The Evaluation of a Standard Treatment Regimen of Anti-tuberculosis Drugs for Patients With MDR-TB: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Evaluation of a Standard Treatment Regimen of Anti-tuberculosis Drugs for Patients With MDR-TB (NCT02409290), a randomized phase 3 study. The primary outcome is STREAM Stage 2 Primary Outcome Measure (the Proportion of Patients With a Favourable Outcome at Week 76).

2. Clinical Question

The trial record identifies STREAM Stage 2 as a Phase 3, randomized, open-label, parallel trial. The central statistical question is about the frequency of a favorable Week 76 status in the named analysis population, not about a change in symptoms or a time-to-event outcome. The registry record does not report the detailed criteria used to classify an outcome as favorable, or the clinical components that could lead to that classification.

3. Trial Design

The registry record describes a randomized, open-label, parallel Phase 3 trial of treatments for MDR-TB. It lists enrollment as 588 and status as completed. The recorded start is 2016-03, and the primary completion date is 2022-05-13. IUATLD, Inc is listed as the sponsor. These registry entries describe the trial record; they do not establish how many participants contributed to each particular analysis.

The posted primary outcome is the proportion of patients with a favorable outcome at Week 76 in the modified intention-to-treat population. The primary comparison includes Regimen B as the control and reports comparisons for Regimen C and Regimen D. The registry also lists Regimen A, but its primary outcome count and participant number are both recorded as 0. The reason for those zero values is not reported in the registry record, so they cannot be interpreted as evidence about Regimen A's outcomes.

Randomization means participants are assigned to treatment groups by an allocation process intended to limit systematic differences between groups. The allocation sequence, concealment procedures, and the details of how assignment was carried out are not reported in the registry record. Open-label means the trial is listed as having no masking. The record does not specify which people involved in care, outcome assessment, or analysis knew the assigned regimen.

The primary endpoint is recorded at a fixed follow-up time, Week 76, rather than as the time until an event. The registry record does not report the original sample-size justification, the trial's planned statistical analysis in detail, interim analyses, or whether the posted comparisons correspond to all analyses in the protocol. Those details should not be inferred from the completed status or the enrolled total.

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: 1436 participants were screened and 588 were randomly assigned to treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: mITT population

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 STREAM Stage 2 Primary Outcome Measure (the Proportion of Patients With a Favourable Outcome at Week 76), measured over 76 weeks.

The primary endpoint is the proportion of patients in the modified intention-to-treat population with a favorable outcome at Week 76. A proportion describes the number of participants meeting a defined outcome relative to the number assessed in the relevant group. Here, the registry posts counts of favorable outcomes and participants for Regimens B, C, and D, along with a comparison between each of the latter regimens and Regimen B.

A fixed-time binary endpoint gives a direct comparison of how often a specified status was recorded at a chosen follow-up point. It is different from an endpoint that measures how quickly an event occurs, because the posted primary measure does not present event timing. The registry's brief outcome description names favorable status at Week 76 but does not report the detailed outcome-classification rules, how deaths or other events were incorporated into the classification, or how participants without a Week 76 assessment were handled.

Using a defined analysis population matters because the participants contributing to an endpoint can differ from the total enrollment. The registry names the modified intention-to-treat population, but its exact definition is not reported in the registry record. The endpoint can therefore be interpreted as the recorded Week 76 favorable-outcome proportion for that named population, while the precise eligibility and data-handling rules behind the population remain unclear.

The endpoint is used to compare the observed frequency of favorable status across regimens. It does not, by itself, explain why outcomes differed, describe every dimension of treatment experience, or establish how results apply beyond the population and follow-up represented in the record.

