Connect-Home Clinical Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Connect-Home Clinical Trial (NCT03810534), a randomized na study. The primary outcome is Care Transitions Measure-15 Score 7 Days After Skilled Nursing Facility Discharge.

2. Clinical Question

Connect-Home examined transitions from skilled nursing facilities to home for patients described in the registry as seriously ill, and for their caregivers. The central question was whether a behavioral transitional-care intervention addressing the needs of both patients and caregivers was associated with patients reporting greater preparedness for discharge. The registry also describes intended patient outcomes including quality of life, function, and acute care use, as well as caregiver preparedness, burden, and distress.

The primary outcome captures self-reported knowledge and skills for continuing care at home. The registry describes higher scores as being associated with less acute care use after discharge. The trial’s reported primary analysis used a mixed-model method, while the simple recalculation from the posted group summaries gives a different estimate and statistical result. That difference is important when interpreting what the registry record supports: the recalculation is not a substitute for the sponsor’s own analysis.

3. Trial Design

The Connect-Home Clinical Trial is listed as completed and randomized, with a crossover model and single masking. Enrollment was 654. The registry lists the University of North Carolina, Chapel Hill, as sponsor. The study started on 2019-03-01, and primary completion was 2021-07-24. Phase is listed as NA. Further details about how the crossover was implemented, including its periods, sequences, or any washout approach, are not reported in the registry record.

Participants in the intervention group received Connect-Home at the skilled nursing facility and at home. The control group received standard discharge planning at the skilled nursing facility only. The intervention is classified as behavioral. The registry describes the study as testing transitional care for patients and caregivers during the transition from a skilled nursing facility to home; it does not report further operational details about how intervention components were delivered or whether delivery varied by participant.

The primary outcome was the Care Transitions Measure-15 score 7 days after discharge. The registry’s primary outcome population is participants who answered the 7-day data collection phone call. The group summaries include 142 participants in Connect-Home and 141 in control. Those outcome counts are distinct from the overall enrollment figure; reasons for the difference and the flow of participants through each analysis are not reported in the registry record.

The registry lists quality of life, life space, caregiver burden, and distress as secondary outcome areas, measured at 30 or 60 days after discharge. It reports 12 outcome measures and 12 analyses, with 10 secondary outcome results posted. The registry does not describe the sequence of secondary analyses, the model details beyond the named mixed-model method for reported analyses, or the original sample-size justification. These omissions limit how fully the design and analysis choices can be reconstructed.

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.

Standard of Care (5-10 Months)

Pre-Implementation Phase (2 Months)

Connect-Home (5-10 Months)

Analysis population for the primary outcome, as defined in the registry: Data reported for participants who answered the 7-day data collection phone call.

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 Care Transitions Measure-15 Score 7 Days After Skilled Nursing Facility Discharge, measured over 7 Days After SNF Discharge.

The primary endpoint is the Care Transitions Measure-15 score 7 days after skilled nursing facility discharge. The CTM-15 contains 5 items, each on a 4-point scale, and measures the patient’s self-reported knowledge and skills for continuing care at home. The registry reports a summary score range of 0-100 and says higher scores are associated with less acute care use after discharge. This endpoint therefore focuses on patients’ perceptions of their preparedness for the transition, rather than directly counting subsequent hospital visits or measuring clinical function.

Preparedness is relevant to a transitional-care intervention because the intervention is intended to address needs around discharge and continuing care at home. The measure offers a structured way to summarize patients’ responses across several aspects of readiness. Its interpretation should remain tied to what it measures: a self-reported score, collected by phone, among participants who answered the 7-day call. The registry does not report the psychometric properties of the measure in this trial, a threshold for a meaningful change, or whether a particular difference in score was considered important in the study’s protocol.

The posted group means were 75.8 for Connect-Home and 68.33 for control. These summaries describe the participants included in the respective endpoint groups. They do not, by themselves, establish the result of the trial’s prespecified model-based analysis, which the registry reports separately.

