Reducing Preterm Births in Underserved Pregnant Women: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reducing Preterm Births in Underserved Pregnant Women (NCT01344616), a randomized phase 2 study. The primary outcome is Birthweight.

2. Clinical Question

The trial title focuses on reducing preterm births, but the stated primary endpoint is birthweight below 2500 grams. Those concepts are related in the broad clinical context but are not interchangeable outcomes: the registry’s primary result concerns the birthweight threshold, not a reported count of preterm births. Gestational weight is listed as a secondary outcome, while further outcome detail is not reported in the registry record.

The event counts were 26 among 195 participants in the lifestyle counseling arm and 40 among 213 participants in usual clinical care. These are the arm-specific figures attached to the primary outcome, and they are the basis for the posted unadjusted comparison. The population note about sets of twins matters when reading the record: it identifies multiple births in each arm, but does not describe how observations from the same pregnancy were handled in the analysis.

The registry record does not describe more detailed eligibility criteria, the recruitment setting, or whether the computer-based support was delivered in a particular format or intensity. It therefore supports a focused question about the recorded group comparison and endpoint, rather than a broader conclusion about all prenatal programs or all measures of preterm birth.

3. Trial Design

This was a phase 2, randomized, parallel trial with no masking, and its status is completed. The registry lists enrollment of 482. The recorded start was 2010-08, and primary completion was 2013-07. PHCC LP is listed as the sponsor. The trial title is “Reducing Preterm Births in Underserved Pregnant Women.”

The two arms were usual clinical care and lifestyle counseling. The usual-care description says that participants received usual clinical care with no intervention; the lifestyle-counseling description specifies computer-based clinical support to the patient. The intervention is categorized as behavioral. The registry does not report additional details about counseling frequency, duration, adherence, or the care that constituted usual clinical care.

The primary outcome is listed as birthweight, measured in participants, with the description defining the event as birth weight below 2500 grams. The stated timeframe is 9 months. The outcome population description records one set of twins in the lifestyle counseling arm and four sets in usual clinical care. The result is therefore presented as a binary event count, not as a mean birthweight or a continuous distribution of weights.

The arm counts for this endpoint are 195 in lifestyle counseling and 213 in usual clinical care. They do not sum to the registry’s overall enrollment of 482. The registry does not explain this difference, so it is not possible to determine from the record whether it reflects outcome availability, an analysis population, or another reporting issue. It is important not to infer a participant flow or reason for the difference without such documentation.

Although the trial is randomized, the record does not report the sequence-generation method, allocation concealment, stratification, or the planned approach to analysis. It also does not report an original sample-size justification or stopping rules. These details cannot be reconstructed from the posted arm counts or results.

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

Analysis population for the primary outcome, as defined in the registry: One Set of twins in the lifestyle counseling arm and 4 sets of twins in the usual clinical care arm

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 Birthweight, measured over 9 months.

The primary endpoint counts participants with a birthweight below 2500 grams. The registry labels the outcome “Birthweight,” gives the unit as participants, and specifies a 9-month timeframe. The endpoint is therefore treated as binary: each participant is counted according to whether the defined threshold event occurred. The reported counts are 40 in usual clinical care and 26 in lifestyle counseling.

A threshold-based endpoint can make a clinically relevant event easier to summarize than a full distribution of birthweights, and it supports a direct comparison of the proportion of participants experiencing the defined event. That is a general statistical rationale for binary outcomes, not a claim about why the trial’s investigators selected this endpoint. The registry does not report the endpoint-selection rationale.

The threshold also determines what the result can and cannot address. It concerns birthweight below 2500 grams, not the full range of birthweights, and it does not directly report preterm delivery. The title and brief summary discuss preterm births, but the primary result posted for this comparison is specifically the birthweight threshold outcome.