7. Results Posted in the Registry

• Regimen C (Oral Regimen): 82.7% (162/196) versus 71.1% (133/187) on control; risk difference 11.5 percentage points (95% CI 3.1 to 19.9), risk ratio 1.16 (1.04–1.30); unadjusted two-proportion z-test p = 0.007.
• Regimen D (6-month Regimen): 91.0% (122/134) versus 71.1% (133/187) on control; risk difference 19.9 percentage points (95% CI 11.8 to 28.0), risk ratio 1.28 (1.15–1.42); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

A binary outcome records whether each participant did or did not meet a specified condition. In this trial, the condition is a favorable outcome at Week 76. The reported counts of favorable outcomes and participants form the basis for group proportions and comparisons. The registry record does not give the detailed rules for classifying favorable status, so the statistical interpretation concerns the recorded classification rather than any unreported component of it.

What does a risk difference express?

A risk difference subtracts one group’s outcome proportion from the other group’s proportion. It uses the same scale as the proportions themselves and is therefore an absolute measure: it describes the separation between groups in the frequency of the endpoint. A value of zero would represent no difference in proportions. A positive value in these comparisons indicates a higher favorable-outcome proportion in the Regimen C or Regimen D group than in Regimen B, based on the direction used in the posted comparisons. The reported intervals for both estimates remain above zero.

Absolute differences are often useful because they keep the comparison tied to the outcome frequency in the groups being studied. They should be read alongside the underlying group counts and proportions, not as an explanation of why the groups differed. The registry record posts risk-difference estimates to four decimal places, but the numerical estimates are not repeated in this explanation; the posted relative measures and counts give other ways to read the comparison.

How do risk ratios and odds ratios differ?

A risk ratio divides one group’s outcome proportion by the other group’s proportion. A value of 1 represents equal proportions, while a value above 1 represents a larger proportion in the numerator group. The posted risk-ratio estimates are 1.162 for Regimen C versus Regimen B and 1.28 for Regimen D versus Regimen B. Their intervals are 1.039 to 1.299 and 1.152 to 1.423, respectively.

An odds ratio compares odds rather than proportions. Odds are the number meeting the endpoint relative to the number not meeting it. An odds ratio of 1 represents equal odds. The reported odds ratios are 1.935 for Regimen C versus Regimen B and 4.128 for Regimen D versus Regimen B; their intervals are 1.189 to 3.147 and 2.108 to 8.083. When an outcome is common, an odds ratio can be numerically farther from 1 than the corresponding risk ratio. This is a general property of the scales, not evidence that one measure is more important.

What do the p-values say?

The p-values are reported as unadjusted: 0.007 for Regimen C versus Regimen B and < 0.001 for Regimen D versus Regimen B. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null model and test assumptions. It is not the probability that the null hypothesis is true, nor does it quantify the size or practical importance of a difference. The registry record does not report the full testing procedure or whether additional analyses were performed.

The word “unadjusted” is important when more than one comparison is made. In general, testing several hypotheses can increase the chance of at least one apparently unusual result if no allowance is made for multiplicity. The registry record does not report a multiplicity adjustment, so the p-values should be read as the posted unadjusted values rather than assumed to control an overall error rate across comparisons.

What role do confidence intervals play?

In general, a confidence interval gives a range of estimates compatible with the data under a statistical method and its assumptions. A narrower interval indicates greater precision on the relevant scale; a wider interval indicates greater uncertainty. For a risk difference, an interval containing zero is compatible with no group difference at the interval’s stated level, while an interval wholly on one side of zero indicates estimates in one direction across that range. Both posted risk-difference intervals are above zero. The confidence level and the interval-construction method are not reported in the registry record, so they should not be assigned a level that the record does not state.

The relative-measure intervals are also reported, and each lies above 1. In general, 1 is the no-difference value for a risk ratio or odds ratio. These intervals describe uncertainty on relative scales and should not be read as a range of individual participant outcomes. The analysis population definition, handling of missing outcomes, model details, and any covariate adjustment are not reported in the registry record.