7. Results Posted in the Registry

• Connect-Home: mean 75.80 versus 68.33 on control; difference 7.47 (95% CI 3.53 to 11.41), standardized effect d = 0.44; z-test p = <0.001.

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.

McGill Quality of Life Questionnaire-Revised Score 30 Days After Skilled Nursing Facility Discharge

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Connect-Home: mean 7.69 versus 7.69 on control; difference 0.00 (95% CI -0.51 to 0.51), standardized effect d = 0.00; z-test p = 1.000.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Connect-Home and the control. The interval gives the range of effects that remain compatible with the data.

McGill Quality of Life Questionnaire-Revised Score 60 Days After Skilled Nursing Facility Discharge

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Connect-Home: mean 7.81 versus 7.71 on control; difference 0.10 (95% CI -0.40 to 0.60), standardized effect d = 0.06; z-test p = 0.696.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Connect-Home and the control. The interval gives the range of effects that remain compatible with the data.

Life Space Assessment 30 Days After Skilled Nursing Facility Discharge

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Connect-Home: mean 22.18 versus 22.97 on control; difference -0.79 (95% CI -4.77 to 3.19), standardized effect d = -0.05; z-test p = 0.698.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Connect-Home and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the study’s scope beyond discharge preparedness to include patient quality of life and life space, as well as caregiver burden and distress. The available detailed summaries show a mixed pattern of small unadjusted mean differences: the 30-day quality-of-life means are equal, the 60-day quality-of-life estimate is close to zero, and the 30-day life-space estimate is also close to zero relative to its score spread. The recalculated intervals reported for these endpoints include zero. The sponsor’s reported mixed-model p-values for these outcomes are .93, .60, and .53, respectively.

This overall pattern does not provide a clear, consistent secondary-outcome signal that independently resolves the primary discrepancy. It also should not be read as evidence that all secondary outcomes were unchanged. The registry lists multiple patient and caregiver outcomes and reports 10 secondary results, but detailed group summaries for several named outcomes are not reported in the registry record. A complete account of the pattern across every secondary outcome is therefore not possible from the reported detail.

Secondary outcomes help show whether a trial’s findings are consistent across related measures and follow-up times. However, multiple outcomes create multiple opportunities for apparently notable results. In general, when many outcomes are tested, unadjusted p-values carry less weight as confirmatory evidence because the chance of at least one chance finding rises. Whether the trial used a multiplicity adjustment, specified a hierarchy among outcomes, or treated any secondary endpoint as confirmatory is not reported in the registry record.

The secondary results should therefore be interpreted as supportive or exploratory descriptions unless the study’s analysis plan establishes otherwise. They should not be combined selectively with the primary recalculation to override the sponsor’s mixed-model primary result. Nor should the absence of detailed results for some outcomes be treated as evidence of no effect. The available summaries provide limited context around the primary question, while missing analysis-plan and outcome details constrain stronger conclusions.

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

• McGill Quality of Life Questionnaire-Revised Score 30 Days After Skilled Nursing Facility Discharge
• McGill Quality of Life Questionnaire-Revised Score 60 Days After Skilled Nursing Facility Discharge
• Life Space Assessment 30 Days After Skilled Nursing Facility Discharge
• Life Space Assessment 60 Days After Skilled Nursing Facility Discharge
• Zarit Caregiver Burden Scale 30 Days After Skilled Nursing Facility Discharge
• Zarit Caregiver Burden Scale 60 Days After Skilled Nursing Facility Discharge
• Distress Thermometer 30 Days After Skilled Nursing Facility Discharge
• Distress Thermometer 60 Days After Skilled Nursing Facility Discharge

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: each participant contributes a numerical CTM-15 score, and the registry reports group means and standard deviations. A mean describes the group’s average score. A standard deviation describes how spread out individual scores are around that group average. These summaries are useful for describing the observed endpoint groups, but they do not account for every design feature that may have been included in the trial’s analysis.

What does the recalculation represent?

