewer events in the QW group, but these data do not establish a difference between schedules.
The primary analysis is based on relatively few participants with events. When event counts are limited, estimates can be imprecise, and a confidence interval can span meaningfully different possibilities. The point estimate alone should not be treated as a definitive description of the underlying difference. The interval and the number of observed events are essential context.
The primary result is specific to the endpoint definition and observation window: adjudicated, clinically relevant, objectively confirmed events through Week 16. It does not by itself describe all suspected events, events occurring later, or other outcomes such as mortality, hemoglobin response, or transfusions. Those outcomes are separately listed in the registry.
The registry’s secondary results also include comparisons through Week 26. They address related but not identical questions, and they do not replace the prespecified primary endpoint. A pattern across secondary outcomes may provide context, but multiple comparisons and the exploratory nature of unadjusted results limit how strongly they can support an independent conclusion. The registry record does not report whether a multiplicity adjustment was planned or used.
These findings describe the posted analyses in this trial. They do not establish that either schedule is safer or more effective in other populations or settings. The registry record does not report enough detail about missing observations, follow-up completeness, or the full analysis plan to resolve those additional sources of uncertainty.
9. Statistical Methods Explained
The registry reports different statistical methods for the primary and secondary comparisons. These methods answer related but distinct questions. General explanations below describe how the methods are ordinarily interpreted; they do not add assumptions about the study’s unreported analysis plan.
Risk difference
A risk difference compares the proportion of participants with the event in one group with the proportion in the other. A value of zero represents equal event proportions. The primary reported risk difference is -0.018, with a 95% confidence interval from -0.051 to 0.016. The comparison is presented as QW relative to TIW, so the negative point estimate corresponds to a lower observed event proportion in QW. Its interval includes zero, leaving uncertainty about the direction and size of the underlying difference.
A confidence interval gives a range of values compatible with the data and the specified statistical method, subject to its assumptions. It is not a guarantee that the true value lies within that particular interval. A wider interval indicates less precision than a narrower interval. Here, the interval’s inclusion of zero means the reported risk-difference analysis does not rule out no difference.
Chi-squared test and p-value
The primary risk-difference comparison uses a chi-squared method. In general, a chi-squared test evaluates whether observed group counts are compatible with a null hypothesis of no association or difference, under the test’s assumptions. The primary p-value is 0.248. A p-value describes how unusual results at least as discrepant as those observed would be if the null hypothesis and the model assumptions held. It is not the probability that the null hypothesis is true, and it does not measure the clinical importance of an effect.
The p-value should be read alongside the effect estimate and confidence interval. It does not show that the groups are equivalent, nor does it establish that a difference exists. The registry record does not report an equivalence or non-inferiority framework for this primary comparison.
Odds ratio and logistic regression
The second primary analysis uses logistic regression and reports an odds ratio of 0.53 with a 95% confidence interval from 0.18 to 1.58. An odds ratio compares the odds of an outcome between groups; an odds ratio of 1 represents equal odds. The interval includes 1, so the result is compatible with equal odds as well as with a range of differences. Odds and risks are related but are not interchangeable measures, so the odds ratio should not be read as a risk difference.
The registry gives a p-value of 0.254 for this analysis. The record does not report whether the logistic regression included covariates, how any model assumptions were assessed, or whether the analysis adjusted for other factors. Those details are not reported in the registry record.
Time-to-event analysis
For time to first positively adjudicated thrombovascular event, the registry reports a log-rank analysis and a hazard ratio of 0.40, with a 95% confidence interval from 0.14 to 1.13 and a p-value of 0.073. In general, a time-to-event analysis uses information about when an event occurs and can account for participants whose event time is not observed before follow-up ends, subject to assumptions about censoring. A hazard ratio compares event rates over time under the model; it is not the same as a ratio of event counts or risks at a single timepoint. The interval includes 1, so the estimate remains uncertain.
