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 S1415CD, Trial Assessing CSF Prescribing Effectiveness and Risk (TrACER) (NCT02728596), a randomized na study. The primary outcome is Percentage of Participants With CSF Prescribed as Primary Prophylaxis.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02728596 |
| Phase | NA |
| Condition | Febrile Neutropenia, Stage 0 Breast Cancer, Stage 0 Colorectal Cancer |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 3665 |
| Arms in the primary outcome | n/a |
| Lead sponsor | SWOG Cancer Research Network |
| Results posted | Yes |
2. Clinical Question
Population
Participants were receiving chemotherapy for breast, colorectal, or non-small cell lung cancer and were considered at risk for febrile neutropenia. The registry lists cancer stages across these cancer types. The primary analysis concerned participants registered in the trial’s intervention and usual-care components. One participant in Clinic Group 4 had an unavailable primary outcome and was excluded from that analysis, leaving 972 analyzable participants in that group rather than 973.
Intervention
The trial evaluated clinic-based approaches to prescribing primary prophylactic colony stimulating factor (PP-CSF). Clinic Group 3 used automated system recommendations to recommend CSF for drugs with intermediate or high febrile-neutropenia risk, but not low risk. Clinic Group 4 used automated recommendations for high-risk drugs, but not intermediate- or low-risk drugs. The registry describes these as intervention groups.
Comparator
Clinic Group 1 had an existing automated system that recommended CSF for high-risk drugs, but not low-risk drugs. Clinic Group 2 had no automated system and based prescribing on existing clinical practice guidelines. The registry identifies these as active comparator groups. Each group represents a distinct prescribing context, so the results would need to be interpreted in relation to the specific comparison being made.
Primary question
The primary question was whether the use of PP-CSF differed according to recommended clinical practice guidelines between participants registered in intervention components and those in usual-care components. PP-CSF was defined as initiating granulocyte CSFs during the first cycle of myelosuppressive systemic therapy, 24 to 72 hours after systemic therapy ended. The primary endpoint was the percentage of participants prescribed PP-CSF. The endpoint values and the estimated comparison between groups are not reported in the registry record.
3. Trial Design
NCT02728596, titled S1415CD, Trial Assessing CSF Prescribing Effectiveness and Risk (TrACER), was a randomized, parallel-group trial sponsored by SWOG Cancer Research Network. The registry lists the phase as NA and masking as none. Enrollment was 3665. The recorded start date was 2016-10-07, primary completion was 2021-04-30, and status is completed.
The trial included four clinic groups with differing prescribing approaches. Clinic Group 1 had an existing automated system recommending CSF for drugs with high febrile-neutropenia risk, but not low-risk drugs. Clinic Group 2 had no automated system and followed existing clinical practice guidelines. Clinic Group 3 used an automated system recommending CSF for drugs with intermediate or high risk, but not low risk. Clinic Group 4 used an automated system recommending CSF for high-risk drugs, but not intermediate- or low-risk drugs.
The registry describes the overall comparison as PP-CSF use under guideline-based clinical practice among participants in intervention components versus usual-care components. It also says that separate mixed effects logistic models will be fit to assess the effect on PP-CSF use. The registry record does not report the randomization unit, allocation procedure, or the exact mapping of clinic groups into each analysis comparison. Those details should not be inferred solely from the four group descriptions.
The trial’s stated purpose extended beyond the primary prescribing endpoint. Secondary outcomes included febrile-neutropenia incidence and related health-related quality of life among low-risk participants; adherence to PP-CSF prescriptions; change in knowledge of PP-CSF indications; completion of initial systemic therapy at planned duration and dose intensity; antibiotic use; emergency department visits and hospitalizations related to febrile neutropenia; and health-related quality of life among intermediate-risk participants. The registry lists 14 outcome measures and 11 analyses, but the results for these outcomes are not reported in the registry record.
