Nurse-Family Partnership Impact Evaluation in South Carolina: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Nurse-Family Partnership Impact Evaluation in South Carolina (NCT03360539), a randomized na study. The primary outcome is Number of Participants With a Composite Birth Outcome.

2. Clinical Question

This is an evaluation of assignment to a home-visiting program against access to usual care and community services, with outcomes obtained from administrative records. The trial was randomized, but the registry record describes the primary population as narrower than the randomized sample. Its primary question concerns the composite outcome, not any one of its separate components. It also lists outcomes related to birth size, birth weight, preterm birth, and gestational age, among other measures. A result for the composite should therefore be read as evidence about the combined definition, not as a direct answer about every component or every other outcome.

3. Trial Design

The registry record describes an individual-level, randomized, parallel-group trial in South Carolina. Masking is listed as none, and phase is listed as not applicable. The study is sponsored by Harvard School of Public Health (HSPH). Its recorded start date is 2016-04-01, its primary completion date is 2023-04-01, and its status is active, not recruiting.

The registry lists enrollment of 5670. Its brief description gives a planned allocation of 4000 low-income, first-time mothers and their children to the Nurse-Family Partnership intervention group and 2000 to control. These are planning figures in the brief description; the primary-outcome results use the analyzed group sizes reported for that outcome. The record does not explain the difference between those planned counts and the enrollment figure, so the reason is not reported in the registry record.

The intervention is a behavioral program. Nurses begin home visits as early in pregnancy as possible, supporting the mother-to-be in making informed choices; the registry description says visits continue regularly but does not give complete details of their duration or schedule. Control participants have access to standard care and other community programs and services. Outcomes are evaluated using administrative records.

The primary result is based on the number of participants with a composite birth outcome. The registry lists 60 outcome measures and 48 analyses, and 57 secondary outcomes are posted. The primary analysis population excludes participants who withdrew, experienced pregnancy loss, or could not be matched to an index pregnancy. The number excluded for each reason and the methods used to address missing or unmatched records are not reported in the registry record. The statistical methods for the sponsor-reported analyses are also not specified.

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: Differs from randomized sample as it excludes participants who withdrew, who experienced a pregnancy loss, and who could not be matched to an index pregnancy.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: The total number of participants in the baseline analysis includes randomized participants who were successfully matched to an index birth or fetal death and excludes those who were not. This sample also excludes participants who withdrew from the study, experienced a pregnancy loss/miscarriage, or who did not receive the treatment as assigned.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With a Composite Birth Outcome, measured over 0-7 days after index birth; vital records, mortality records, and fetal death records.

The primary endpoint counts participants with at least one event in a composite of adverse birth outcomes. A participant is counted if the index birth meets at least one of these criteria: small for gestational age, low birth weight under 2500 grams, preterm birth under 37 weeks' gestation by obstetric estimate, or perinatal mortality. Perinatal mortality is defined as fetal death at or after 20 weeks of gestation or mortality in the first 7 days of life. The outcome is assessed from 0-7 days after index birth using vital records, mortality records, and fetal death records.

A composite endpoint can capture several related outcomes in a single summary measure. In general, this can be useful when the research question concerns whether any of several specified events occurred, and when a single measure is needed for a group comparison. It also means that the summary does not say which component accounts for the observed count or whether the components follow the same pattern. The composite result should not be treated as though it were a separate estimate for each component.

The endpoint is recorded as a count of participants with an event, while the comparison expresses the difference in the observed proportions between groups. The registry specifies that the analyzed population excludes participants who withdrew, experienced pregnancy loss, or could not be matched to an index pregnancy. Thus, the reported comparison describes the eligible, matched primary-outcome population rather than every person randomized. Reasons for those exclusions and the number affected are not reported in the registry record.

7. Results Posted in the Registry

• Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership: 26.9% (892/3319) versus 26.1% (430/1647) on control; risk difference 0.8 percentage points (95% CI -1.8 to 3.4), risk ratio 1.03 (0.93–1.14); unadjusted two-proportion z-test p = 0.565.