The registry record does not report the event-time distributions, censoring details, or checks of the time-to-event model assumptions. The hazard ratio should therefore be interpreted as the reported analysis, not as a fully assessable description of the underlying event-time data.
Planning illustrations
The sample-size illustrations use observed effects to show how participant requirements could vary if a future study aimed to detect a specified fraction of an observed effect. Their stated assumptions include a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustration is not the trial’s own design justification and does not describe the study’s actual power. Sample-size requirements based on observed effects are planning illustrations, not observed power. The registry record does not report the original sample-size justification or its assumptions.
For time-to-event illustrations, the stated method uses the total number of events; the other illustrations express participants per arm. General statistical reasoning explains why smaller target effects require larger studies: a smaller separation between groups is harder to distinguish from random variation. The registry’s planning figures should be understood in that limited, hypothetical sense, not as evidence that the completed study was adequately powered for its endpoints.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Chi-squared | Risk Difference (RD) -0.018 | -0.051 to 0.016 | 0.248 |
| Regression, Logistic | Odds Ratio (OR) 0.53 | 0.18 to 1.58 | 0.254 |
The registry reports two sponsor analyses for the primary endpoint. The chi-squared analysis gives a risk difference of -0.018, with a 95% confidence interval from -0.051 to 0.016 and a p-value of 0.248. The logistic regression analysis gives an odds ratio of 0.53, with a 95% confidence interval from 0.18 to 1.58 and a p-value of 0.254. Both intervals include their respective null values, and both analyses leave uncertainty about the size and direction of the group difference.
A simple unadjusted recalculation is not reported for the primary endpoint because the available endpoint data are classified as a count outcome for which a two-arm recalculation was not estimable. There is therefore no recalculated primary estimate to compare numerically with the sponsor’s analyses. The sponsor analyses are the registry’s reported inferential results for this endpoint.
For some secondary outcomes, the registry also contains sponsor analyses and simple unadjusted recalculations. Those methods and estimates need not be identical, and their differences should not be treated as a correction to the sponsor’s analysis. The primary interpretation should remain anchored to the endpoint-specific sponsor results and their uncertainty, rather than to secondary recalculations.
11. Multiplicity, Missing Data and Analysis Populations
The registry lists 8 outcome measures and 9 analyses. The primary endpoint has two reported analyses, using a chi-squared method and logistic regression. Several secondary outcomes also have reported analyses. Testing multiple outcomes or using multiple analyses creates a multiplicity issue: when many hypotheses are tested, the chance of at least one apparently unusual result can increase. A prespecified adjustment or testing hierarchy can address this concern, but the registry record does not report whether such a strategy was planned or applied.
Without information on the intended hierarchy, it is not possible to determine which analyses were confirmatory and which were exploratory. Secondary p-values should therefore be interpreted cautiously, particularly when they are considered alongside several other outcomes. The registry record does not report a comprehensive statistical analysis plan, interim analysis plan, or stopping rule.
The primary analysis population is defined as randomized participants who received at least 1 dose of study drug. This modified intent-to-treat definition is narrower than an analysis of every randomized participant. The registry record does not report how many randomized participants did not receive a dose, how many participants had incomplete follow-up, or how missing endpoint assessments were handled.