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 | Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) |
|---|---|---|---|---|
| STARTED | 707 | 644 | 1281 | 973 |
| Treatment Completed | 665 | 602 | 1158 | 905 |
| COMPLETED | 660 | 596 | 1151 | 898 |
| NOT COMPLETED | 47 | 48 | 130 | 75 |
Reasons for not completing this period:
| Reason | Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) |
|---|---|---|---|---|
| Refusal unrelated to adverse event | 3 | 6 | 7 | 7 |
| Death | 44 | 42 | 123 | 68 |
Analysis population for the primary outcome, as defined in the registry: There was one participant in Arm 4, that was deemed as having an unavailable outcome and was excluded from this analysis population. Leaving Arm 4 with 972 analyzable participants rather than 973.
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 | Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) | Total |
|---|---|---|---|---|---|---|
| Participants analysed | 707 | 644 | 1281 | 973 | 3605 | |
| Age, Continuous | years | 58.8 (25.8 to 94.8) | 59.9 (23.6 to 86.1) | 59.6 (18.8 to 91.5) | 58.9 (22.5 to 89.7) | 59.3 (18.8 to 94.8) |
| Sex: Female, Male: Female | Participants | 594 | 508 | 937 | 788 | 2827 |
| Sex: Female, Male: Male | Participants | 113 | 136 | 344 | 185 | 778 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 14 | 26 | 227 | 110 | 377 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 692 | 599 | 1012 | 846 | 3149 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 1 | 19 | 42 | 17 | 79 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 1 | 0 | 22 | 17 | 40 |
| Race (NIH/OMB): Asian | Participants | 10 | 15 | 45 | 24 | 94 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 | 1 | 7 | 6 | 14 |
| Race (NIH/OMB): Black or African American | Participants | 62 | 113 | 115 | 118 | 408 |
| Race (NIH/OMB): White | Participants | 622 | 500 | 1059 | 733 | 2914 |
| Race (NIH/OMB): More than one race | Participants | 6 | 1 | 4 | 8 | 19 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 6 | 14 | 29 | 67 | 116 |
| Febrile Neutropenia Risk Level: Low | Participants | 121 | 128 | 249 | 209 | 707 |
| Febrile Neutropenia Risk Level: Intermediate | Participants | 202 | 207 | 542 | 304 | 1255 |
| Febrile Neutropenia Risk Level: High | Participants | 384 | 309 | 490 | 460 | 1643 |
| Cancer Type: Breast | Participants | 508 | 397 | 731 | 639 | 2275 |
| Cancer Type: Colorectal | Participants | 136 | 181 | 368 | 245 | 930 |
| Cancer Type: NSCLC | Participants | 63 | 66 | 182 | 89 | 400 |
6. Primary Endpoint
The registry lists the primary outcome as Percentage of Participants With CSF Prescribed as Primary Prophylaxis, measured over Baseline to up to 14 days.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | percentage of participants |
| Endpoint classification | Binary Endpoint |
The primary endpoint measures the percentage of participants prescribed primary prophylactic colony stimulating factor. The registry defines PP-CSF as the initiation of granulocyte CSFs during the first cycle of myelosuppressive systemic therapy, given 24 to 72 hours after systemic therapy ends. The stated timeframe is baseline to up to 14 days.
This endpoint focuses on prescribing behavior rather than on febrile-neutropenia events themselves. Its purpose is to assess whether clinic prescribing approaches are associated with use of PP-CSF in accordance with recommended clinical practice guidelines. A prescribing percentage can describe how often the preventive medication was initiated in a defined group, but by itself it does not establish whether prescribing changed a patient’s risk of febrile neutropenia or other outcomes.
For a percentage to be interpretable, the numerator and denominator must correspond to the same outcome definition, population, and time window. The registry describes the rate as the number of participants prescribed CSF as primary prophylaxis divided by the total number of participants in each arm. It notes that one participant in Clinic Group 4 had an unavailable outcome and was excluded, leaving 972 analyzable participants rather than 973 in that group. The endpoint values and group-level comparisons are not reported in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record establishes the trial’s question, the definitions of the clinic prescribing approaches, and the planned primary endpoint. It does not report the primary endpoint percentages, the estimated differences between groups, confidence intervals, or p-values. As a result, the posted information does not establish the direction or size of any observed prescribing contrast, nor how precisely that contrast was estimated. No conclusion about the primary outcome can be drawn from the design description alone.