The recalculated mean difference is the Connect-Home group mean minus the control group mean, using the posted group summaries. For the primary outcome, this unadjusted calculation is 7.47. Its confidence interval runs from 3.53 to 11.41, and its unadjusted p-value is p < 0.001. This approach treats the two group summaries as the basis for a direct comparison. It does not reproduce a mixed model, adjust for covariates, or account for unreported features of the crossover design. It should be read as a simple comparison of the available summaries, not as the trial’s definitive analysis.

In general, an unadjusted comparison is useful for checking how the reported averages relate to each other. It is not necessarily equivalent to a prespecified analysis that models repeated observations, clustering, baseline measures, or other design features. Which features were included in the sponsor’s model is not reported in the registry record. The registry does identify the sponsor’s method as Mixed Models Analysis and its parameter as Mean Difference (Final Values), but it does not give the model equation, covariates, correlation structure, or handling of incomplete observations.

What does the mixed-model result add?

A mixed model is a general statistical approach that can represent outcomes with dependencies or multiple sources of variation, rather than assuming every observation is unrelated to every other observation. The precise structure matters: different model specifications can produce different estimates and uncertainty. Here, the sponsor’s reported primary estimate is -1.64, with p = .62. Since the registry does not report the details needed to reconstruct that analysis, the result cannot be independently matched to the simple recalculation from the group means. The discrepancy flag means the sponsor’s analysis takes precedence in interpreting the trial’s reported result.

In general, a p-value summarizes how compatible the observed data are with a specified null hypothesis, assuming the chosen model and its assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. A small p-value from the simple recalculation should not override the trial’s own model-based result when the two conflict. The reported p = .62 is not proof of no difference; it is a result that does not establish a difference under the sponsor’s analysis.

What do the effect estimates and standardized values mean?

The primary recalculation also reports a standardized effect, d = 0.442. In general, standardization expresses a mean difference relative to the spread of scores, which can help describe magnitude on a scale independent of the original units. It does not solve the discrepancy between the simple calculation and the mixed-model analysis. The posted score difference remains interpretable in CTM-15 points, while the standardized statistic is a supplementary description of the unadjusted group summaries.

Standard errors are also reported for the group means: 1.49 for control and 1.349 for Connect-Home. In general, a standard error describes the estimated sampling variability of a statistic, such as a mean, and is used in constructing inferential summaries. A standard error is not the same as the standard deviation: the standard deviation concerns variation among participant scores, while the standard error concerns uncertainty in an estimated mean. Neither quantity, in isolation, describes all uncertainty in the sponsor’s mixed-model estimate.

Why does the trial model matter?

The registry labels the design as crossover, but further details about the crossover structure and its analytic treatment are not reported in the registry record. In general, a crossover design can involve observations linked to the same participant or ordered treatment periods, so the analysis may need to reflect those dependencies and order. That general consideration cannot be used to infer how Connect-Home handled periods, sequence, carryover, or participant-level dependence. No such protocol or analysis specifics should be assumed from the design label alone.

Random allocation is reported, but details of sequence generation and allocation concealment are not. Single masking is recorded, without identifying who was masked. These design labels give useful context but are not a full account of trial conduct. The registry also does not report the original sample-size justification or whether any interim analyses were planned or conducted.

How should multiple outcomes be handled?

The registry lists 12 outcome measures and 12 analyses, and 10 secondary results are posted. In general, conducting multiple tests increases the chance that at least one result may appear notable by chance when no multiplicity adjustment is used. Whether an adjustment was planned or applied is not reported in the registry record. The recalculated secondary p-values are unadjusted, and they should not be treated as independently confirmatory evidence without knowledge of the analysis plan.

For these reasons, the statistical methods support a cautious reading: describe the group summaries and recalculated quantities as such; distinguish them from the sponsor’s mixed-model result; and avoid claiming that unreported model choices, assumptions, or corrections were used. The registry record does not contain enough methodological detail to reproduce the sponsor’s analysis or fully assess its assumptions.

11. Confidence Intervals and Precision

A confidence interval describes a range of estimates compatible with the observed data under the calculation’s assumptions. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty about the effect size. An interval is not a guarantee that the true effect lies inside it, and it does not establish clinical importance. The registry does not report the confidence level for its comparison intervals, so no particular confidence level should be assigned to them.