Missing-data methods can matter because excluding participants with unavailable outcomes may affect the comparison if outcome availability differs between groups or is related to event risk. The registry record does not report imputation methods, sensitivity analyses for missing data, or follow-up completeness for the primary endpoint. It also does not report whether the primary comparison was adjusted for baseline characteristics. These omissions limit assessment of the robustness of the reported estimates, beyond the uncertainty already reflected in their confidence intervals.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Epoetin Alfa QW | 53 | 242 | 21.9% |
| Epoetin Alfa TIW | 64 | 262 | 24.4% |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Epoetin Alfa QW | 157 of 242 |
| Epoetin Alfa TIW | 151 of 262 |
Most frequent serious adverse event terms
| Event term | Organ system | Epoetin Alfa QW | Epoetin Alfa TIW |
|---|---|---|---|
| Malignant neoplasm progression | Neoplasms benign, malignant and unspecified (incl cysts and polyps) | 19 of 242 | 20 of 262 |
| Pneumonia | Infections and infestations | 2 of 242 | 9 of 262 |
| Deep vein thrombosis | Vascular disorders | 3 of 242 | 4 of 262 |
| Anaemia | Blood and lymphatic system disorders | 4 of 242 | 2 of 262 |
| Neutropenia | Blood and lymphatic system disorders | 2 of 242 | 3 of 262 |
| Thrombocytopenia | Blood and lymphatic system disorders | 1 of 242 | 4 of 262 |
| Febrile neutropenia | Blood and lymphatic system disorders | 4 of 242 | 0 of 262 |
| Leukopenia | Blood and lymphatic system disorders | 3 of 242 | 1 of 262 |
| Pulmonary embolism | Respiratory, thoracic and mediastinal disorders | 0 of 242 | 4 of 262 |
| Pancytopenia | Blood and lymphatic system disorders | 2 of 242 | 1 of 262 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Epoetin Alfa QW | Epoetin Alfa TIW |
|---|---|---|---|
| Neutropenia | Blood and lymphatic system disorders | 81 of 242 | 72 of 262 |
| Thrombocytopenia | Blood and lymphatic system disorders | 52 of 242 | 54 of 262 |
| Leukopenia | Blood and lymphatic system disorders | 51 of 242 | 44 of 262 |
| Nausea | Gastrointestinal disorders | 27 of 242 | 40 of 262 |
| Vomiting | Gastrointestinal disorders | 13 of 242 | 24 of 262 |
| Diarrhoea | Gastrointestinal disorders | 26 of 242 | 10 of 262 |
| Asthenia | General disorders | 21 of 242 | 14 of 262 |
| Fatigue | General disorders | 17 of 242 | 16 of 262 |
| Pyrexia | General disorders | 15 of 242 | 18 of 262 |
| Anorexia | Metabolism and nutrition disorders | 13 of 242 | 11 of 262 |
Serious events with the highest arm-level rate: malignant neoplasm progression (7.9%), pneumonia (3.4%), anaemia (1.7%), febrile neutropenia (1.7%), thrombocytopenia (1.5%).
For the serious adverse event summary, 53 of 242 participants in the QW group and 64 of 262 in the TIW group were reported as affected. The registry also reports other adverse events in 157 of 242 participants in the QW group and 151 of 262 in the TIW group. These are participant counts, not the primary endpoint’s adjudicated thrombovascular-event counts. A participant may have experienced more than one adverse event, so the figures should not be interpreted as counts of all events.
The registry lists malignant neoplasm progression (7.9%), pneumonia (3.4%), anaemia (1.7%), febrile neutropenia (1.7%), thrombocytopenia (1.5%), and deep vein thrombosis (1.5%) among the top serious terms. The record does not specify an arm for these percentages in the available details, so they should not be attributed to one treatment group or treated as arm-specific comparisons.
Deaths are not reported in the available safety results. These summaries describe events recorded during the study; by themselves, they do not establish that treatment caused an event or explain how events differed in severity, timing, or duration.
13. What the Result Does, and Does Not, Mean
The primary result describes a comparison of two dosing schedules among participants with cancer and chemotherapy-related anemia who received at least 1 dose of study drug. The endpoint is narrowly defined: at least 1 clinically relevant and objectively confirmed thrombovascular event from randomization through Week 16. The posted counts were 5 in the QW group and 10 in the TIW group.
The sponsor’s estimated risk difference and odds ratio both point toward fewer observed primary events in the QW group, but their confidence intervals include the respective null values. The data therefore do not establish a difference between schedules on this endpoint. They also do not establish equivalence: a result that does not rule out no difference is not proof that the schedules have the same risk. The intervals remain compatible with a range of effects.
