A Study of Management of Tarceva - Induced Rash in Patients With Non-Small Cell Lung Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Management of Tarceva - Induced Rash in Patients With Non-Small Cell Lung Cancer. (NCT00531934), a randomized phase 2 study. The primary outcome is Percentage of Participants With at Least One Skin Rash (Folliculitis) of Any Grade During the First 4 Months of Treatment.

2. Clinical Question

The question concerns prevention of a treatment-associated skin outcome, rather than treatment of the cancer itself. The primary endpoint includes folliculitis and specified skin findings, and was assessed at scheduled visits through the first 4 months. The trial also recorded rash outcomes after that period and measures related to the time until rash. Those additional measures provide context about timing and persistence, but they do not replace the designated primary comparison.

The reported primary percentages were 71.2% for erlotinib plus doxycycline and 80.8% for erlotinib alone. The unadjusted comparison does not establish a clear difference: the reported p-value is 0.175, and the recalculated confidence interval for the difference includes no difference. These results describe the observed groups and the uncertainty around their comparison; they do not establish that the treatments are equivalent.

3. Trial Design

The registry describes a randomized, open-label, parallel-group phase 2 study sponsored by Hoffmann-La Roche. Enrollment was 147, and the study status is completed. The study began in 2007-10, with primary completion in 2010-08. Participants were assigned to erlotinib plus doxycycline or erlotinib without preventative treatment. The arm descriptions are not reported in the registry record, but the brief summary describes the assigned preventative-treatment approaches.

The summary states that doxycycline was given at 100mg po daily and that all participants received Tarceva at 150mg/kg po daily. It describes anticipated treatment continuing until disease progression or intolerable toxicity. The registry identifies erlotinib [Tarceva] and doxycycline as drug interventions. As written, the erlotinib dose in the summary is 150mg/kg po daily; no alternative dose or clarification is reported in the registry record.

No masking was used, so the study was open label. The record identifies random allocation and a parallel model. Details about the randomization procedure, allocation concealment, stratification, protocol deviations, or adherence are not reported in the registry record. The planned sample-size rationale and original design assumptions are also not reported in the registry record. The separate planning illustrations are based on the observed effect and should not be read as the trial's original enrollment target or design calculation.

The primary endpoint assessed the percentage of participants with at least one rash of any grade during the first 4 months. Scheduled assessments were on Days 0, 14, and 28 and Months 2, 3, and 4. The registry identifies the primary population as intention-to-treat and notes that primary data for one participant in the erlotinib group were missing. Enrolled participants, outcome-specific denominators, and safety denominators do not always match, so each result should be interpreted using its stated analysis population and denominator.

The record lists 32 outcome measures and 5 analyses. It does not report whether interim analyses were planned or conducted, whether the primary analysis was prespecified in a statistical analysis plan, or how secondary analyses were prioritized. Those details cannot be inferred from the completed status or the results posted in the registry.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population; data for 1 participant in the erlotinib treatment group were missing.

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.

Population: Intent-to-Treat population (ITT): all randomized participants who received at least 1 dose of erlotinib.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With at Least One Skin Rash (Folliculitis) of Any Grade During the First 4 Months of Treatment, measured over Days 0, 14, 28 and Months 2, 3, and 4.

The primary endpoint is the percentage of participants with at least one skin rash (folliculitis) of any grade during the first 4 months of treatment. The registry specifies assessments on Days 0, 14, and 28 and Months 2, 3, and 4. Rash descriptions included erythema, papulo-pustules, nodule, and crust, assessed according to the Common Terminology Criteria for Adverse Events (CTCAE) version 3 scale. Medical photographs of the face and regions with skin lesions were obtained and reviewed by a centralized evaluation committee.

A binary percentage endpoint records whether each participant experienced at least one qualifying event in the stated interval. It gives a direct summary of how many participants had the outcome, but does not by itself describe the number of rash episodes, their duration, or their severity beyond the endpoint's any-grade definition. Separate secondary outcomes address some other aspects, including rash type, maximal intensity, later events, and time-related outcomes.

