Prothrombin Times Outside the Therapeutic Range in Otherwise Stable Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prothrombin Times Outside the Therapeutic Range in Otherwise Stable Patients (NCT00814177), a randomized phase 2 study. The primary outcome is Number of Patients With Prothrombin Time Results Within the Therapeutic Range After 2 Weeks.

2. Clinical Question

The clinical question concerns what to do after a moderately out-of-range result in a patient whose warfarin dose has otherwise been stable. The registry summary describes possible reasons for such results, including missed doses, changes in food or herbal agents, other drugs, and laboratory error. It also notes that fluctuations may occur between monitoring visits. These possibilities provide context for the comparison, but the record does not say that any particular cause was established for participants in this trial.

The outcome focuses on whether a follow-up result was within range, rather than on the size or direction of the change in the prothrombin time. It therefore addresses short-term return to the therapeutic range as recorded at the follow-up assessment. It does not, by itself, describe longer-term control, the causes of out-of-range results, or outcomes beyond the stated endpoint.

3. Trial Design

This was a phase 2, randomized, parallel-group trial with no masking. The registry reports an enrollment of 160 and describes two groups: no change in warfarin dose and a dose change based on whether the initial result was subtherapeutic or supratherapeutic. Each group has 80 participants recorded for the primary outcome. The sponsor was Hamilton Health Sciences Corporation. The trial status is completed, with a start date of 2006-07 and a primary completion date of 2009-01.

The comparison is between a management strategy involving a one-dose adjustment and a strategy of leaving the dose unchanged. Because the trial was unmasked, participants and study personnel were not blinded according to the registry’s masking description. Further details about the randomization procedure, allocation concealment, recruitment sites, or how treatment assignment was implemented are not reported in the registry record.

The trial concerned patients taking warfarin who were otherwise stable but had a prothrombin time outside the therapeutic range. The listed conditions are atrial fibrillation, venous thromboembolism, ischemic stroke, and myocardial infarction. The registry record does not provide further eligibility criteria, the clinical thresholds used to define an out-of-range result, or details about the time since the prior dose adjustment.

The primary outcome was assessed at 2 weeks. The record identifies one posted outcome measure and no posted secondary outcomes. Although the trial is labeled phase 2, the registry record does not describe the intended role of that phase designation or the original design assumptions. The trial’s own sample-size justification, target effect, and planned analysis strategy 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.

Before assignment: The study was planned for the sample size of 326 but due to competing studies, testing new dosing algorithms, the inclusion rate was much lower than expected.

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 Number of Patients With Prothrombin Time Results Within the Therapeutic Range After 2 Weeks, measured over 2 weeks.

The primary endpoint is the number of participants with follow-up prothrombin time results within the therapeutic range after 2 weeks. The registry describes the measure as comparing participants with follow-up results in range after a single dose was skipped, reduced, or added with participants whose dose was unchanged.

This is a binary outcome at the participant level: each participant either has a follow-up result recorded within the therapeutic range or does not. A count in each group provides a direct summary of how often the defined short-term status was observed. It is not a measure of the average prothrombin time, the amount by which a result moved, or the length of time spent in range.

The endpoint is relevant to the trial’s question because the competing strategies differ in whether a temporary dose adjustment is made after an out-of-range result. The registry does not define the therapeutic range numerically, describe how borderline results were classified, or state how results that were unavailable at follow-up were handled. The meaning of the outcome therefore depends on the registry’s stated classification, without enough detail to independently assess the threshold or measurement process.

The endpoint captures status at the stated follow-up, not longer-term anticoagulation control or clinical events. The registry record does not report a broader outcome measure alongside this primary endpoint.

7. Results Posted in the Registry

• Change: 56.2% (45/80) versus 60.0% (48/80) on control; risk difference -3.7 percentage points (95% CI -19.0 to 11.5), risk ratio 0.94 (0.72–1.22); unadjusted two-proportion z-test p = 0.631.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a risk difference describe?

A risk difference compares the proportion with an outcome in one group with the proportion in another. Its scale is an absolute difference in outcome frequency. A value below zero for the change group relative to the no-change group points toward a lower observed outcome proportion in the change group; a value above zero points toward a higher proportion. The registry’s posted estimate is negative. The participant counts also show a numerically lower in-range count in the change group, 45 versus 48, with 80 recorded in each group.

As a general statistical principle, an absolute difference can be easier to relate to the frequency of an outcome than a relative measure, but it does not remove uncertainty. The point estimate is the best single summary under the reported calculation, not a guarantee of the true difference.