8. Secondary Efficacy Endpoints

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

Number of Participants With an Infant Born Small for Gestational Age (SGA)

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership: 16.8% (553/3295) versus 16.2% (265/1637) on control; risk difference 0.6 percentage points (95% CI -1.6 to 2.8), risk ratio 1.04 (0.91–1.19); unadjusted two-proportion z-test p = 0.597.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With an Infant Born Large for Gestational Age (LGA)

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership: 5.0% (166/3295) versus 5.8% (95/1637) on control; risk difference -0.8 percentage points (95% CI -2.1 to 0.6), risk ratio 0.87 (0.68–1.11); unadjusted two-proportion z-test p = 0.258.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With an Infant Born With Low Birth Weight (LBW)

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership: 13.2% (436/3295) versus 13.4% (219/1637) on control; risk difference -0.1 percentage points (95% CI -2.2 to 1.9), risk ratio 0.99 (0.85–1.15); unadjusted two-proportion z-test p = 0.887.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Treatment - Participation in Nurse Family Partnership Nurse-Family Partnership and the control. The interval gives the range of effects that remain compatible with the data.

The registry posts 57 secondary outcomes, including measures of birth size, birth weight, preterm birth, and gestational age. Numeric comparisons are recorded for several birth outcomes. Across the secondary comparisons with numeric results, the estimated directions are not uniform, and the reported confidence intervals include the null value. This broad pattern does not show a consistent directional separation across the measures for which these results are posted.

These analyses are related to the primary composite but do not duplicate it. The composite counts participants with at least one of several specified events, while individual secondary outcomes focus on particular birth characteristics or events. They can help describe the broader set of outcomes being evaluated, but they cannot by themselves explain the primary composite result or identify its contributing components. Each outcome can also have its own analyzed sample, definition, and precision.

General explanation: when many outcomes are examined, some apparently notable results can arise by chance even if there are no underlying differences. Statistical procedures that account for multiple comparisons can help control this issue, but the registry record does not report whether a multiplicity adjustment or a hierarchy for interpreting secondary outcomes was used. The secondary results should consequently be interpreted with more caution than a single primary comparison, particularly when considering isolated findings. Their unadjusted status should not be mistaken for evidence that every secondary outcome has been tested under a prespecified family-wise plan.

The registry lists 60 outcome measures and 48 analyses, but it does not describe how those analyses were selected or how the family of outcomes was handled. The posted secondary pattern is therefore descriptive rather than a substitute for a documented multiplicity strategy. No claim about unreported secondary results or results beyond the numeric comparisons in the record can be made.

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

• Number of Participants With an Infant Born Small for Gestational Age (SGA)
• Number of Participants With an Infant Born Large for Gestational Age (LGA)
• Number of Participants With an Infant Born With Low Birth Weight (LBW)
• Number of Participants With an Infant With Very Low Birth Weight
• Birth Weight (Continuous)
• Number of Participants With a Preterm Birth
• Number of Participants With an Extremely Preterm Birth
• Gestational Age at Birth

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary at the participant level: each participant either meets at least one component of the composite or does not. Although the registry describes the outcome as a number of participants, the comparison can be expressed as a risk difference between the intervention and control groups. The posted recalculation gives a difference in proportions, with the control proportion listed as 0.261 and the intervention proportion as 0.269. The estimated risk difference is 0.008. These quantities describe the observed event frequency in the analyzed groups; they are not adjusted for other variables.

What does a risk difference express?

General explanation: a risk difference subtracts the event proportion in one group from the event proportion in the other. A value of zero corresponds to equal observed proportions. Positive and negative values indicate different directions according to the subtraction order. Here, the reported comparison is for the intervention against control, and the recalculated estimate is positive. The size of the estimate is expressed on the proportion scale, rather than as a relative comparison. It does not, on its own, establish a causal explanation for any difference.

How should the confidence interval be read?