The registry describes participants as the measurement unit and notes sets of twins in both arms. It does not say whether both infants from a multiple pregnancy were included in the participant count, how birthweights within a twin set were mapped to participants, or how any dependence between twins was accommodated. This missing detail limits assessment of the endpoint’s exact construction and the independence assumptions used in analysis.

7. Results Posted in the Registry

• Lifestyle Counseling: 13.3% (26/195) versus 18.8% (40/213) on control; risk difference -5.4 percentage points (95% CI -12.5 to 1.6), risk ratio 0.71 (0.45–1.12); unadjusted two-proportion z-test p = 0.136.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is a comparison of a binary outcome: whether a participant had a birthweight below 2500 grams. Binary outcomes are often summarized as event counts and arm totals, then compared using measures such as a risk difference, risk ratio, or odds ratio. These measures answer related but distinct questions. A risk difference expresses the absolute separation between event proportions, while a risk ratio compares them multiplicatively and an odds ratio compares the odds. These are general definitions; the registry does not describe which measure was designated as primary in the trial’s analysis plan.

What does the risk difference describe?

The posted risk-difference estimate is negative for lifestyle counseling relative to usual clinical care. In the posted comparison, that direction corresponds to a lower observed event proportion in lifestyle counseling. The confidence interval runs from a negative value to a positive value, crossing zero, which represents no absolute difference. A confidence interval expresses uncertainty around an estimate under the method used to construct it. It is not a guarantee that the true effect lies within the displayed limits, and its coverage interpretation depends on the statistical procedure and assumptions.

The registry also includes a risk ratio and an odds ratio with confidence limits, but it does not state that either was the designated primary measure. These alternative scales can be useful because relative and absolute comparisons answer different questions. However, neither scale removes uncertainty or resolves the missing analysis details. No adjusted model or covariate-adjusted estimate is reported in the registry record.

How should the p-value be understood?

The posted unadjusted p-value is 0.136. In general, a p-value is calculated under a statistical model that assumes a specified null hypothesis, often no difference between groups. It describes the compatibility of the observed data with that model, considering outcomes at least as extreme according to the test. It is not the probability that the null hypothesis is true, and it does not say whether an effect is clinically meaningful.

The record does not identify the test used to obtain the p-value, its exact null hypothesis, or any covariate adjustment. The planning illustration lists a two-sided significance level of 0.05, but that is an assumption for the illustration and not a documented statement of the trial’s original testing plan. The result should therefore be described as the posted unadjusted p-value, without attributing an unreported method to the trial.

Why does over-dispersion or dependence matter in general?

For a simple binary comparison, standard methods often rely on assumptions about independent observations and the variability of event counts. Over-dispersion means that observed variability is greater than a model expects; ignoring it can make standard errors too small and intervals too narrow. Dependence can have a similar consequence when outcomes within a cluster are correlated. These are general statistical concerns, not findings established for this trial.

The endpoint description identifies sets of twins in both arms. Outcomes from twins in a shared pregnancy may be dependent, so the way the analysis accounts for multiple births could matter. The registry does not report whether the analysis treated pregnancies as clusters, selected one birth per pregnancy, or used another approach. No conclusion about the magnitude or direction of any resulting effect can be drawn from the posted information.

What does randomization contribute?

Random allocation is intended, in general, to make treatment groups comparable on average with respect to both measured and unmeasured characteristics. It does not guarantee exact balance in a particular trial, nor does the word “randomized” establish how allocation was generated or concealed. Those procedural details are not reported in the registry record. The record also says masking was none, so the trial is listed as unmasked; it does not explain who knew each assignment or whether outcome assessment was blinded.

The registry describes a parallel design, meaning the arms are compared as concurrent groups rather than as successive treatment periods for each participant. The posted comparison is unadjusted and based on the primary outcome counts. The record does not report an intention-to-treat definition, per-protocol population, model specification, covariates, or methods for missing outcomes. These omissions restrict a full audit of the analysis.

What can recalculation establish?