The endpoint is used to compare the frequency of experiencing any qualifying rash between the randomized groups over a defined treatment period. In this trial, the posted percentages were 71.2% with erlotinib plus doxycycline and 80.8% with erlotinib. The intention-to-treat description is qualified by the registry's note that data for one participant in the erlotinib group were missing. The precise handling of that missing primary outcome is not reported in the registry record.

7. Results Posted in the Registry

• Erlotinib: 80.8% (59/73) versus 71.2% (52/73) on control; risk difference 9.6 percentage points (95% CI -4.2 to 23.4), risk ratio 1.13 (0.94–1.36); unadjusted two-proportion z-test p = 0.175.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Percentage of Participants With at Least One Skin Rash (Folliculitis) of Any Grade After the First 4 Months of Treatment

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Erlotinib: 1.4% (1/74) versus 2.7% (2/73) on control; risk difference -1.4 percentage points (95% CI -6.0 to 3.2), risk ratio 0.49 (0.05–5.32); unadjusted two-proportion z-test p = 0.552.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erlotinib and the control. The interval gives the range of effects that remain compatible with the data.

Time Free From Skin Rash (Folliculitis) During the First 4 Months of Treatment - Number of Participants With an Event

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Erlotinib: 79.7% (59/74) versus 71.2% (52/73) on control; risk difference 8.5 percentage points (95% CI -5.3 to 22.3), risk ratio 1.12 (0.93–1.35); unadjusted two-proportion z-test p = 0.231.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erlotinib and the control. The interval gives the range of effects that remain compatible with the data.

Time Free From Skin Rash (Folliculitis) During the First 4 Months of Treatment - Time to Event

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Erlotinib + Doxycycline vs Erlotinib: hazard ratio 0.76 (95% CI 0.53–1.11), reported p = 0.143 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Erlotinib + Doxycycline vs Erlotinib and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the primary question by looking at rash after the first 4 months, event counts, rash type and intensity, and time-related outcomes. Together, they address whether the primary any-rash comparison is accompanied by differences in when rash appears, how it is characterized, or whether it occurs later in treatment. The posted secondary results do not establish a clear, consistent separation between the treatment groups across these different measures.

The time-to-event analysis is a distinct type of evidence because it uses event timing, whereas the primary endpoint is a percentage over a fixed period. Later-period outcomes also concern a different follow-up interval and may have different denominators. These outcomes can help describe the overall pattern, but they should not be treated as repeated tests of exactly the same hypothesis. Each addresses a related but separate outcome.

Secondary results generally carry less confirmatory weight when many outcomes are examined and no multiplicity procedure is reported. When multiple comparisons are made, the chance of finding an apparently unusual result by chance can increase. A prespecified multiplicity strategy can address that issue, but one is not reported in the registry record. The record also does not report a hierarchy identifying which secondary outcomes were intended to support confirmatory conclusions.

Accordingly, the secondary outcomes are best regarded as contextual and exploratory evidence alongside the primary result. Their varied definitions and follow-up periods make a single pooled interpretation inappropriate. Their role is to describe additional aspects of rash in this trial, not to overturn or independently confirm the primary comparison.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Skin Rash (Folliculitis) Events During the First 4 Months of Treatment
• Percentage of Participants With Skin Rash (Folliculitis) During the First 4 Months of Treatment By Type
• Percentage of Participants With Skin Rash (Folliculitis) During the First 4 Months of Treatment By Maximal Intensity
• Percentage of Participants With at Least One Skin Rash (Folliculitis) of Any Grade After the First 4 Months of Treatment
• Number of Skin Rash (Folliculitis) Events After the First 4 Months of Treatment
• Number of Participants With Skin Rash (Folliculitis) After the First 4 Months of Treatment By Type
• Number of Participants With Skin Rash (Folliculitis) After the First 4 Months of Treatment By Intensity
• Time Free From Skin Rash (Folliculitis) During the First 4 Months of Treatment - Number of Participants With an Event

9. Statistical Interpretation

10. Statistical Methods Explained

What does a percentage endpoint represent?