General explanation: a confidence interval describes a range of values that are reasonably compatible with the data and the statistical procedure used to calculate it. A narrower interval indicates greater precision, while a wider interval indicates greater uncertainty. The primary recalculated 95% confidence interval is -0.018 to 0.034. Because it includes zero, it is compatible with no difference as well as differences in either direction. The interval is not a range containing a known fixed proportion of individual participants, and it does not assign a probability to a particular treatment effect after the data have been observed.

What does the p-value add?

General explanation: a p-value measures how unusual data at least as incompatible with a null hypothesis would be if that null hypothesis and the test assumptions were correct. The recalculated unadjusted p-value is 0.565. It is not a measure of the probability that the intervention has no effect, and it does not tell readers whether a difference would be important in practice. It is best interpreted with the estimate and confidence interval, which give the direction, scale, and precision of the result.

Why are there other effect measures?

The recalculated comparison also lists a risk ratio of 1.029, with a 95% confidence interval from 0.933 to 1.136, and an odds ratio of 1.04, with a 95% confidence interval from 0.910 to 1.189. General explanation: a risk ratio compares the event proportions as a ratio, whereas an odds ratio compares the odds of the event. These measures use different scales from the risk difference, so their numerical values should not be compared as if they were the same quantity. Both intervals include the value corresponding to no relative difference. The registry's unadjusted recalculations are simple comparisons of posted summary statistics; they do not replace the sponsor's own analysis.

The sponsor-reported primary analysis is labeled an adjusted between-group difference. Its estimate is 0.5, and its 95% confidence interval is -2.1 to 3.1. The method, adjustment variables, and unit are not specified in the registry record, and no p-value is reported for this analysis. Without those details, readers cannot determine precisely how the adjusted statistic was modeled or whether its scale matches any of the recalculated measures. The adjusted and unadjusted estimates answer related but not necessarily identical statistical questions.

Why can the analyzed population differ from the randomized groups?

The primary outcome analysis excludes participants who withdrew, experienced pregnancy loss, or could not be matched to an index pregnancy. This means the analyzed sample sizes are 3319 in the intervention group and 1647 in control, rather than a count of everyone enrolled or randomized. General explanation: excluding participants can affect the composition of the groups if the probability of exclusion is related to assignment or outcome. The registry does not report the number excluded, reasons by group, or sensitivity analyses addressing this issue. The effect of the exclusions on the estimate therefore cannot be assessed from the registry record.

For a binary outcome, the observed number of events and the number analyzed in each group contribute to the estimate and its uncertainty. General explanation: more information generally leads to greater precision when other conditions are similar, while a limited event count or a restricted analysis population can leave a wider range of effects compatible with the data. The primary confidence interval remains the direct summary of precision for this posted comparison; it should not be replaced by a judgment based only on the number of participants.

The registry identifies 60 outcome measures and 48 analyses, but does not describe the statistical analysis plan, the model used for the adjusted results, the handling of multiple outcomes, or any interim analyses. Those procedures are not reported in the registry record. General statistical principles can explain why these details matter, but their absence means readers should not infer which procedures were used. The sample-size illustrations are separate planning calculations based on observed effects and stated assumptions; they do not reveal the trial's original design justification.

11. Confidence Intervals and Precision

The primary unadjusted risk-difference estimate is 0.008, with a 95% confidence interval from -0.018 to 0.034. This interval crosses zero, so the data are compatible with no difference. It also includes values on both sides of zero, indicating that the estimate does not precisely pin down the direction of the group difference. The point estimate alone is not a sufficient description: the uncertainty interval is essential for understanding how much the posted data constrain plausible values.

General explanation: precision concerns the spread of uncertainty around an estimate, not whether the estimated effect is important or whether a result is valid for every population. A confidence interval is influenced by the amount of information in the analyzed data and by the statistical method used. In this trial, the primary outcome analysis uses 3319 intervention participants and 1647 control participants after exclusions specified in the endpoint population. The reasons and counts for exclusion are not reported, so their contribution to uncertainty or possible selection effects cannot be determined.