How should the confidence interval be understood?

A confidence interval describes the range of values reasonably compatible with the data under the method used to calculate it. In repeated use of a method, intervals constructed in this way would cover the underlying value at the method’s stated confidence level at the corresponding frequency. It is not a range containing a known fixed proportion of individual participant outcomes, and it is not a probability statement that the true effect lies within this particular interval.

For this comparison, the reported risk-difference interval includes zero and extends into both negative and positive values. Zero represents no difference on the risk-difference scale. This means the data do not rule out no difference and remain compatible with effects in either direction. The interval’s span also indicates that the estimate is not precise enough to pin down a narrow range of possible differences.

What does the p value contribute?

The reported unadjusted p value is 0.631. A p value describes how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor does it give the probability that the result occurred by chance. A larger p value does not establish that groups are the same.

Here, the p value is consistent with the confidence interval’s inclusion of no difference: the comparison does not provide strong evidence against a no-difference null on the reported analysis. The p value does not indicate whether a possible difference would matter clinically. That judgment would require a clinically meaningful threshold, which is not reported in the registry record.

Why are relative measures also reported?

A risk ratio compares outcome proportions as a ratio, while an odds ratio compares the odds of the outcome. Both are relative scales and answer a different descriptive question from an absolute risk difference. A ratio near one represents little relative separation; values on either side of one represent different directions. Their confidence intervals quantify uncertainty on their respective scales. The registry reports these measures, but they are not separate outcomes or independent evidence: they use the same participants and endpoint as the risk-difference comparison.

Odds and risks are not interchangeable concepts. The odds are formed by comparing the number with the outcome to the number without it, whereas risk refers to the proportion with the outcome. The distinction can matter when interpreting the size of a relative estimate. The registry record does not identify a preferred effect scale beyond posting the risk difference, risk ratio, and odds ratio.

What does randomization contribute?

In general, random assignment is intended to make comparison groups similar, on average, with respect to measured and unmeasured characteristics at the time of assignment. It supports a causal interpretation when the trial is conducted and analyzed appropriately, but does not guarantee balance in a particular trial. The record identifies the allocation as randomized but does not describe the sequence generation, concealment, adherence, or analysis population. Those details cannot be assessed from the registry entry.

The trial was unmasked. In general, lack of masking can affect behavior, co-interventions, or assessment, depending on the study procedures and outcome. The endpoint is based on a laboratory result, but the registry does not report how laboratory assessment was conducted or whether those assessing results knew group assignment. It would be inappropriate to infer protections against bias that are not described.

Why can variation matter?

As a general statistical concept, outcome variation affects how precisely a study can estimate a group difference. For a binary endpoint, the observed number with the outcome and the number assessed in each group contribute to uncertainty. When the estimate is uncertain, a confidence interval can include a range of possible effects even if the point estimate is near no difference. The registry’s interval crosses zero, so the direction is not resolved by the posted comparison.

These methods describe the statistical information in the recorded endpoint. They do not establish the reason for a participant’s follow-up result, the biological mechanism behind a change, or the best clinical response to an out-of-range result. The registry record does not report further model details, adjustment variables, or sensitivity analyses.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate the effect estimate. The posted risk-difference confidence interval spans values below and above zero. Because zero is included, the data are compatible with no difference as well as with a difference favoring either group. The interval’s breadth means that the point estimate should not be read as a precise measure of the underlying group difference.

The result is based on the posted endpoint counts: 48 of 80 in the no-change group and 45 of 80 in the dose-change group. Counts alone do not show how much uncertainty surrounds the comparison; the confidence interval adds that context. The reported unadjusted p value, 0.631, is also consistent with the absence of strong statistical evidence against no difference. Neither the interval nor the p value proves that the groups have identical outcomes.

Relative measures are reported as well, but their intervals also express uncertainty. A risk ratio and an odds ratio use different scales from a risk difference, so their values should be interpreted according to their definitions. The registry’s results do not identify a clinically important threshold for the endpoint. Without such a threshold, statistical precision cannot be translated into a judgment about whether a possible difference would matter clinically.

The registry record does not report handling of missing endpoint data, so the precision of the posted comparison cannot be evaluated in light of any unavailable follow-up results. It also does not report an analysis population or sensitivity analyses. The confidence interval therefore describes uncertainty under the posted comparison, not every possible source of uncertainty in trial conduct or outcome measurement.