A percentage endpoint treats each participant as having either experienced the defined outcome during the specified period or not. The primary endpoint counted participants with at least one rash of any grade during the first 4 months. This makes the result relatively straightforward to interpret as a proportion of participants, but it does not capture repeated episodes as separate events. Nor does the any-grade endpoint alone describe how long rash lasted or how severe it became. The registry lists other outcomes addressing event counts, types, intensity, and timing.

For a binary outcome, a treatment comparison can be expressed as a risk difference, a risk ratio, or an odds ratio. These measures answer related but distinct questions. A risk difference compares the proportions by subtraction; a risk ratio compares them as a ratio; and an odds ratio compares the odds of the outcome. The recalculation uses the posted summary counts and is unadjusted. It does not incorporate participant-level covariates or any modeling choices that are not available in the registry record.

How should the chi-squared result be read?

The reported primary analysis used a chi-squared test. In general, this test assesses whether the observed distribution of a categorical outcome across groups is compatible with a null hypothesis of no association between group and outcome. Its p-value summarizes how unusual the data would be under that null model, subject to the test's assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference.

The registry reports p = 0.175 for the primary comparison. The unadjusted recalculation based on the posted counts also gives p = 0.175. This agreement indicates that the simple calculation reproduces the posted p-value at the precision reported. It does not show whether the original analysis used additional procedures, because any such details are not reported in the registry record.

What does a confidence interval add?

A confidence interval expresses uncertainty around an estimate under the assumptions of the calculation. For the primary risk difference, the recalculated interval includes zero, the value corresponding to no difference in proportions. This means the data are compatible with a range that includes no difference as well as values on either side. The interval is more informative than the point estimate alone because it makes the uncertainty visible. It does not assign a probability to a fixed treatment effect or guarantee that the true value lies within its endpoints.

For the time-to-event outcome, the registry reports a hazard ratio with a 95% confidence interval. A hazard ratio compares the event rates over time under the analysis model; it is not the same as comparing the final percentages. General interpretation depends on the event definition, comparison direction, and assumptions of the time-to-event method. The registry reports a log-rank test for this comparison. The hazard-ratio interval includes 1, the null value for a ratio measure, and the reported p-value is 0.143.

Why does the time-to-event result differ from a simple percentage?

Time-to-event methods use information about when an event occurred, rather than only whether it occurred by the end of a period. They can also accommodate participants whose event time is not observed through the full follow-up, under assumptions about censoring. The registry reports median time free from rash and a log-rank comparison. Details about censoring rules, proportional-hazards assessment, and other model checks are not reported in the registry record, so those assumptions cannot be evaluated from the posted information.

The median is the time by which half of the relevant event-time distribution has experienced the event, as estimated by the analysis. The reported medians are 14 days for erlotinib plus doxycycline and 13 days for erlotinib. Their reported confidence intervals overlap, and the hazard-ratio confidence interval includes its null value. Those features describe uncertainty; they do not establish that the time distributions are identical.

What is the role of the recalculation?

A recalculated estimate based on group-level counts is a useful check on the arithmetic and gives readers a familiar unadjusted contrast. It is not a replacement for a prespecified analysis using the trial's complete participant-level data. For the primary outcome, the registry's reported chi-squared p-value and the recalculated p-value agree. No discrepancy is flagged. When the registry supplies an analysis, that remains the record of the sponsor's reported method and result; the recalculation provides an additional unadjusted perspective, not an alternative claim about the trial's analysis plan.