The recalculated 95% confidence intervals for the relative measures also include their respective null values: the risk ratio estimate is 1.029 with interval 0.933 to 1.136, and the odds ratio estimate is 1.04 with interval 0.910 to 1.189. These are different scales, but each interval indicates uncertainty that includes no relative difference. The unadjusted p-value of 0.565 is consistent with the interval crossing the null value, though the p-value and interval convey different aspects of the analysis.

The sponsor-reported adjusted estimate has a 95% confidence interval from -2.1 to 3.1 around an estimate of 0.5. It too includes zero. Its unit and method are not reported in the registry record, so the numerical width should not be compared directly with the risk-difference interval. In general, an interval that includes zero does not prove exact equivalence or rule out all meaningful differences; it indicates that the estimate remains uncertain over the interval's range. The registry does not specify an equivalence margin or a threshold for a clinically important difference.

12. Planning a Similar Trial: Sample Size

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

The planning figures are illustrations based on the observed effect, not the trial's original design calculations. For these binary outcomes, the listed quantities are participants per arm, not total events. The assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The target effect is the stated fraction of the observed effect. The trial's own sample-size assumptions and justification are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm that the stated calculation associates with an 80% chance of detecting the assumed effect under its planning assumptions, if that effect and assumptions were correct. The 90% power column gives the corresponding illustration for a higher chance of detecting that effect. These are design-planning quantities, not observed results or guarantees. They are not observed power.

Rows that plan for 100% of the observed effect use the full estimated effect as the target; rows planning for 75% or 50% use smaller target effects. Smaller effects are harder to distinguish from random variation and therefore require more participants under otherwise comparable assumptions. This is why the illustrative participant requirements increase as the target fraction decreases. The calculations are especially sensitive to the observed effect used as their starting point, so they should be read as scenarios rather than forecasts of what a future study would achieve.

The posted planning values also should not be confused with the enrollment recorded for this trial or with the analyzed sample sizes for the primary outcome. The illustration uses equal allocation, while the study description gives unequal planned group counts. The figures are not a retrospective assessment of whether the trial had adequate power, and they do not tell readers whether the original study met its design goals. Such original planning details are not reported in the registry record.

13. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Control proportion p₀ = 0.261; treatment proportion p₁ = 0.269; average p̄ = 0.265.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6241 = 1.2232.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6241 = 0.5252.
4. n = (1.2232 + 0.5252)² / (0.0077)² = 51,895.6, rounded up to 51,896 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 <- 0.261080752884032                          # control proportion
p1 <- 0.268755649291955                          # 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(3360539)
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 = 51,896 (80% power), 69,473 (90% power)
 75% of observed effect: n per arm = 92,046 (80% power), 123,223 (90% power)
 50% of observed effect: n per arm = 206,621 (80% power), 276,606 (90% power)
power.prop.test: 51895.6 
Simulated power at n = 51,896 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.

14. Analyses Reported by the Sponsor

The registry reports a sponsor analysis labeled as an adjusted between-group difference. For the primary endpoint, its estimate is 0.5, with a 95% confidence interval from -2.1 to 3.1. The registry does not list a p-value for this result. It also does not identify the model, adjustment variables, or unit for the estimate, which limits direct interpretation of its scale.

The unadjusted recalculation uses the posted event counts and analyzed group sizes. It expresses the comparison as a risk difference of 0.008, with a 95% confidence interval from -0.018 to 0.034 and an unadjusted p-value of 0.565. This is not the same statistical quantity as an adjusted difference unless the underlying scale and method match, and the record does not establish that they do. The two estimates should therefore be described separately rather than treated as interchangeable or as competing calculations of precisely the same target.

Both reported intervals include zero, but their numerical scales are not documented as equivalent. The adjusted result is the sponsor's reported analysis; the recalculation is a simple, unadjusted analysis of the posted summary statistics. The registry record does not provide enough methodological detail to reproduce the adjusted result or explain how its estimate relates to the unadjusted risk difference. No conclusion about the model's adequacy or the role of covariate adjustment can be drawn from the information recorded.