For the unadjusted primary recalculation, the mean difference is 7.47 and the interval extends from 3.53 to 11.41. That interval is entirely above zero, consistent with the small unadjusted p-value. This apparent precision pertains to the recalculation from the posted group summaries. It does not resolve the conflict with the sponsor’s mixed-model estimate of -1.64, for which the registry reports p = .62 and no confidence interval. The sponsor result therefore cannot be described with interval precision from the information in the registry record.

The secondary recalculations illustrate why the interval matters alongside the point estimate. For quality of life at 60 days, the reported recalculated difference is 0.1, with an interval from -0.401 to 0.601. For life space at 30 days, the difference is -0.79, with an interval from -4.775 to 3.195. These intervals include zero, so the corresponding unadjusted estimates are compatible with effects in either direction under their calculation assumptions. The quality-of-life result at 30 days has a recalculated difference of 0.0, with an interval from -0.511 to 0.511.

These intervals are based on unadjusted comparisons, not the sponsor’s mixed-model analyses. Model-based confidence intervals for the reported analyses are not reported in the registry record. Precision should therefore be assessed separately for the posted summaries and for the sponsor analysis; the latter has no reported interval with which to judge the range of values consistent with that model.

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 a description of the trial’s original design calculation. For the primary outcome, the figures are participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; each scenario sets the effect to detect as the stated fraction of the observed effect. The registry’s original sample-size assumptions and justification are not reported in the registry record.

The 80% power column gives the number of participants per arm that the planning calculation indicates would be needed to have an 80% chance of detecting the specified effect under those assumptions, if that effect were the true one. The 90% power column gives the corresponding per-arm planning number for a higher chance of detection. These are prospective planning quantities, not a statement about the probability that the completed trial found a real effect.

For the primary endpoint, planning for 100% of the observed effect gives 81 participants per arm for 80% power and 108 for 90% power. Planning for 75% of the observed effect gives 143 per arm for 80% power and 192 for 90% power. Planning for 50% of the observed effect gives 322 per arm for 80% power and 431 for 90% power. Smaller effects require more participants because they are harder to distinguish from sampling variation under the same assumptions.

These values are not observed power. They are illustrative calculations tied to the unadjusted observed effect and the stated planning assumptions. They do not establish what power the trial had for the sponsor’s mixed-model analysis, or what effect the trial was originally designed to detect. The planning assumptions do not account for design or analysis details that 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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 7.470; pooled standard deviation σ = 16.901.
2. n = 2 × (1.96 + 0.842)² × 16.901² / 7.470² = 80.4, rounded up to 81 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
delta <- 7.47                      # observed difference in means
sigma <- 16.9013167167995                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3810534)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 81 (80% power), 108 (90% power)
 75% of observed effect: n per arm = 143 (80% power), 192 (90% power)
 50% of observed effect: n per arm = 322 (80% power), 431 (90% power)
power.t.test: 81.3 
Simulated power at n = 81 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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 sponsor’s primary analysis is reported as a Mixed Models Analysis of Mean Difference (Final Values). Its estimate is -1.64, with p = .62. The registry does not report a confidence interval for that result. In contrast, the unadjusted recalculation from the posted group means gives a mean difference of 7.47, an interval from 3.53 to 11.41, and p < 0.001. The recalculation and the sponsor analysis therefore differ in both direction and statistical result.

The recalculation uses only posted group summary statistics. The sponsor reports a mixed-model method, but the model details and analysis population are not described sufficiently to reproduce it or explain the discrepancy. It would be speculative to attribute the difference to any particular adjustment, participant subset, crossover feature, or data-handling decision.

