Colorectal Cancer: Screening vs. Non-Screening: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Colorectal Cancer: Screening vs. Non-Screening (NCT02727894), a registered phase not stated study. The primary outcome is Colorectal Cancer Stage (pTNM).

2. Clinical Question

The clinical question is about stage at the time cancer was diagnosed, not about survival, recurrence, or response to treatment. The study description frames earlier-stage findings as a possible indicator of screening-programme efficiency. The primary endpoint is colorectal cancer stage (pTNM), recorded as participant counts across stage categories. The record defines stage 0 as Tis N0 M0; stage I as T1-2 N0 M0; stage II as T3-4 N0 M0; stage III as T1-4 N1-2 M0; and stage IV as T1-4 N1-2 M1.

This comparison describes how patients entered diagnosis and what stage their cancer was at that point. It does not compare randomly assigned screening with no screening. Patients in the screening group were without symptoms when invited for examination, whereas patients in the non-screening group were symptomatic. Those group definitions are central to interpreting any difference: factors related to symptoms, access to screening, or the opportunity for screening to find a cancer could also be related to stage. The registry record does not report the stage distribution or an estimated difference between groups, so the clinical question cannot be answered from its posted results.

3. Trial Design

The registry record describes a completed study sponsored by the Institute for Clinical and Experimental Medicine. Its listed enrollment is 276. The recorded start is 2013-03, and the primary completion is 2015-12. The study title contrasts colorectal cancer screening with non-screening, while the brief summary describes comparing cancer characteristics at diagnosis in screening-detected and symptom-detected cases.

The comparison is nonrandomized. The record does not state how patients were allocated to groups; the groups instead reflect whether cancer was diagnosed by screening or after symptoms. The screening group includes patients diagnosed by primary screening colonoscopy or colonoscopy after a positive immunochemical-based faecal occult blood test, when patients without symptoms had been invited according to the national screening programme policy. The non-screening group includes patients with symptomatic colorectal cancer. The registry record explicitly cautions that the arms are not a like-for-like control comparison.

The primary outcome is stage at diagnosis. The listed timeframe is “at time of diagnosis,” so the endpoint concerns disease classification at that point rather than a later clinical event. Secondary outcomes concern colorectal cancer grade, resection margins, metastasis, surgery, and palliative therapy. The record lists six outcome measures in total, but it does not provide outcome counts or findings for these measures.

No randomization method, masking approach, recruitment procedures, eligibility criteria beyond the group descriptions, or protocol details are reported in the registry record. A detailed statistical analysis plan, including the planned handling of confounding and the method for comparing stage categories, is also not reported. The listed status is completed, but completion alone does not establish that every planned analysis or outcome is available in the registry. The record lists no analyses. Accordingly, the study design and endpoint can be described, but its comparative result cannot be reconstructed from the registered information.

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.

Before assignment: Patients with diagnosed colorectal carcinoma were enrolled (276). 11 patients were excluded because of the inability to categorise those patients into the screening or non-screening group.

Overall Study

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.

6. Primary Endpoint

The registry lists the primary outcome as Colorectal Cancer Stage (pTNM), measured over at time of diagnosis.

The primary endpoint is colorectal cancer stage (pTNM) among participants at the time of diagnosis. It is expressed as counts of participants in stage categories. The registry record defines stages 0 through IV using tumour, node, and metastasis classifications. It identifies stages 0, I, and II as having a better outcome, in contrast with stages III and IV.

Stage at diagnosis is used as a clinical-pathological measure for comparing the screening and non-screening groups. The study description asks whether cancers detected through screening have more favourable characteristics at diagnosis than cancers diagnosed after symptoms. Stage provides a structured way to describe the extent of disease at that point and to examine whether the stage distribution differs across the groups. It is not itself a measure of survival or a direct estimate of the effect of screening on later health outcomes.