A simple unadjusted calculation from posted summary counts can describe the observed group contrast, but it cannot reproduce a trial’s full analysis unless the underlying methods and data are known. It cannot account for unreported stratification, covariate adjustment, clustering, outcome adjudication, or missing data. In this record, the listed discrepancy indicator is false, but that does not supply the missing protocol details. The posted summary should be treated as the available result, with the boundaries of what it documents kept clear.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate an effect estimate. The primary comparison’s confidence interval extends from a negative risk difference to a positive risk difference. Because zero lies within that interval, the posted data do not distinguish clearly between a reduction, no absolute difference, or an increase in the event proportion. The interval’s breadth also indicates substantial uncertainty about the size of any effect.

A confidence interval is not a range of values that are all equally plausible, and it is not a probability statement about the particular trial effect. In general, the method is designed to have a stated long-run coverage rate when its assumptions hold. The confidence level and precise interval method for the trial’s own analysis are not reported in the registry record. Accordingly, the displayed interval should be interpreted as the interval accompanying the posted comparison, not as evidence that every assumption behind its construction has been confirmed.

The event counts are finite, and uncertainty is expected when comparing proportions. General statistical reasoning says that intervals tend to be narrower when estimates are based on more information, but the actual precision also depends on event frequency, design, and analysis choices. The registry gives an overall enrollment of 482 and primary-outcome arm totals of 213 and 195; it does not explain the difference between those totals and enrollment. The extent to which missing or unavailable outcomes affect precision cannot be determined from the record.

The twin-set note raises a further general consideration: if outcomes within a pregnancy are related, treating all observations as independent may affect standard errors and interval width. Since the statistical handling is not reported, the effect of that issue on precision remains unknown. The most defensible reading is therefore that the posted estimate has uncertainty across both sides of no difference, with further uncertainty about unreported aspects of data handling and analysis.

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 effect, not the trial’s original design assumptions. They are expressed as participants per arm for each specified planning scenario. The assumptions listed for the illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario uses a stated fraction of the observed effect as the effect to detect. The trial’s own original sample-size justification is not reported in the registry record.

The 80% power column describes the number of participants per arm that the illustration estimates would be needed to have an 80% chance of detecting the selected effect under its assumptions. The 90% power column uses a higher target chance of detection and therefore gives a larger planning requirement for the same effect scenario. These are design calculations, not descriptions of what happened in the completed trial.

The rows correspond to planning for 100% of the observed effect, planning for 75% of the observed effect, and planning for 50% of the observed effect. The associated figures are 713 and 953 participants per arm for the full observed effect; 1,311 and 1,754 participants per arm for 75% of the observed effect; and 3,045 and 4,075 participants per arm for 50% of the observed effect, with the first number in each pair corresponding to 80% power and the second to 90% power. Smaller effects require more participants to detect under otherwise unchanged assumptions.

The curve is another way to read the same planning relationship across effect sizes: as the target effect becomes smaller, the required sample size rises. The figures are not observed power and do not show the probability that the completed trial found a true effect. They are illustrations conditional on the stated assumptions, and they should not be mistaken for the trial’s enrollment target or for evidence about the trial’s design process.

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 = 0.188; treatment proportion p₁ = 0.133; average p̄ = 0.161.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5192 = 1.0176.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5178 = 0.4358.
4. n = (1.0176 + 0.4358)² / (-0.0545)² = 712.2, rounded up to 713 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 <- 0.187793427230047                          # control proportion
p1 <- 0.133333333333333                          # 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(1344616)
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 = 713 (80% power), 953 (90% power)
 75% of observed effect: n per arm = 1,311 (80% power), 1,754 (90% power)
 50% of observed effect: n per arm = 3,045 (80% power), 4,075 (90% power)
power.prop.test: 712.2 
Simulated power at n = 713 per arm: 0.804 (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 secondary outcome, Gestational Weight, and states that there are two outcome measures. It also lists zero analyses. That count does not establish how many statistical tests were performed, whether any were planned in advance, or whether analyses were conducted but not posted. The analysis details are not reported in the registry record.