15. Multiplicity, Missing Data and Analysis Populations

The registry record describes a primary composite endpoint and a broad set of secondary measures, with 60 outcome measures and 48 analyses listed. It does not report a multiplicity strategy, such as an adjustment for testing multiple outcomes or an order in which the secondary results should be interpreted. General explanation: examining many outcomes increases the opportunity for chance findings, so the context of a result among the full set of analyses matters. Without a reported plan, isolated secondary results warrant cautious interpretation.

The record states that the primary-outcome population excludes participants who withdrew, experienced a pregnancy loss, or could not be matched to an index pregnancy. It does not give the number excluded for each reason or by randomized group. It also does not describe whether missing records were imputed, whether unmatched records were handled in another way, or whether sensitivity analyses were conducted. These missing details limit assessment of how the analysis population may differ from the randomized sample.

The registry identifies the sponsor's primary analysis as an adjusted between-group difference but does not report the method, covariates, or model assumptions. The methods for the reported secondary analyses are also not specified. The registry does not report the trial's original sample-size justification, its statistical analysis plan, or whether interim analyses were performed. These items should be regarded as unreported rather than inferred from the posted estimates.

The primary comparison is based on participants with an outcome record who met the stated population criteria; the record does not say that every randomized participant was included in an intention-to-treat analysis. General explanation: analysis-population definitions matter because they determine who contributes data and what group comparison is being estimated. Randomization supports comparable groups at assignment, but exclusions after assignment can change the composition of the analyzed groups. The consequences cannot be quantified from the registry information alone. These limitations affect how fully the statistical results can be independently assessed; they do not change the numerical estimates recorded.

16. 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.

The registry lists 3319 participants at risk in the intervention group and 1647 in control for adverse-event reporting. Serious adverse events are recorded as affecting 0 participants in each group, and other adverse events are also recorded as affecting 0 participants in each group. The record lists 5 deaths in the intervention group and 2 in control.

These figures should be read as registry counts, not as evidence about cause or attribution. The record does not provide event descriptions, causes of death, timing, or an assessment of whether any event was related to the intervention. It also does not explain the reporting conventions behind the zero affected counts alongside the reported deaths. The death counts should therefore not be assumed to represent deaths caused by the program, nor should the adverse-event listing be used to infer a complete account of safety outcomes.

General explanation: adverse-event counts describe events recorded under a study's reporting framework; interpretation depends on definitions, follow-up, ascertainment, and attribution procedures. Those details are not reported in the registry record. The posted figures are limited to the counts listed and do not support broader claims about comparative safety.

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

The posted primary result estimates the difference in the frequency of a composite birth outcome between the analyzed intervention and control groups. The unadjusted recalculation is close to zero on the risk-difference scale, but its confidence interval includes zero and extends in both directions. This means the estimate is uncertain: the posted data do not precisely establish the direction or size of a group difference. The sponsor's adjusted estimate also has an interval that includes zero, although its scale and method are not described sufficiently for direct comparison with the recalculation.

This result does not establish that the groups are exactly equivalent. A confidence interval that includes zero is not proof that the true difference is zero, and the registry does not state an equivalence margin or a threshold for a practically important difference. Nor does the result establish that every component of the composite has the same pattern. The composite combines small for gestational age, low birth weight, preterm birth, and perinatal mortality; separate outcome comparisons are needed to describe those components.

The result also does not answer every question about the Nurse-Family Partnership. The record concerns this randomized evaluation in South Carolina, against a control condition that included standard care and community services, and analyzes a population excluding withdrawals, pregnancy losses, and pregnancies that could not be matched. Findings for that population should not automatically be generalized to other settings, populations, service environments, or program implementations.

Finally, the posted analyses do not establish a mechanism for any observed difference, and the registry record does not contain enough detail to assess all aspects of analysis conduct. It does not report the sponsor's statistical method, adjustment variables, missing-data procedures, multiplicity strategy, or original design assumptions. The appropriate reading is a measured comparison with substantial uncertainty and documented population limits, not a categorical judgment about the program's effects.