10. Confidence Intervals and Precision

The posted risk-difference intervals for Regimen C versus Regimen B and Regimen D versus Regimen B are both entirely above zero. This indicates that the reported interval ranges are consistent with a positive difference in favorable-outcome proportions under the methods used. Their numerical bounds are reported in the registry record, but are not repeated because they have more than three decimal places. The confidence level and interval method are not reported in the registry record.

Precision is not the same as whether an estimate points in a particular direction. In general, an interval shows how much uncertainty remains around an estimate under a statistical method. Narrower intervals represent greater precision on that scale; wider intervals represent less precision. The posted relative-measure intervals also lie above their no-difference value of 1: for the risk ratios, the ranges are 1.039 to 1.299 for Regimen C and 1.152 to 1.423 for Regimen D; for the odds ratios, they are 1.189 to 3.147 and 2.108 to 8.083, respectively.

These intervals do not account for every possible source of uncertainty. Their interpretation depends on the analysis, the endpoint definition, the analysis population, and assumptions about the data. The registry record does not report the method used to construct the intervals, the confidence level, the missing-data approach, or any adjustment for multiple comparisons. The estimates should therefore be read as the posted statistical summaries, not as guarantees about the range of outcomes in future participants.

Counts are also relevant to precision: the primary outcome record includes 187 participants in Regimen B, 196 in Regimen C, and 134 in Regimen D. The total enrollment is listed separately as 588. The difference between total enrollment and the counts in these outcome groups cannot be explained from the registry record alone, and no reason for that difference should be assumed.

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 sample-size figures are planning illustrations based on the observed effects in the posted comparisons, not the original trial design. For this binary outcome, the figures mean participants per arm. They are not total event counts. The calculation 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 set to the stated fraction of the observed effect. The trial’s own original sample-size assumptions and justification are not reported in the registry record.

The 80% power column gives the illustrated number of participants per arm under which the planning calculation would have an 80% chance of detecting the selected effect if that effect were the true one and the stated assumptions held. The 90% power column gives the corresponding number for a 90% chance. These are planning concepts, not guarantees for a particular study. The displayed values are not observed power.

For Regimen C versus Regimen B, planning for 100% of the observed effect gives 209 participants per arm for 80% power and 279 for 90% power. Planning for 75% of the observed effect gives 388 and 519 participants per arm; planning for 50% gives 908 and 1,215. For Regimen D versus Regimen B, the corresponding figures are 60 and 80 for 100% of the observed effect, 118 and 157 for 75%, and 287 and 384 for 50%.

Smaller effects are harder to distinguish from random variation, so the planning illustration requires more participants per arm as the targeted fraction of the observed effect decreases. These calculations are hypothetical and use the observed effect as their starting point; they do not describe the trial’s original design or establish what power the completed trial had.

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 = 0.711; treatment proportion p₁ = 0.827; average p̄ = 0.769.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5962 = 1.1685.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5906 = 0.4970.
4. n = (1.1685 + 0.4970)² / (0.1153)² = 208.6, rounded up to 209 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 <- 0.711229946524064                          # control proportion
p1 <- 0.826530612244898                          # 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(2409290)
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 = 209 (80% power), 279 (90% power)
 75% of observed effect: n per arm = 388 (80% power), 519 (90% power)
 50% of observed effect: n per arm = 908 (80% power), 1,215 (90% power)
power.prop.test: 208.6 
Simulated power at n = 209 per arm: 0.803 (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 posts two primary comparisons: Regimen C versus Regimen B and Regimen D versus Regimen B. It reports an unadjusted p-value for each. Whether a multiplicity procedure was planned or applied is not reported in the registry record. In general, testing more than one hypothesis can affect the interpretation of nominal p-values; without a reported adjustment plan, the posted values should not be described as controlling a family-wise error rate.

The registry record lists no posted secondary outcome results. It does not describe whether other outcomes were collected, how many analyses were planned, or how the primary comparison was selected from any analysis plan. The original statistical analysis plan and interim analyses are not reported in the registry record. Accordingly, the posted comparisons can be described, but their relationship to every planned analysis cannot be established from this record.