The endpoint is categorical: the stages distinguish groups rather than representing a continuous measurement with equal intervals between values. Treating the categories as an ordered scale, combining categories, or testing each category separately would require an analysis choice. The registry record does not state which comparison method was planned, whether stage would be analysed as a full distribution or collapsed into more favourable and less favourable categories, or how missing stage information would be handled.

The stage counts and group-specific results are not reported in the registry record. Consequently, it is not possible to describe which stage was most common, how the distributions compared, or how precisely any difference was estimated. The endpoint definition establishes what was intended to be measured; it does not by itself establish the result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a set of participant counts across colorectal cancer stage categories at diagnosis. The record defines stage using pTNM descriptions and identifies stages 0, I, and II as having a better outcome. The registry record does not state the statistical test or model used to compare the groups, and it lists no analyses. The following concepts explain general approaches that could be relevant; they do not describe an analysis confirmed as having been conducted in this study.

How can stage categories be compared?

Because stage is categorical, an analysis can compare the distribution of participants across categories between groups. A test of the full distribution asks whether the pattern of counts differs, without necessarily reducing stage to a simple favourable-versus-unfavourable contrast. Alternatively, an analysis could combine categories, but that changes the question and should be justified in the analysis plan. The registry record identifies stages 0, I, and II as having a better outcome but does not say whether those stages were combined for the primary comparison.

A count is interpretable alongside its denominator: the number of participants contributing information to that group and outcome. If denominators are absent, a count alone cannot show the proportion of a group in a given category. The registry record gives overall enrollment but not group sizes, stage counts, or the number with evaluable stage data. No proportions should therefore be inferred. The counts and denominators are also needed to assess whether expected counts are adequate for a particular categorical test.

What does an effect estimate add?

A test statistic or p-value addresses compatibility between observed data and a specified statistical model or null hypothesis; it does not measure the size or clinical importance of a difference. An effect estimate describes the contrast in a defined scale, such as a difference in proportions or an odds ratio, depending on the chosen analysis. A confidence interval expresses the range of estimates reasonably compatible with the data and model under the method’s assumptions. A narrower interval indicates greater precision, while a wider interval indicates more uncertainty. These are general interpretations; the registry record reports no effect estimate, confidence interval, or p-value for the primary outcome.

For a multi-category endpoint, the effect measure should match the question. An overall comparison of distributions may not reveal which categories account for a difference. Category-specific comparisons can be informative, but they introduce multiple opportunities to find apparently notable results. Combining categories can simplify interpretation but may obscure variation within the combined groups. The registry record does not report the planned contrast, whether category-specific tests were intended, or how any resulting multiplicity would be addressed.

Why does the study design matter?

Random allocation can, under appropriate conditions, balance measured and unmeasured baseline characteristics across groups. This record describes groups defined by screening-based or symptom-based diagnosis and notes that they are not a like-for-like control comparison. Consequently, a group difference cannot automatically be attributed to the route of diagnosis. General approaches to observational comparisons can include adjustment for prespecified characteristics, stratification, or sensitivity analyses, but the registry record does not report whether any such methods were planned or used. The necessary participant-level covariates and model results are not reported either.

Screening and symptom-based diagnosis can involve different pathways to detecting disease. As general considerations, lead-time bias can make the time from diagnosis appear longer simply because diagnosis occurs earlier, and length bias can make screening more likely to identify some slower-growing conditions. These concepts are especially relevant when interpreting survival or disease-course comparisons. They do not establish that either bias occurred in this study. The primary endpoint is stage at diagnosis, and the registry record does not provide enough result or design detail to evaluate the effects of these mechanisms.

What do the planning calculations mean?

The planning illustration is described as using the observed effect, a two-sided significance level of 0.05, equal allocation between the arms, and a normal approximation. It is explicitly not the trial’s own design calculation. A planning calculation asks how much information a future study might require under stated assumptions about the effect to detect, the significance level, allocation, and statistical model. Its answer depends on those assumptions and does not demonstrate that the completed study had adequate power. It is not observed power, which cannot resolve uncertainty in a result after the fact.