Different analyses can yield different estimates or uncertainty when they use different populations, adjustments, outcome definitions, or missing-data assumptions. The primary record identifies the intention-to-treat population but notes missing data for one participant in the erlotinib group. The precise procedure for handling that missing value is not reported in the registry record. Therefore, the recalculation should be understood in the context of the posted summary data, not as a reconstruction of every analytic decision.

11. Confidence Intervals and Precision

For the primary outcome, the posted percentages are 71.2% with erlotinib plus doxycycline and 80.8% with erlotinib. The recalculated unadjusted confidence interval for their risk difference includes zero. This is an important precision point: the estimate has uncertainty broad enough to include no between-group difference, as well as differences in either direction. The p-value of 0.175 is consistent with that reading. The point estimates alone should not be treated as a precise estimate of the underlying contrast.

A confidence interval is a range produced by a statistical procedure that, under its assumptions, reflects the uncertainty associated with an estimate. Wider intervals indicate less precision than narrower intervals. The interval does not mean that every value inside it is equally plausible, and it is not a guarantee about the true effect. It also does not resolve whether an observed difference would be important to patients or clinicians; that requires a relevant clinical context and thresholds that are not reported in the registry record.

The time-to-event comparison reports a hazard ratio of 0.763, with a 95% confidence interval from 0.525 to 1.109. The interval includes 1, the no-difference value for a hazard ratio. The log-rank p-value is 0.143. The median estimates are 14 days and 13 days, with respective confidence intervals of 13 to 22 days and 8 to 16 days. The uncertainty around these estimates cautions against interpreting their point-estimate separation as conclusive.

For the later-period rash outcome, only a small number of events were recorded, and the recalculated interval for the difference includes zero. Sparse outcomes can produce imprecise estimates, particularly when the event count is low. General statistical interpretation should therefore consider both the estimate and its interval, rather than treating a small observed percentage as proof of a precisely known rate.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning illustrations are not the trial's original design calculation. The registry's original sample-size justification and assumptions are not reported in the registry record. These illustrations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They use a stated fraction of the observed effect as the effect to detect. The time-to-event illustrations are expressed as total events; the other illustrations are expressed as participants per arm.

The 80% power column represents the planning sample size associated with an 80% chance of detecting the specified effect under the stated assumptions, if that effect were the true effect and the test assumptions held. The 90% power column uses a higher target chance of detecting that effect, and therefore requires a larger planning sample in the displayed illustrations. Power is a design property conditional on assumptions, not a measure of how likely the observed result is to be true.

The rows that plan for 100%, 75%, and 50% of the observed effect illustrate how the required sample changes when the target effect is smaller. Smaller effects need more participants to distinguish them from random variation at the same significance level and target power. For example, for the primary endpoint, planning for 100% of the observed effect gives 310 participants per arm for 80% power and 415 for 90% power; planning for 50% gives 1,325 participants per arm for 80% power and 1,773 for 90% power. These are planning illustrations, not observed power. They should not be interpreted as proof that the completed trial was adequately or inadequately designed, because its original assumptions are not reported.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.712; treatment proportion p₁ = 0.808; average p̄ = 0.760.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6038 = 1.1833.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5999 = 0.5049.
4. n = (1.1833 + 0.5049)² / (0.0959)² = 310.0, rounded up to 310 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.712328767123288                          # control proportion
p1 <- 0.808219178082192                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(531934)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 310 (80% power), 415 (90% power)
 75% of observed effect: n per arm = 571 (80% power), 764 (90% power)
 50% of observed effect: n per arm = 1,325 (80% power), 1,773 (90% power)
power.prop.test: 310 
Simulated power at n = 310 per arm: 0.805 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry reports a chi-squared analysis of the primary comparison with p = 0.175. The unadjusted recalculation from the posted group counts also gives p = 0.175. There is no discrepancy between the reported p-value and this simple recalculation at the displayed precision.