The named analysis population is modified intention-to-treat. The precise definition of that population is not reported in the registry record. The record also does not explain the handling of participants with missing Week 76 assessments, withdrawals, protocol deviations, or other incomplete outcome data. No imputation method, sensitivity analysis, or missing-data assumption is reported. As a result, the available account does not show how the posted group counts relate to participants who may have lacked an assessable endpoint.

Randomization is listed, but the record does not provide allocation concealment details or the randomization method. Masking is listed as none. The registry does not report whether outcome classification was independently reviewed or whether the analysis adjusted for baseline characteristics. These omissions do not change the posted counts, but they limit the ability to reconstruct or fully assess the analysis process.

The enrollment total and the primary outcome group counts are not identical sets of figures in the record. No explanation for the difference is reported, so it should not be attributed to loss to follow-up, exclusions, or any other specific cause. The Regimen A primary-outcome row contains zero participants, but the reason is likewise not reported. These are relevant boundaries on interpretation, rather than evidence about the unreported data handling.

14. Safety Results

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

Deaths

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: gastrointestinal disorders (6.2%), psychiatric disorders (6.2%), infections and infestations (5.9%), hearing and vestibular disorders (smq) (4.9%), general disorders and administration site conditions (4.2%).

The registry reports serious adverse event counts and deaths separately from the primary efficacy outcome. For serious adverse events, it lists 44 affected participants among 202 at risk in Regimen B, 44 among 211 in Regimen C, and 33 among 143 in Regimen D. Regimen A is listed with 8 affected among 32 at risk. These are the counts reported in the safety record; they are not the primary endpoint counts, and the different safety and efficacy group sizes should not be conflated.

The death counts are 8 for Regimen B, 11 for Regimen C, 2 for Regimen D, and 0 for Regimen A. The record does not provide a comparative statistical analysis of these death counts or explain how they relate to the Week 76 favorable-outcome classification. They should therefore be read as descriptive registry counts, not as evidence of a treatment effect.

The registry also lists serious adverse event terms with percentages, including gastrointestinal disorders and psychiatric disorders at 6.2%, infections and infestations at 5.9%, and hearing and vestibular disorders at 4.9%. The record does not specify in this listing the denominator or whether these terms overlap within participants. The percentages should not be added or treated as mutually exclusive categories. The counts and terms alone do not provide a complete comparative safety analysis.

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

The results describe the recorded frequency of a favorable Week 76 outcome in the modified intention-to-treat population. The posted group counts are 133 of 187 for Regimen B, 162 of 196 for Regimen C, and 122 of 134 for Regimen D. For both Regimen C and Regimen D, the registry reports a risk-difference interval entirely above zero and an unadjusted p-value below the conventional 0.05 threshold, with the reported values being 0.007 and < 0.001, respectively. These are statistical summaries of the comparisons, not a complete account of treatment effects.

The posted evidence does not establish the clinical importance of the difference. The registry does not report a threshold for a clinically meaningful difference, detailed endpoint classification rules, or a broader assessment of benefits and harms. Nor does the result explain the mechanism behind the group differences. Statistical evidence about a group comparison is not a prediction of what will happen to any individual participant.

The comparisons do not directly test Regimen C against Regimen D. Each is compared with Regimen B. The record also does not explain the zero participant count in the Regimen A primary outcome row, so the primary outcome information cannot be used to make a meaningful comparison involving Regimen A.

Interpretation is limited by information not reported in the registry record, including the exact modified intention-to-treat definition, handling of missing Week 76 outcomes, confidence level and interval method, and multiplicity plan. The trial is listed as open-label, and the record does not specify how outcome classification or analysis may have been protected from knowledge of assigned regimen. The results should be understood within these limits and the population and follow-up represented by the posted record.

Nothing in these statistical summaries is a recommendation about treatment. They report what the registry records for this trial’s selected endpoint and comparisons; they do not settle questions about other outcomes, later follow-up, or populations not represented in the analysis.