The primary endpoint is a binary outcome at the participant level: a participant either met the registry definition of receiving PP-CSF or did not. Group-level percentages are useful for describing prescribing frequency, while a model-based comparison can account for the structure specified by the analysis plan. The registry says separate mixed effects logistic models will be fit, but does not report the model estimates or the details needed to reproduce them. General interpretation of a logistic model depends on its outcome coding, covariates, random effects, and target contrast; those specifications are not reported in the registry record.
The four clinic groups use different approaches, including guideline-based usual care and automated systems with different febrile-neutropenia risk thresholds. That means a result for one group contrast should not automatically be treated as a result for every other contrast or as a single effect of “automation.” The registry’s primary question refers to intervention components versus usual-care components, but the precise comparisons and their estimates are not reported. The group descriptions are important context for interpreting any eventual contrast.
One participant in Clinic Group 4 was excluded from the primary analysis because the outcome was unavailable; the analysis population for that group was 972 rather than 973. This identifies one documented instance of outcome unavailability, but it does not describe missingness in the other groups or outcomes. The registry does not report how unavailable outcome data were handled beyond that exclusion, or whether sensitivity analyses were conducted.
General statistical interpretation: a confidence interval, when reported, expresses uncertainty around an estimated effect under the model and assumptions used. Its width reflects how precisely the effect is estimated, while its location indicates the range of effects compatible with the analysis. Because estimates and intervals are not reported in the registry record, there is no basis to characterize statistical precision or whether any estimate is compatible with no difference. Statistical significance, even if reported elsewhere, would not by itself describe the size or practical meaning of a prescribing difference.
The trial also lists secondary outcomes covering febrile-neutropenia events, health-related quality of life, antibiotic use, treatment completion, and health-care utilization. Those outcomes address different questions from the prescribing endpoint. The registry record does not report their results in the information. The completed status and total enrollment do not substitute for outcome estimates and cannot establish what the trial observed.
9. Statistical Methods Explained
The registry describes the primary outcome as a percentage and states that separate mixed effects logistic models will be fit. It does not report the fitted model results or a complete statistical analysis specification. The concepts below explain the methods in general terms and should not be read as confirmation that any unreported modeling choices were used.
What does a binary endpoint represent?
A binary endpoint records whether each participant meets a defined condition. For this trial, the condition was receiving PP-CSF according to the registry’s definition. The resulting group percentage summarizes the proportion of participants who meet that condition. The registry defines the denominator as the total number of participants within each arm and defines PP-CSF in relation to the first cycle of myelosuppressive systemic therapy and the interval after treatment cessation.
General explanation: logistic regression is commonly used when the outcome has two possible values. It models the probability of the outcome through a transformed scale and can estimate associations between group membership and the probability of the event. A model’s effect estimate is not necessarily identical to a simple difference between observed percentages. The exact effect measure, covariates, and reported scale for this trial are not reported in the registry record.
Why might a model be mixed effects?
General explanation: mixed effects models can represent more than one level of variation, such as participant-level outcomes occurring within clinics. A random effect can account for correlation among observations sharing a cluster, when that structure is part of the design and model. This can affect uncertainty estimates because observations within a cluster may be more alike than observations from different clusters. The registry states that mixed effects logistic models will be used, but it does not report the random-effects structure, the clustering unit, or the model specification.
The trial is described as randomized and parallel group, and its four arms are clinic groups with distinct prescribing approaches. However, the randomization unit and allocation process are not reported in the registry record. It would therefore be unwarranted to assert a particular cluster-randomized design or to state precisely how clinic-level grouping entered the model. The registry’s model description signals an intended mixed-effects analysis, but does not by itself reveal all assumptions or implementation details.
What does a comparison between groups require?
General explanation: an estimated group contrast depends on which groups are compared and how the analysis population is defined. When several arms represent different strategies, a single overall contrast may not describe every pairwise difference. The trial describes two comparator settings and two automated-system settings, with different risk thresholds for CSF recommendations. The exact primary contrast estimates and their interpretation are not reported in the registry record. Readers should distinguish the stated question—intervention components versus usual-care components—from any specific pairwise comparison unless the analysis definition is available.