When a study examines multiple outcomes, time points, subgroups, or statistical models, the chance of at least one apparently unusual result can increase. This is a general multiplicity concern. Methods such as adjusting significance thresholds or designating one primary analysis can address different aspects of multiplicity, but no such method or plan is reported for this trial. The posted p-value should therefore be read as an unadjusted comparison, not as evidence that a multiplicity strategy was applied.

Missing data are also insufficiently described. The registry reports enrollment of 482 but gives primary-outcome arm totals of 213 and 195, without a participant-flow account or reasons for the difference. It does not state how many participants had missing birthweight information, whether missingness differed by arm, or how missing outcomes were handled. No imputation, complete-case method, or sensitivity analysis is reported. Without these details, the potential impact of unavailable outcomes cannot be assessed.

The analysis population is not defined. The record does not say whether the comparison followed an intention-to-treat principle, included only participants with available outcomes, or used another population. Randomization does not by itself determine how participants with missing or incomplete data enter a specific analysis. The twin-set note also leaves unclear whether the unit of analysis was the participant, infant, or pregnancy for every part of the analysis.

These gaps do not establish that the analysis was inappropriate; they limit what can be verified from the registry record. A fuller statistical analysis plan and participant-flow description would be needed to assess multiplicity, exclusions, missingness, and the precise analysis population.

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.

The serious-adverse-event figures are reported separately from the primary birthweight outcome. For usual clinical care, 244 participants are listed as at risk and zero as affected by a serious adverse event. For lifestyle counseling, 233 are listed as at risk and zero as affected. The record lists no serious-event terms among the top terms.

These figures describe what is posted in the registry; they do not establish that no serious events occurred outside the recorded categories or that every participant was observed for the same duration. The at-risk counts differ from the primary-outcome arm counts, and the registry does not explain the difference. It also does not provide a detailed adverse-event collection method, follow-up window, or event definitions beyond the serious-event category.

Other-event fields show zero at risk and zero affected in both arms, so they do not provide a basis for a meaningful comparison of non-serious events. Death counts are not reported in the registry record. Zero recorded serious events should therefore be read as a description of the posted counts, not as proof of absence of risk or as a complete safety assessment.

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

The result describes a randomized comparison of lifestyle counseling and usual clinical care on one defined endpoint: participant counts for birthweight below 2500 grams. The event counts were lower in the lifestyle-counseling arm in the posted data, and the risk-difference estimate points in that direction. However, the interval spans no difference and extends to effects in the opposite direction. The unadjusted p-value is 0.136. Together, these details describe an uncertain estimate rather than a precise demonstration of a difference.

The endpoint is not a direct count of preterm births, despite the trial title and the brief summary’s discussion of preterm birth. It is a birthweight threshold outcome. A finding about this endpoint should not be substituted for a finding about gestational age at delivery, all birthweights, or other maternal and infant outcomes. The registry lists Gestational Weight as a secondary outcome, but does not provide a detailed result for it in the registry entries.

The result also does not establish why the event counts differ, whether the difference is attributable to the intervention, or whether any observed contrast would apply to populations beyond those represented in this trial. The registry’s outcome totals do not cover the full enrollment count, and the reason is not reported. Missing-outcome handling, analysis populations, and the treatment of twin sets are also not described.

In general, randomized allocation can support causal comparisons when assignment and follow-up are conducted and analyzed appropriately. The registry identifies the design as randomized, but does not report enough detail to evaluate allocation procedures or all analysis assumptions. Thus, the posted result can be described, but its causal interpretation depends on aspects of trial conduct and analysis that are not documented in the record.

The planning calculations do not change this interpretation. They are illustrations based on the observed effect and stated assumptions, not a reconstruction of the trial’s intended enrollment or an estimate of observed power. The result is best read as a comparison with an observed direction but considerable uncertainty and incomplete reporting about outcome availability and analysis methods.