The record says the effect to detect is a stated fraction of the observed effect, but does not report the underlying primary outcome counts or a numerical observed effect. The numerical planning results are not reported in the registry record. These planning assumptions therefore cannot support a numerical interpretation of the primary comparison, and they should not be mistaken for evidence that the original study was designed around that calculation.

What can be concluded from the posted analysis information?

The record lists no analyses and gives no primary endpoint result. This means an analyst cannot check the chosen model, assumptions, handling of sparse stage categories, or agreement between a planned and reported analysis. It also means no uncertainty estimate can be interpreted. The endpoint definition and design description support an understanding of the intended comparison, but they do not substitute for analysis results. Any statistical conclusion about differences in stage would require the group-specific data or a reported estimate with its method and uncertainty.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Screening Group", "Non-Screening Group")
n      <- c(73, 192)
events <- c(46, 85)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Screening Group  73  46  63.0  50.9  74.0  51.5  73.2  51.5  73.2
ARM  Non-Screening Group  192  85  44.3  37.1  51.6  37.4  51.3  37.4  51.3

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists six outcome measures in total: the primary stage outcome and secondary measures concerning colorectal cancer grade, resection margins, metastasis, surgery, and palliative therapy. Considering multiple outcomes or multiple contrasts can increase the chance of a false-positive finding if each is tested without an appropriate plan. General methods for addressing this issue include designating a primary hypothesis in advance, adjusting inference across a defined family of tests, or treating secondary analyses as exploratory. The record does not report a multiplicity strategy or the planned hierarchy among secondary outcomes.

The record lists no analyses and does not report results for the primary or secondary outcomes. It is not possible to determine which outcomes were analysed, whether all planned outcomes were reported, or whether any adjustment was applied. The absence of a stated adjustment is not proof that none was used; the method is not reported in the registry record.

Missing-data handling is also not reported. The registry record does not say whether stage or secondary outcome information was missing, how missing observations were treated, or whether any sensitivity analysis was planned. Overall enrollment cannot establish how many participants contributed to each endpoint. Group-specific denominators and the analysis population are not reported.

The record does not describe whether analyses used all enrolled participants, only participants with available endpoint data, or another defined population. Nor does it give a participant flow by group. These omissions limit assessment of completeness and possible selection related to missing information. They do not establish that data were missing or that an analysis was biased; they mean the relevant details are unavailable in the registry record.

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

The registry record establishes the question, the group definitions, and the stage categories of interest. It does not report the stage counts or a comparative estimate. As a result, it supports no conclusion about whether screening-diagnosed cancers were at earlier stages in this study, how large any difference was, or how precisely it was estimated. A missing result is not a finding of no difference.

Even if an association between diagnostic route and stage were available, the nonrandomized design would matter. The groups were defined by screening-based versus symptomatic diagnosis, and the registry record cautions that they are not like-for-like controls. A difference could reflect the screening pathway, differences between the people in the groups, or other factors related to diagnosis. Without a reported analysis and relevant design details, the relative contribution of these possibilities cannot be assessed.

Stage at diagnosis is also not interchangeable with later outcomes. A difference in stage distribution would describe disease classification at diagnosis; it would not alone show an effect on survival, mortality, quality of life, or the consequences of treatment. General considerations such as lead-time and length bias can complicate interpretation of screening comparisons, particularly for outcomes measured after diagnosis. These are methodological cautions, not evidence that a particular bias occurred in this study.

The study’s stated purpose includes assessing whether screening-diagnosed cancer has more favourable clinical-pathological characteristics and confirming programme efficiency. Those aims should not be read as results. The record lists secondary outcomes related to grade, margins, metastasis, surgery, and palliative therapy, but does not report their findings. No medical decision or claim about the effectiveness of screening follows from the available registry information. A sound interpretation would require the outcome data, a clearly described analysis, and an account of uncertainty and the observational design.