Because the discrepancy flag is true, the sponsor’s own analysis takes precedence. This means the unadjusted recalculation should be presented as a descriptive check on the group summaries, not as the primary inferential conclusion. The registry record supports reporting the sponsor’s estimate and p-value while also noting that the public information is insufficient to reconcile them. It does not support selecting whichever result appears more favorable or treating the recalculation as a correction of the sponsor’s analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 12 outcome measures and 12 analyses, with 10 secondary results posted. This means the evidence is distributed across several endpoints and follow-up times rather than concentrated in a single test. In general, multiple testing can increase the probability of obtaining at least one apparently unusual result by chance. A prespecified outcome hierarchy or multiplicity adjustment can address that concern, but whether either was planned or applied is not reported in the registry record. The secondary recalculations are explicitly unadjusted, so their p-values should not be interpreted as if they account for the full family of outcomes.

Missing-data handling is also important because the primary endpoint population consists of participants who answered the 7-day data collection phone call. The group summaries include 142 Connect-Home participants and 141 control participants, while overall enrollment was 654. The registry does not explain the relationship between these counts, give the number eligible for each assessment, or describe why outcomes were unavailable. It does not report whether missing observations were handled by imputation, complete-case analysis, likelihood-based methods, or another approach. The consequences of missing outcome data therefore cannot be assessed from the record.

Analysis populations are not fully specified. The registry does not report whether the sponsor’s mixed-model analysis used an intention-to-treat population, a per-protocol population, or another defined set. It also does not explain how crossover allocation and participant-level observations were represented in the analysis. Those details matter for understanding the estimate and for judging whether the sponsor analysis and summary groups refer to the same participants.

The registry reports the method name for the sponsor analyses but not the statistical analysis plan, model specification, covariates, or assumptions. Interim analyses, stopping rules, and the trial’s original sample-size justification are not reported in the registry record. These are gaps in the available description, not evidence that a particular method was or was not used. Together with the discrepancy between the primary recalculation and sponsor analysis, they limit independent verification and make the sponsor-reported result the appropriate primary reference.

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.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: hospitalization due to fracture subsequent to fall (4.8%), hospitalization due to pneumonia (3.7%), hospitalization due to urinary tract infection (3.0%), hospitalization due to shortness of breath (2.5%), hospitalization reason undisclosed (1.8%).

The registry reports serious adverse events separately for Connect-Home patients, Connect-Home caregivers, control patients, and control caregivers. Serious events were recorded for 38 of 165 Connect-Home patients and 34 of 162 control patients. Deaths were recorded for 6 Connect-Home patients and 8 control patients. No serious adverse events or deaths were recorded in either caregiver group, each of which had 165 Connect-Home caregivers or 162 control caregivers at risk. Other adverse events were recorded as 0 in all four groups.

These are registry counts for the listed groups, not a complete account of event timing, severity beyond the serious-event classification, participant-level follow-up, or causal attribution. The registry does not report whether events were judged related to the intervention, how long participants were followed for safety, or how differences in time at risk were handled. The at-risk counts do not explain how they relate to the overall enrollment or the primary endpoint groups.

The listed leading serious-event terms include hospitalization due to fracture subsequent to fall at 4.8%, hospitalization due to pneumonia at 3.7%, hospitalization due to urinary tract infection at 3.0%, hospitalization due to shortness of breath at 2.5%, hospitalization reason undisclosed at 1.8%, and hospitalization due to skin infection at 1.8%. These terms describe reported serious events; they do not, by themselves, establish that the intervention caused or prevented any event. Comparisons should be cautious because the registry does not provide a detailed safety analysis or context for event ascertainment.

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

The registry supports a careful description of the trial’s measurements and reported analyses, not a definitive statement about whether Connect-Home changed patient or caregiver outcomes. For the primary endpoint, the unadjusted group summaries yield a positive difference, while the sponsor’s mixed-model analysis reports a negative estimate with p = .62. Because the discrepancy is true, the sponsor’s own analysis takes precedence. The available record does not explain why the estimates differ, and the model-based confidence interval is not reported.

A non-significant sponsor p-value is not proof that the groups were equivalent or that an effect was absent. Establishing equivalence or ruling out effects of meaningful size would require a suitable design and analysis with a prespecified margin, neither of which is reported in the registry record. Likewise, the recalculated p-value below 0.001 does not supersede the sponsor result: it pertains to a simple unadjusted comparison of the posted means and does not reproduce the reported mixed model.