11. 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 figures are participants per arm, not total events. They are illustrations based on the observed effect, not the trial’s original design calculation. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The trial’s own original design assumptions are not reported in the registry record.

For planning for 100% of the observed effect, the illustration gives 2,716 participants per arm for 80% power and 3,636 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 4,814 participants per arm for 80% power and 6,444 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 10,795 participants per arm for 80% power and 14,451 participants per arm for 90% power.

In plain language, the 80% power column describes the planning size associated with an 80% chance of detecting the specified effect under the stated assumptions if that effect were present; the 90% column gives the corresponding planning size for a higher chance of detection. These are planning illustrations, not observed power. Smaller effects require more participants to detect under the same assumptions, which is why the figures increase as the planned effect becomes a smaller fraction of the observed effect.

The curve should be read as a visual representation of the same planning relationship, not as a separate result from the trial. These calculations do not show that the trial was designed with these assumptions or that its observed finding would recur in a future study.

12. 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 = 0.600; treatment proportion p₁ = 0.562; average p̄ = 0.581.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6977 = 1.3675.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6972 = 0.5868.
4. n = (1.3675 + 0.5868)² / (-0.0375)² = 2,715.8, rounded up to 2,716 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 <- 0.6                          # control proportion
p1 <- 0.5625                          # 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(814177)
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 = 2,716 (80% power), 3,636 (90% power)
 75% of observed effect: n per arm = 4,814 (80% power), 6,444 (90% power)
 50% of observed effect: n per arm = 10,795 (80% power), 14,451 (90% power)
power.prop.test: 2715.8 
Simulated power at n = 2,716 per arm: 0.798 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and no posted secondary outcomes. The record does not report whether other outcomes were collected but not posted, whether the primary outcome was selected before the data were examined, or whether any multiplicity adjustment was planned. Therefore, it is not possible to assess from this entry whether the reported p value accounts for other statistical tests that may have been conducted.

The comparison is reported with an unadjusted p value of 0.631. The registry record does not describe a prespecified statistical analysis plan, model details, covariate adjustment, interim analyses, or rules for choosing among analyses. These items are not reported in the registry record and should not be inferred from the presence of a posted estimate.

Missing-data handling is also not reported in the registry record. It is unclear whether every randomized participant had a 2-week result, how unavailable results were classified, or whether any imputation or other method was used. The posted group counts and denominators are the information available for the primary endpoint, but they do not, on their own, describe the disposition of participants or the handling of missing observations.

The analysis population is not reported in the registry record. In particular, the entry does not state whether the comparison followed an intention-to-treat approach, included only participants with observed follow-up, or used another population definition. The randomization label alone does not answer this question. Without population and missing-data details, the extent to which the posted comparison preserves the randomized groups cannot be assessed.

These omissions limit assessment of robustness. They do not prove that the analysis was inappropriate; they mean the registry entry does not provide enough information to evaluate these aspects. No additional analytical conclusions should be inferred from the absence of posted details.

14. 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 non-serious adverse event terms

The posted safety figures record zero serious adverse events in the no-change group and zero in the dose-change group. For other adverse events, the record lists 2 in the no-change group and 1 in the dose-change group. Each group has 80 participants recorded as at risk for these safety summaries. The registry provides no serious adverse event terms in its listed top serious terms, and deaths are not reported in the registry record.

These figures are counts in the registry record, not a complete description of the nature, timing, severity, or relationship of events to warfarin or the dose strategy. The entry does not give definitions for serious or other adverse events, explain how events were collected, or describe the follow-up period for safety reporting. A zero posted count should be understood as zero reported in this record, not as proof that no event could have occurred.

The small counts should not be read as a reliable comparison of safety between groups. The registry does not report a formal safety analysis or uncertainty estimates for these event counts. These data are descriptive and do not establish that one strategy is safer than the other.

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

The primary result describes a short-term laboratory endpoint: after 2 weeks, the registry records 48 participants with prothrombin time results within the therapeutic range in the no-change group and 45 in the dose-change group. The reported comparison has a negative risk-difference point estimate for dose change relative to no change, but its confidence interval includes zero and the unadjusted p value is 0.631. The data therefore do not provide strong statistical evidence of a difference in this endpoint.

This is not proof of equivalence. A non-significant result can arise when the groups truly differ, when they are similar, or when the available data do not estimate the difference precisely enough. The confidence interval crossing zero illustrates that uncertainty. The registry does not report a clinically meaningful difference threshold or an equivalence or non-inferiority objective.