16. Why This Trial Matters Statistically

This trial is statistically useful as an example of why the direction of an estimate, its precision, and its p-value should be read together. The observed event counts are lower in lifestyle counseling, but the confidence interval crosses no difference and the unadjusted p-value is 0.136. Reporting only the direction would omit uncertainty; reporting only the p-value would omit the estimated direction and the range of effects compatible with the interval.

It also illustrates how the endpoint definition shapes interpretation. The title emphasizes preterm births, while the primary posted endpoint is birthweight below 2500 grams. These outcomes may be relevant to a shared clinical question, but they are not the same measurement. Clear distinction between a study’s motivation and its registered endpoint is important when translating results into a statistical claim.

Another statistical lesson concerns denominators. The registry gives overall enrollment of 482, primary-outcome totals of 213 and 195, and serious-event at-risk totals of 244 and 233. Those figures refer to different reported quantities, and the record does not explain the participant flow behind them. Readers should not assume that a denominator from one outcome applies to another or invent a reason for differing totals.

The multiple-birth note is also relevant to design and analysis. Twins from the same pregnancy can produce correlated observations in general, making the unit of analysis and handling of dependence important. The registry records the presence of twin sets but does not describe the statistical response to that feature.

Finally, planning illustrations based on the observed effect can help show how sample-size needs depend on the effect targeted and desired power, but they are not the trial’s original design record. Keeping planning calculations separate from observed results prevents a common misreading of power and sample-size displays.

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 limitations arise from what is not documented. The registry identifies the study as randomized and unmasked, but allocation generation, concealment, and the roles of masking are not detailed. The record therefore does not support a complete assessment of protection against selection or measurement bias.

The primary-outcome totals are 213 for usual clinical care and 195 for lifestyle counseling, while overall enrollment is 482. The reason for this difference is not reported in the registry record. There is no participant-flow explanation, missing-outcome count, or description of exclusions. As a result, the extent and possible consequences of incomplete endpoint data cannot be determined.

The outcome description notes one set of twins in lifestyle counseling and four sets in usual clinical care. The registry does not explain whether the unit of analysis was the pregnancy, participant, or infant, or whether any correlation within twin sets was addressed. This limits assessment of the independence assumptions and of the uncertainty attached to the comparison.

The analysis plan is not described. The record does not report the statistical test, model, adjustment variables, analysis population, or method for handling missing data. It also does not report the original sample-size justification, interim analyses, stopping rules, or multiplicity strategy. The planning illustration is not a substitute for these trial-specific details.

The primary endpoint is a birthweight threshold, while the title focuses on preterm births. The registry does not provide a primary result for preterm delivery itself in the registry record. The result therefore answers a narrower question than the title might suggest. These reporting gaps are limitations on interpretation; they do not, by themselves, demonstrate that trial conduct or analysis was flawed.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Reducing Preterm Births in Underserved Pregnant Women — NCT01344616.
• Registry record last updated 2014-11-24; status last verified 2014-11.
• Results first posted 2014-11-24.
• 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

NCT01344616 is a completed phase 2 randomized, parallel trial with no masking, titled “Reducing Preterm Births in Underserved Pregnant Women.” The registry lists enrollment of 482, a start of 2010-08, primary completion in 2013-07, and PHCC LP as sponsor.

The primary endpoint is participants with birthweight below 2500 grams during 9 months. The posted counts are 40 of 213 in usual clinical care and 26 of 195 in lifestyle counseling. The unadjusted risk-difference estimate favors a lower observed event proportion in the counseling arm, but its interval crosses no difference; the unadjusted p-value is 0.136.

Gestational Weight is listed as a secondary outcome. The record does not explain why primary-outcome totals differ from overall enrollment, how missing outcomes or twin sets were handled, or what statistical methods were prespecified. The planning sample-size figures are illustrations based on the observed effect, not the trial’s original design assumptions.