The recalculation supplies a transparent comparison based on the posted summary statistics, but it does not reconstruct participant-level analysis or undisclosed modeling choices. It is unadjusted, and the registry does not report additional details about the primary statistical plan, missing-data procedure, or any covariate adjustment. The registry's chi-squared analysis is therefore the sponsor's recorded analysis; the recalculation is a check and a complementary description rather than a replacement.

The posted arm percentages are 71.2% and 80.8%. The confidence interval for the recalculated difference includes no difference, and the p-value does not provide evidence of a clear difference under the reported test. These statistical features support cautious interpretation of the observed contrast. They do not establish equivalence, and they do not determine whether any difference would be clinically meaningful.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, time points, or comparisons. Even if each test is considered separately, the chance of at least one apparently notable result can increase as more tests are examined. Statistical approaches can control or otherwise address this issue, for example through an adjustment or a prespecified testing hierarchy. Whether such methods were used in this study is not reported in the registry record.

The registry lists 32 outcome measures and 5 analyses. Several outcomes concern related rash features over different periods, while the primary endpoint is the percentage of participants with at least one rash during the first 4 months. Without a reported multiplicity plan, secondary p-values or estimates should not be interpreted as if they were the sole prespecified test. The status of each secondary analysis as confirmatory or exploratory is also not reported in the registry record.

The primary endpoint is described for the intention-to-treat population, with a note that data for one participant in the erlotinib group were missing. The registry does not report the exact missing-data method, whether the participant was excluded from the primary calculation, or whether any imputation was used. The posted outcome-specific denominators should therefore be considered alongside the population description rather than assuming that every endpoint used an identical set of participants.

Details about protocol deviations, adherence, interim analyses, and the statistical analysis plan are not reported in the registry record. It is also not reported whether analyses were adjusted for baseline characteristics. These omissions limit the ability to assess analytic flexibility and the robustness of secondary findings. They do not invalidate the posted numbers, but they constrain how strongly conclusions can be drawn from the collection of results.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: general physical health deterioration (8.1%), sepsis (5.4%), pulmonary embolism (4.1%), abdominal pain (2.7%), diarrhoea (2.7%).

The registry lists serious adverse events and other adverse events separately for the treatment groups. Serious adverse events affected 18 of 73 participants in the erlotinib plus doxycycline group and 24 of 74 in the erlotinib group. Other adverse events were recorded for 71 of 73 and 69 of 74 participants, respectively. These are descriptive counts with their stated denominators; they are not a randomized comparison adjusted for follow-up or other factors.

The registry lists serious-event terms including general physical health deterioration, sepsis, pulmonary embolism, abdominal pain, diarrhoea, and myocardial infarction, with percentages reported for those terms. The record does not establish from these listings that an event was caused by a particular treatment. Death counts are not reported in the registry record. Definitions, exposure-adjusted rates, timing, and detailed event-level comparisons are not reported in the registry record, so the figures should be read as the posted safety summaries rather than as a complete causal assessment.

17. What the Result Does, and Does Not, Mean

The primary result describes the frequency of recorded rash during the first 4 months among participants in this randomized study. The observed percentages were lower in the erlotinib plus doxycycline group than in the erlotinib group, but the uncertainty around the unadjusted comparison includes no difference. The reported p-value is 0.175. The data therefore do not demonstrate a clear difference for the primary endpoint, while also not establishing that the approaches are equivalent.

The result applies to the study population and outcome definition recorded in the registry: people with non-small cell lung cancer after first-line chemotherapy failure, and at least one rash of any grade during the specified assessment period. It does not by itself answer questions about rash severity, duration, quality of life, cancer response, or outcomes beyond those posted. Related secondary measures provide additional descriptions, but the registry does not report a multiplicity plan that would support treating them as a set of confirmatory tests.

Statistical uncertainty should not be confused with evidence that there is no effect. The confidence interval includes the null, but it also allows values on either side of that null. Conversely, the difference between the observed percentages should not be taken as proof of a treatment effect. A confidence interval and p-value describe evidence under statistical assumptions; they do not determine clinical importance or establish causality on their own.