18. Why This Trial Matters Statistically

This trial is statistically notable for evaluating an established home-visiting program through individual-level randomization and administrative records in a modern South Carolina setting. The control group retained access to standard care and community services, making the contrast one of program participation versus the services otherwise available rather than intervention versus no care. That distinction is central to defining the comparison that the estimates address.

The primary endpoint combines several related birth outcomes into one participant-level measure. This creates a clear overall question about whether at least one specified event occurred, while also making component-specific interpretation important. A composite can be informative as a summary, but it can conceal differing patterns among its parts. The registry's additional birth outcomes provide related measures, although the breadth of posted outcomes also makes transparent multiplicity methods important.

The analysis illustrates why effect estimates and uncertainty should be read together. The recalculated risk difference is 0.008, but its 95% confidence interval includes zero and spans both directions. The result is not meaningfully summarized by a p-value alone. Its interpretation also depends on the analyzed population, which excludes withdrawals, pregnancy losses, and unmatched index pregnancies.

The record is informative about the trial's design and outcome definitions, but leaves important statistical details unreported. In particular, the adjusted-analysis method, covariates, missing-data procedures, multiplicity approach, and original sample-size justification are not available in the registry record. Those gaps constrain independent assessment of the analysis without erasing the value of the randomized comparison. The study offers a useful example of how design, endpoint construction, analysis populations, and precision jointly shape the meaning of clinical-trial results.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary endpoint population is not identical to the randomized sample. Participants who withdrew, experienced pregnancy loss, or could not be matched to an index pregnancy are excluded. The registry record does not give the number excluded by reason or group, nor does it report sensitivity analyses for those exclusions. This makes it difficult to assess possible selection effects or how the findings might differ under another analysis population.

The sponsor-reported adjusted difference is not accompanied by its statistical method, adjustment variables, or unit. As a result, it cannot be fully reconstructed or placed on the same scale as the unadjusted risk difference. The unadjusted recalculations use posted summary counts and do not account for covariates. Neither set of estimates should be treated as a complete description of the other.

The registry lists many outcome measures and analyses but does not report a multiplicity plan. It also does not report the original sample-size justification, the full statistical analysis plan, or whether interim analyses were conducted. The sample-size illustrations based on observed effects are not a substitute for those design details.

The intervention description is incomplete, and the registry record does not provide full information about visit frequency or program duration. The trial is unmasked, and the control condition includes usual care and available community services; variation in those services is not described. The record's adverse-event listing contains limited detail about event definitions, causes, timing, and attribution. Finally, the available primary result concerns a composite outcome and does not, on its own, characterize each component or all 57 posted secondary outcomes. These limitations define the scope of interpretation, rather than invalidate the reported numbers.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Nurse-Family Partnership Impact Evaluation in South Carolina — NCT03360539.
• Registry record last updated 2025-12-12; status last verified 2025-12.
• Results first posted 2025-12-12.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

The registry record describes NCT03360539, a randomized, parallel-group evaluation of Nurse-Family Partnership in South Carolina. It lists no masking, phase not applicable, enrollment of 5670, and active-not-recruiting status. The intervention is a nurse-led prenatal and infancy home-visiting program; control participants have access to standard care and community services.

The primary endpoint is the number of participants with at least one component of a composite birth outcome, assessed from 0-7 days after index birth. Its analysis excludes withdrawals, pregnancy losses, and pregnancies that could not be matched. The unadjusted recalculated risk difference is 0.008, with a 95% confidence interval from -0.018 to 0.034 and p = 0.565. The sponsor-reported adjusted difference is 0.5, with a 95% confidence interval from -2.1 to 3.1; its method and unit are not reported in the registry record.

The record posts numerous secondary outcomes but does not report a multiplicity strategy, full missing-data methods, or the original sample-size justification. The primary estimate is therefore best read alongside its uncertainty, population exclusions, and the limits of the statistical detail recorded.