The result also does not establish that either approach is beneficial or ineffective. It reports the observed outcome counts and a statistical comparison; it does not show what would happen under other care strategies, at other follow-up times, or in populations with eligibility criteria different from those used in the trial. The registry gives no further detail about the sample’s clinical characteristics or the definition of stable patients.

The endpoint should not be confused with longer-term anticoagulation control or clinical outcomes. It records whether a result was within range at the stated follow-up, not the duration of control or the occurrence of thrombosis, bleeding, or other events. The registry record does not report such secondary outcomes. It also does not provide enough information to judge missing-data handling, analysis population, or the measurement process.

Accordingly, the posted result is best interpreted as an uncertain comparison of short-term in-range counts in this randomized trial. It neither demonstrates a difference nor rules one out, and it is not a basis for a treatment recommendation.

16. Why This Trial Matters Statistically

This trial addresses a statistical problem common to clinical monitoring: a measurement can fall outside a target range even when a patient has otherwise been stable. The comparison asks whether a temporary adjustment is associated with a different short-term in-range result from leaving the dose unchanged. The randomized design makes the question experimentally testable, while the unmasked design and limited registry detail leave important aspects of conduct and analysis unreported.

Its main statistical lesson is the distinction between a point estimate and the uncertainty around it. The observed counts differ, and the reported risk-difference estimate points toward a lower outcome proportion in the change group. Yet the confidence interval includes no difference and values favoring either group. A point estimate alone would give an incomplete account; the interval shows why the direction and size of the underlying difference remain uncertain.

The result also illustrates why a non-significant comparison should not be translated into proof that strategies are identical. The reported p value is not a measure of the probability that the approaches are equivalent, and the trial is not identified as an equivalence or non-inferiority study. The registry record does not report the clinically meaningful threshold that would help interpret the practical importance of possible effects.

Finally, the endpoint is deliberately narrow: it records whether the follow-up result was within range after 2 weeks. That makes the comparison interpretable as a short-term outcome, but it cannot answer questions about sustained control or clinical events. The distinction between what the endpoint measures and what remains unknown is central to sound statistical interpretation.

17. 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 registry record offers limited detail about the trial population. It names several conditions and describes patients as otherwise stable, but does not report the full eligibility criteria, the definition of stable status, or the threshold defining an out-of-range prothrombin time. This makes it difficult to judge how closely the participants represent other clinical settings.

The endpoint is recorded at 2 weeks, and the therapeutic-range definition is not reported in the registry record. The entry does not describe the assay or assessment procedures, how borderline values were classified, or whether outcome assessors were masked. Although the trial is listed as unmasked, the consequences of that design feature for outcome assessment cannot be determined from the available description.

Important analysis details are also absent. The original sample-size rationale, prespecified analysis plan, analysis population, missing-data methods, and any interim analyses are not reported in the registry record. These gaps prevent assessment of whether the posted comparison followed a prespecified plan and how unavailable outcomes, if any, were handled.

The posted result concerns one primary outcome and no secondary outcomes are listed. There is no registry detail on longer-term follow-up or clinical outcomes, so the result cannot address sustained control or broader consequences. Safety reporting is limited to event counts without descriptions, collection methods, or a formal comparison.

Finally, the confidence interval includes no difference and effects in either direction, leaving substantial uncertainty about the size and direction of the underlying difference. The observed counts are descriptive of this trial’s posted endpoint, not conclusive evidence that the groups are the same or that one strategy is preferable.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Prothrombin Times Outside the Therapeutic Range in Otherwise Stable Patients — NCT00814177.
• Registry record last updated 2012-08-01; status last verified 2012-07.
• Results first posted 2009-11-25.
• 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.

21. Record Summary

NCT00814177 is a completed, randomized, unmasked, parallel-group phase 2 trial sponsored by Hamilton Health Sciences Corporation. It enrolled 160 and compared no change in warfarin dose with a one-dose adjustment for otherwise stable patients whose prothrombin time was outside the therapeutic range. The registry lists atrial fibrillation, venous thromboembolism, ischemic stroke, and myocardial infarction as conditions.

The primary endpoint was the number of participants with a follow-up result within the therapeutic range after 2 weeks. The posted counts were 48 of 80 for no change and 45 of 80 for dose change. The reported risk-difference interval includes zero, and the unadjusted p value is 0.631. The registry lists no secondary outcomes and provides limited detail about analysis methods, missing data, and safety-event ascertainment.