In an unadjusted comparison, the group percentages are compared directly. A model-based comparison can incorporate the design and specified covariates or random effects. These approaches can yield different estimates, so it matters whether a reported result is crude or adjusted. The registry indicates mixed effects logistic models but does not state which covariates were included, how effects were summarized, or whether any adjusted and unadjusted estimates were both reported.
How should precision be assessed?
General explanation: an effect estimate describes the estimated magnitude and direction of a contrast, while its confidence interval describes uncertainty under the analysis assumptions. A narrow interval indicates greater precision than a wide interval, but neither precision nor statistical significance alone determines whether an observed difference is important. The registry record does not report primary outcome estimates, confidence intervals, or p-values, so precision cannot be assessed from the reported trial details.
Planning calculations are distinct from observed results. Any sample-size illustration based on an observed effect is a planning exercise, not an estimate of what the trial proved and not “observed power.” Such illustrations depend on assumptions about the target effect, allocation, significance level, and statistical approximation. The trial’s own sample-size justification and design assumptions are not reported in the registry record. No sample-size rationale should be inferred from enrollment alone.
What role does the outcome window play?
The registry gives the primary outcome timeframe as baseline to up to 14 days, and defines PP-CSF in relation to the first chemotherapy cycle. A prespecified outcome window determines which prescriptions count toward the endpoint. General explanation: changing a measurement window can change which events are included and therefore can alter a reported proportion. The registry’s definition should be retained when interpreting the endpoint rather than treating all CSF use during treatment as equivalent to primary prophylaxis.
The registry describes 14 outcome measures and 11 analyses. It does not report the full analysis plan, any hierarchy among outcomes, or whether particular comparisons were designated as primary or secondary beyond the primary outcome description. The planned model description, outcome definition, and study groups are available, but many implementation details and all primary outcome estimates remain unreported in the registry record.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 14 outcome measures and 11 analyses. It identifies one primary outcome, percentage of participants with CSF prescribed as primary prophylaxis, and lists secondary outcomes addressing febrile neutropenia, health-related quality of life, adherence, knowledge, treatment completion, antibiotic use, emergency visits, and hospitalizations. The registry record does not report whether the analysis plan used a hierarchy, adjusted for multiple comparisons, or designated which secondary analyses were confirmatory. With multiple outcomes or comparisons, the chance of at least one statistically notable result can increase; general methods for addressing this include prespecification and multiplicity adjustment. Whether such methods were used in this trial is not reported in the registry record.
The primary analysis population excluded one participant in Clinic Group 4 because the outcome was unavailable. The registry states that this left 972 analyzable participants in that group rather than 973. It does not report the reasons for this unavailable outcome, the amount or handling of missing data in other groups, or whether missingness was assessed for other endpoints. The record also does not state whether an intention-to-treat population, per-protocol population, or additional sensitivity analyses were used.
General explanation: missing outcome data can affect interpretation if the participants with unavailable outcomes differ systematically from those with observed outcomes. The potential impact depends on why data are missing and how the analysis handles them. A single documented exclusion does not establish the overall missing-data pattern or demonstrate that missingness could not affect an estimate. Since the registry provides no outcome estimates and limited missing-data detail, the direction or size of any possible impact cannot be assessed.
The registry says separate mixed effects logistic models will be fit, but does not report the complete analysis plan, model details, or results. The sponsor’s sample-size justification, interim-analysis plan, and statistical multiplicity procedures are not reported in the registry record.