The result does not answer every safety question. Secondary outcomes cover other event definitions and a longer timeframe, and the safety summary includes adverse events beyond thrombovascular events. No single endpoint represents the full safety profile. The registry record also does not report all details needed to assess missing observations, follow-up completeness, multiplicity, or the planned analysis strategy.
These results should not be generalized automatically to other populations, treatment settings, or dosing regimens. They describe the registry’s reported comparison for this trial and its defined analysis population. They do not, by themselves, support medical advice or a conclusion about treatment benefit or harm outside the measured outcomes.
14. Why This Trial Matters Statistically
This trial illustrates why a safety comparison should be read through its endpoint definition, analysis population, and observation window. The primary outcome is not a broad count of every adverse event: it requires a clinically relevant thrombovascular event that is objectively confirmed and adjudicated. That specificity helps define what is being compared, while also limiting the result to those qualifying events.
The example also shows why point estimates need their uncertainty. The observed primary counts differ between groups, and the estimated direction favors fewer events in QW. But the reported confidence intervals include the null values, so the apparent difference is not a precise or conclusive estimate of a treatment contrast. Sparse outcome counts can leave substantial uncertainty even in a randomized study.
Different statistical methods express the comparison on different scales. The risk difference describes an absolute contrast in event proportions; the odds ratio compares odds; and a hazard ratio from a time-to-event analysis compares event rates over time under its assumptions. These estimates are not interchangeable, and each depends on the endpoint and analysis used.
Finally, registry records can leave important analytical details unreported. Here, the record identifies the primary analysis population and gives sponsor estimates, but does not report a complete multiplicity plan, missing-data approach, or sample-size justification. Recognizing those gaps is part of responsible interpretation: the posted results can be described precisely without claiming more certainty than they support.
15. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
The primary endpoint is based on few observed participants with events, which limits precision. Its result applies only to events meeting the stated clinical relevance and objective confirmation criteria during the primary timeframe. Suspected events and events recorded later are separate outcomes and should not be merged with the primary result.
The study was open-label. The primary event definition includes objective confirmation and adjudication, but the registry record does not report whether event assessors were masked to treatment assignment or how adjudication disagreements were resolved. It also does not report detailed randomization procedures or allocation concealment. These details would help readers assess how the design and outcome assessment were implemented.
The registry record does not report missing-data handling, completeness of follow-up, baseline group characteristics, or whether analyses adjusted for other factors. It also lacks a full statistical analysis plan and information on multiplicity control. These omissions limit assessment of potential bias and the robustness of the reported analyses.
The primary analysis includes randomized participants who received at least 1 dose, rather than all randomized participants, so the exact relationship between this analysis population and the full randomized cohort is not described. The record does not report the protocol’s sample-size justification or planned power for the primary endpoint. Sample-size planning illustrations based on observed effects cannot fill that gap or establish the study’s original design rationale.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: A Safety Study of Epoetin Alfa in Patients With Cancer Who Have Chemotherapy-Related Anemia — NCT01394991.
- Registry record last updated 2012-04-05; status last verified 2012-04.
- Results first posted 2011-07-15.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
NCT01394991 was a completed, randomized, open-label, multicenter, parallel-group phase 4 study of epoetin alfa in patients with cancer and chemotherapy-related anemia. It compared 450 IU/kg once a week with 150 IU/kg 3 times a week. Enrollment was 504; the study started in 2006-01 and primary completion was 2009-09.
The primary endpoint was the number of participants with at least 1 clinically relevant, objectively confirmed thrombovascular event from randomization through Week 16. The modified intent-to-treat population included randomized participants who received at least 1 dose. The registry reports 5 events in QW and 10 in TIW, with a risk difference of -0.018 (95% confidence interval -0.051 to 0.016; p = 0.248) and an odds ratio of 0.53 (95% confidence interval 0.18 to 1.58; p = 0.254).
The reported estimates do not establish a difference between schedules. Secondary outcomes and adverse-event summaries provide additional context, but the registry record does not report several details needed for a complete assessment of multiplicity, missing data, and the analysis plan.