13. Why This Trial Matters Statistically

This study illustrates a central issue in clinical biostatistics: a well-defined clinical question does not by itself make a comparison causal or provide evidence about its result. Screening-based and symptom-based diagnoses form groups through different clinical pathways, not through randomized assignment. Statistical interpretation must therefore distinguish a descriptive association from an effect attributable to screening.

The endpoint also highlights the importance of matching the analysis to the scientific question. Stage is recorded in categories, and the registry identifies some categories as having a better outcome. An analysis of the complete stage distribution and an analysis combining selected categories would answer related but not identical questions. A clear prespecified contrast, appropriate uncertainty estimates, and transparent reporting would help readers understand what was compared.

Reporting completeness is just as important as endpoint selection. The registry record gives enrollment and an endpoint definition, but not group sizes, stage counts, effect estimates, confidence intervals, or analysis methods. Without these, readers cannot assess the magnitude or precision of a difference, evaluate the contribution of individual categories, or judge how missing information was handled. The record lists no analyses, so it does not support reconstruction of the result.

Finally, the planning illustration should be kept separate from evidence about the completed study. A sample-size calculation based on an observed effect can be useful for planning under explicit assumptions, but it is not the study’s original design justification and does not establish observed power. This distinction prevents planning quantities from being mistaken for findings. The case demonstrates why transparent outcome reporting, design context, and careful separation of association, inference, and planning are essential to interpreting clinical research.

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 main limitation for interpreting this trial is that the registry record does not report the primary endpoint results. Stage-specific counts, group-specific denominators, and a comparison estimate are unavailable. Precision cannot be assessed because no confidence interval or other uncertainty measure is reported, and the record lists no analyses. The enrollment figure alone cannot fill these gaps or show how participants were distributed across groups or stages.

The comparison is nonrandomized, and allocation is not stated. The screening and non-screening groups are defined by diagnostic route and symptom status, and the registry record explicitly says the arms are not a like-for-like control comparison. Any observed difference could therefore be influenced by group characteristics or the process leading to diagnosis. The registry record does not describe adjustment for potential confounders or provide information with which to evaluate such adjustment.

The outcome is stage at diagnosis, not a later outcome. The record does not report survival or other follow-up results for the primary endpoint, and the stated timeframe is at diagnosis. Thus, even a reported stage difference would not alone establish a difference in subsequent health outcomes. General screening-related biases may also require consideration, but the registry information does not establish whether or how they affected this study.

Additional limitations include the lack of reported eligibility and recruitment details, group sizes, missing-data procedures, analysis populations, statistical methods, and multiplicity handling. The record lists secondary outcomes but does not report their results. These are reporting gaps; they should not be interpreted as proof that the study lacked a protocol or that particular methods were not used. They do limit independent assessment of the study and its conclusions.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Colorectal Cancer: Screening vs. Non-Screening — NCT02727894.
• Registry record last updated 2019-03-05; status last verified 2019-03.
• Results first posted 2019-03-05.
• 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.

18. Record Summary

NCT02727894 is a completed study titled “Colorectal Cancer: Screening vs. Non-Screening,” sponsored by the Institute for Clinical and Experimental Medicine. The registry record lists enrollment of 276, a start of 2013-03, and primary completion of 2015-12.

The study compares colorectal cancer diagnosed through screening with cancer diagnosed in symptomatic patients. Its primary endpoint is participant counts by pTNM stage at diagnosis; stages 0, I, and II were considered to have a better outcome. The record also lists secondary outcomes concerning grade, resection margins, metastasis, surgery, and palliative therapy.

The comparison is nonrandomized, allocation is not stated, and the record cautions that the groups are not like-for-like controls. It lists no analyses and does not report stage counts, group-specific denominators, effect estimates, confidence intervals, p-values, missing-data handling, or analysis populations. The design and question are described, but the posted registry information does not establish the comparative result.