11. 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 |
|---|---|---|---|
| Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | 3 | 456 | 0.7% |
| Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | 0 | 350 | 0.0% |
| Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | 0 | 694 | 0.0% |
| Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) | 1 | 448 | 0.2% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | 44 of 707 |
| Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | 42 of 644 |
| Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | 123 of 1281 |
| Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) | 68 of 973 |
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 |
|---|---|
| Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | 146 of 456 |
| Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | 123 of 350 |
| Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | 243 of 694 |
| Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) | 171 of 448 |
Most frequent serious adverse event terms
| Event term | Organ system | Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) |
|---|---|---|---|---|---|
| Sepsis | Infections and infestations | 1 of 456 | 0 of 350 | 0 of 694 | 1 of 448 |
| Neoplasms benign, malignant and unspecified - Other | Neoplasms benign, malignant and unspecified (incl cysts and polyps) | 1 of 456 | 0 of 350 | 0 of 694 | 0 of 448 |
| Respiratory failure | Respiratory, thoracic and mediastinal disorders | 1 of 456 | 0 of 350 | 0 of 694 | 0 of 448 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Active Comparator: Clinic Group 1 (Clinics With Existing Automated System for CSF Prescribing) | Active Comparator: Clinic Group 2 (Clinics With no Automated System for CSF Prescribing) | Experimental: Clinic Group 3 (Clinics With Automated System for CSF Prescribing) | Experimental: Clinic Group 4 (Clinics With Automated System for CSF Prescribing) |
|---|---|---|---|---|---|
| Bone pain | Musculoskeletal and connective tissue disorders | 110 of 456 | 101 of 350 | 205 of 694 | 121 of 448 |
| Arthralgia | Musculoskeletal and connective tissue disorders | 51 of 456 | 44 of 350 | 81 of 694 | 76 of 448 |
| Mucositis oral | Gastrointestinal disorders | 10 of 456 | 13 of 350 | 44 of 694 | 27 of 448 |
| Myositis | Musculoskeletal and connective tissue disorders | 6 of 456 | 2 of 350 | 24 of 694 | 23 of 448 |
Serious events with the highest arm-level rate: sepsis (0.2%), neoplasms benign, malignant and unspecified - other (0.2%), respiratory failure (0.2%).
The registry lists serious adverse events and other adverse events by clinic group, along with deaths. For Clinic Group 1, 3 serious adverse events are listed among 456 participants at risk, 146 other adverse events among 456, and 44 deaths. For Clinic Group 2, the corresponding counts are 0 among 350, 123 among 350, and 42. For Clinic Group 3, they are 0 among 694, 243 among 694, and 123. For Clinic Group 4, they are 1 among 448, 171 among 448, and 68.
These are registry-listed counts, not evidence that an event was caused by a particular prescribing approach. The record does not report the event assessment method, follow-up duration for these safety summaries, or whether participants could appear in more than one adverse-event category. Death counts should not be treated as equivalent to treatment-related deaths; attribution is not reported. The most frequent serious-event terms listed are sepsis, neoplasms benign, malignant and unspecified - Other, and respiratory failure, each at 0.2%.
General explanation: safety counts require context, including the observation period, event definitions, severity criteria, and whether events are participant counts or repeated events. Comparisons between groups should not be made from raw counts alone when group sizes or follow-up differ. The registry lists participants at risk and affected counts, but does not report enough detail to determine causal relationships or make a comparative safety interpretation.
12. What the Result Does, and Does Not, Mean
The trial’s primary question concerned whether PP-CSF prescribing differed between intervention components and usual-care components. The registry gives a clinically interpretable definition of PP-CSF and describes four clinic prescribing approaches. It does not report the primary endpoint values or statistical comparisons. Therefore the available record does not establish whether prescribing was more or less frequent in any particular group, or by how much.
The endpoint concerns prescribing, not direct evidence of prevention of febrile neutropenia. Even a reported difference in the percentage receiving PP-CSF would describe medication use under the defined prescribing strategies; it would not alone determine effects on febrile-neutropenia incidence, quality of life, antibiotic use, or treatment completion. Those outcomes are separately listed as secondary measures, and their results are not reported in the registry record.
General explanation: a randomized design can support a causal comparison when the allocation, analysis, outcome ascertainment, and handling of missing data are appropriate to the study question. The registry identifies the design as randomized and parallel group, but does not report the randomization unit or detailed allocation procedures. The result’s interpretation would also depend on which groups were compared and the statistical model used. The available description is not enough to infer those results.
The trial status is completed and enrollment is 3665, but completion and enrollment do not communicate the direction, magnitude, or precision of an outcome. The registry also records one unavailable primary outcome in Clinic Group 4, excluded from the analysis population. This limited missing-data note is not sufficient to determine whether missing data affected the findings.