The open-label design and missing primary-outcome data for one participant in the erlotinib group are relevant context. The record does not report how missing data were handled or provide a full protocol and statistical analysis plan. Safety listings are descriptive and do not establish whether reported events were caused by treatment. No medical recommendation follows from these registry results.

18. Why This Trial Matters Statistically

This trial offers a useful statistical example of why the observed difference between group percentages should be considered together with uncertainty. The primary percentages point in one direction, but the confidence interval includes no difference and the reported p-value is 0.175. Looking only at the point estimates could overstate what the comparison establishes; looking only at the p-value could obscure the range of effects compatible with the data.

It also illustrates how distinct outcome definitions answer distinct questions. The primary endpoint asks whether participants experienced any qualifying rash during a fixed period. Event counts and type or intensity outcomes describe other dimensions, while the time-to-event analysis uses information about when rash occurred. Agreement or disagreement across such measures must be interpreted in light of their definitions, follow-up windows, and uncertainty, not as though all were interchangeable versions of one endpoint.

The registry makes the distinction between sponsor-reported analysis and a simple unadjusted recalculation visible: in this case both give p = 0.175. Such agreement can reassure readers that the posted group summaries are consistent with the reported test, while still leaving important details unknown, including missing-data handling and multiplicity procedures.

Finally, the planning illustrations show how sample-size requirements depend on the effect chosen for detection and the desired power. Because the trial's original design assumptions are not reported, those illustrations should remain separate from claims about the study's original planning or adequacy. This distinction is important for interpreting registry results without attributing intentions or methods that the record does not document.

19. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

Several limitations are specific to the information recorded for this trial. It was open label, with no masking. The registry does not report the randomization procedure, allocation concealment, adherence, protocol deviations, or a full statistical analysis plan. As a result, readers cannot evaluate those design and conduct details from the registry record.

The primary outcome is a broad any-grade indicator of rash during the first 4 months. It captures whether a participant had at least one qualifying event, but it does not by itself represent event burden, duration, or severity. Other posted outcomes address some of these dimensions, but the record does not report a multiplicity plan or a hierarchy for their interpretation.

The registry notes missing primary data for one participant in the erlotinib group, but the handling of that missing outcome is not reported. The primary population is identified as intention-to-treat, while outcome-specific denominators vary across the posted results. This makes the exact analytic population for each result important, and the record does not supply enough detail to reconstruct all analysis decisions.

The original sample-size justification and interim-analysis plans are not reported. The planning illustrations are based on the observed effect and cannot establish what the trial was designed to detect. In addition, the serious adverse-event information is descriptive; the record does not provide detailed causal assessments or exposure-adjusted comparisons. These limitations narrow the conclusions that can be drawn from the registry record, without changing the reported results themselves.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Management of Tarceva - Induced Rash in Patients With Non-Small Cell Lung Cancer. — NCT00531934.
• Registry record last updated 2015-02-25; status last verified 2015-02.
• Results first posted 2014-08-26.
• 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.

23. Record Summary

NCT00531934 is a completed, randomized, open-label, parallel-group phase 2 study sponsored by Hoffmann-La Roche. It enrolled 147 participants with non-small cell lung cancer after first-line chemotherapy failure. The comparison was erlotinib plus doxycycline versus erlotinib without preventative treatment.

The primary outcome was the percentage with at least one skin rash of any grade during the first 4 months. The posted percentages were 71.2% and 80.8%, respectively. The registry reports a chi-squared p-value of 0.175; the unadjusted recalculation agrees, and its confidence interval includes no difference. The record also includes secondary rash and time-to-event outcomes and descriptive adverse-event summaries.

Missing-data handling, multiplicity procedures, the original sample-size rationale, and several details of the statistical plan are not reported in the registry record. The results describe the posted outcomes and their uncertainty; they do not establish equivalence or a causal safety conclusion.