16. Why This Trial Matters Statistically

This trial illustrates why both absolute and relative measures are useful when interpreting a binary clinical outcome. The risk difference describes separation in favorable-outcome proportions on an absolute scale, while the risk ratio and odds ratio express the comparison on relative scales. Because the endpoint is a proportion and the outcome is common in the posted groups, the numerical values of the odds ratios are notably different in scale from the risk ratios. Reading more than one measure helps avoid treating a relative effect as if it were an absolute change.

The example also shows why the analysis population and endpoint definition matter. The primary outcome is named for a modified intention-to-treat population at Week 76, but the registry does not give the population’s exact definition or detailed classification rules for favorable status. Without those details, readers can interpret the posted counts and estimates but cannot fully reconstruct who qualified for analysis or how every outcome was classified.

Two comparisons share Regimen B as the control. This creates a multiplicity question: whether the analysis plan accounted for more than one comparison. The p-values are identified as unadjusted, while the registry record does not report a multiplicity plan. This does not erase the reported estimates, but it is important context when interpreting evidence across multiple comparisons.

The trial is listed as randomized and open-label. These design features are relevant to statistical interpretation: randomization is intended to support fair group comparisons, while lack of masking can matter for assessment or conduct depending on the outcome and procedures. The registry does not provide enough information to assess those processes in detail. Overall, this record offers a useful case for distinguishing what the numerical results directly describe from what cannot be resolved without fuller methods and outcome reporting.

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.

Several trial-specific gaps constrain interpretation. The registry names the primary endpoint and its Week 76 timeframe, but does not provide the detailed rules used to define favorable status. The exact modified intention-to-treat population definition is also not reported. These omissions make it difficult to determine how participants were classified and which participants were eligible for the posted analysis.

The reason for the zero participant and zero favorable-outcome values in the Regimen A primary outcome row is not reported. The record also lists 588 participants as enrollment, while the primary outcome counts are reported for Regimens B, C, and D. It does not explain how the enrolled total relates to those analysis counts. No cause, such as exclusions or missing follow-up, should be assumed.

The registry record does not report the handling of missing outcomes, the original statistical analysis plan, interim analyses, allocation concealment, detailed randomization procedures, or a multiplicity adjustment. It reports unadjusted p-values, but does not state the confidence level or construction method for the intervals. These details would help assess the robustness and intended interpretation of the comparisons.

Masking is listed as none. The record does not say whether outcome assessment or analysis was conducted by people unaware of treatment assignment. The safety counts are descriptive and have different group denominators from the primary efficacy counts; the record does not provide a comparative safety analysis or specify the denominator for the listed serious-event term percentages. Finally, no direct comparison of Regimen C with Regimen D is posted, and no secondary outcome results are listed.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: The Evaluation of a Standard Treatment Regimen of Anti-tuberculosis Drugs for Patients With MDR-TB — NCT02409290.
• Registry record last updated 2023-09-28; status last verified 2023-09.
• Results first posted 2023-09-28.
• 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

The registry record describes NCT02409290, a completed, randomized, open-label, parallel Phase 3 trial sponsored by IUATLD, Inc. Enrollment is listed as 588. The recorded start is 2016-03, and primary completion is dated 2022-05-13.

The primary outcome is favorable status at Week 76 in the modified intention-to-treat population. The posted comparisons are Regimen C and Regimen D versus Regimen B. The registry reports favorable-outcome counts, risk differences, relative measures, and unadjusted p-values for those comparisons. Regimen A has zero participants listed in the primary outcome record, with no explanation reported.

The record also includes serious adverse event counts, deaths, and selected serious-event terms. It does not report the original sample-size justification, detailed statistical analysis plan, multiplicity approach, missing-data handling, or the exact analysis-population definition. These absences limit reconstruction of the full analysis and should remain part of any interpretation of the posted results.