Accordingly, the appropriate interpretation is bounded: the trial was designed to examine clinic prescribing approaches, and the registry defines the primary measure and analysis approach. The observed primary prescribing result, its uncertainty, and the results for the secondary clinical outcomes are not reported in the registry record. No claim about benefit, lack of benefit, or comparative clinical effect follows from the design information alone.
13. Why This Trial Matters Statistically
This trial addresses a statistical distinction between changing a prescribing process and changing a clinical outcome. The primary endpoint measures whether participants received PP-CSF, while the listed secondary outcomes include febrile-neutropenia incidence, related quality of life, antibiotic use, and health-care utilization. Keeping those endpoint types separate is essential: a prescribing percentage describes implementation, whereas clinical event outcomes address what happened to participants.
The four clinic groups also illustrate why intervention labels need careful interpretation. The automated systems did not all use the same threshold: the registry describes approaches covering high-risk drugs only or intermediate- and high-risk drugs. One comparator group had an existing automated system, while another used clinical practice guidelines without an automated system. These differences mean that an analysis should make the specific contrast clear rather than treating all automated prescribing as one uniform strategy.
The registry states that mixed effects logistic models were planned. General statistical reasoning supports attention to clustering when outcomes may be related within clinics, but the randomization unit and model structure are not reported. This makes the model description important yet incomplete for evaluating precision and reproducibility.
The case also shows why outcome estimates, uncertainty intervals, and missing-data descriptions are necessary for interpreting a completed trial. Enrollment and design details alone do not reveal the observed prescribing pattern. The record documents one unavailable primary outcome in Clinic Group 4, while broader missing-data and multiplicity procedures are not reported. Because the primary result and its precision are absent, the registry description supports understanding the question and design, but not judging the magnitude of the observed statistical contrast.
14. 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’s group values, between-group estimates, confidence intervals, and p-values are not reported in the registry record. This prevents assessment of the direction, magnitude, and precision of the primary prescribing comparison. Results for the listed secondary outcomes are also not reported in the information, so the broader clinical and health-care outcomes cannot be interpreted.
Important design and analysis details are absent. The registry does not report the randomization unit or allocation procedure, the exact group contrasts used in the primary analysis, the full mixed effects logistic model specification, or the covariates and random-effects structure. Masking is listed as none. The trial’s sample-size justification, interim-analysis plan, and multiplicity procedures are not reported in the registry record.
The primary analysis population note identifies one unavailable outcome in Clinic Group 4, leaving 972 analyzable participants rather than 973. Reasons for that missing outcome and the handling of missingness elsewhere are not reported. It is not possible to determine from this note alone whether missing outcome data could have influenced a comparison.
The safety information is presented as counts by group, but the record does not report the safety observation period, event attribution, or details needed to determine whether categories overlap. The listed adverse events and deaths should therefore be read descriptively, not as evidence of a causal effect of a particular clinic approach.
Finally, the four prescribing approaches differ in more than the presence of automation: their risk thresholds and usual-care context differ. Without the exact reported contrasts and results, it is not possible to isolate the effect of automation from the specific prescribing rules or clinical context.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: S1415CD, Trial Assessing CSF Prescribing Effectiveness and Risk (TrACER) — NCT02728596.
- Registry record last updated 2022-10-20; status last verified 2022-09.
- Results first posted 2022-10-20.
- 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.
18. Record Summary
NCT02728596, S1415CD TrACER, was a completed randomized, parallel-group trial sponsored by SWOG Cancer Research Network. Enrollment was 3665; the recorded start date was 2016-10-07 and primary completion was 2021-04-30. The registry lists masking as none and phase as NA.
The trial compared four clinic prescribing approaches for primary prophylactic colony stimulating factor among participants receiving chemotherapy for breast, colorectal, or non-small cell lung cancer. The primary endpoint was the percentage prescribed PP-CSF, defined in relation to the first cycle of myelosuppressive systemic therapy and the interval after treatment ended. The registry states that separate mixed effects logistic models will be fit.
The primary endpoint results, estimates of between-group differences, uncertainty intervals, p-values, and results for secondary outcomes are not reported in the registry record. One unavailable primary outcome was excluded in Clinic Group 4, leaving 972 analyzable participants rather than 973.