The endpoint measures self-reported preparedness for discharge, not all aspects of successful transition. The registry also names quality of life, function, acute care use, caregiver preparedness, burden, and distress as outcomes of interest, but detailed results for every named outcome are not reported. The findings should not be extended to outcomes or participants not represented in the reported analyses.

The trial record gives useful descriptive information about the intervention, comparator, enrollment, outcome measures, and safety counts. It also leaves important uncertainties about missing observations, analysis populations, model specification, multiplicity, and the crossover analysis. Those limitations affect how confidently readers can interpret the results and compare the recalculation with the sponsor’s analysis. The statistics alone do not establish that the intervention works or does not work, and they do not support medical advice.

18. Why This Trial Matters Statistically

This trial illustrates why a posted difference between group averages is not automatically the same as the trial’s primary inferential result. The primary group summaries lead to an unadjusted recalculation of 7.47, with p < 0.001, while the sponsor’s mixed-model estimate is -1.64, with p = .62. The opposing directions make it especially important to distinguish a summary-statistic calculation from the analysis reported by the study sponsor.

For biostatisticians and clinical researchers, the discrepancy emphasizes the value of transparent reporting of analysis populations, model specifications, covariates, missing-data methods, and design-specific features. The registry labels the design crossover and identifies a mixed-model analysis, but it does not provide the details needed to reproduce that analysis. Without those details, readers cannot determine whether the differences arise from a particular modeling choice or another feature of the analysis. Any such explanation would be speculative.

The record also demonstrates why precision and multiplicity need separate attention. A point estimate does not show the range of uncertainty, and the sponsor’s primary analysis has no reported interval. Several outcomes are listed, while multiplicity handling is not described. Secondary findings therefore need cautious interpretation rather than being treated as a set of independent confirmations.

Finally, planning illustrations based on an observed effect are not a record of the original design rationale. The trial’s original sample-size justification is not reported in the registry record. Keeping planning assumptions distinct from observed results prevents retrospective power figures from being mistaken for evidence about the completed trial’s ability to detect an effect.

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.

Several limitations specific to this record affect interpretation. First, the primary recalculation from the posted group means conflicts with the sponsor’s mixed-model estimate in both direction and statistical result. The sponsor’s analysis takes precedence, but the registry does not report enough model or population detail to explain the discrepancy or reproduce the result. Its primary analysis also has no reported confidence interval, limiting assessment of model-based precision.

Second, the primary endpoint is reported for participants who answered the 7-day phone call. Although the endpoint group summaries list 142 participants in Connect-Home and 141 in control, the registry does not report the eligible denominators, reasons for nonresponse, or missing-data methods. Overall enrollment was 654, but the participant flow into the primary and secondary analyses is not described sufficiently to evaluate selection or missingness.

Third, the crossover model label is not accompanied by details about treatment periods, sequence, within-participant observations, or how the design was represented statistically. Single masking is listed, but who was masked is not reported. The registry does not provide the statistical analysis plan, original sample-size justification, or information about interim analyses.

Fourth, the record lists numerous outcomes and analyses, but multiplicity handling is not reported. Detailed secondary group summaries are available for only some listed outcomes, and complete results for all named patient and caregiver outcomes are not reported in the registry record. The safety counts also lack information about event attribution, follow-up duration, and exposure time. These limitations do not establish that any analysis was inappropriate; they mark areas where the public record does not allow independent assessment.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Connect-Home Clinical Trial — NCT03810534.
• Registry record last updated 2022-08-23; status last verified 2021-12.
• Results first posted 2022-08-23.
• 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

Connect-Home was a completed randomized, single-masked trial with a crossover model and enrollment of 654. It compared a behavioral transitional-care intervention delivered at a skilled nursing facility and at home with standard discharge planning at the facility. The primary endpoint was CTM-15 score 7 days after discharge among participants who answered the data collection phone call.

The sponsor reported a mixed-model mean difference of -1.64 with p = .62. A simple unadjusted recalculation from the posted group means gives a different result: 7.47, with an interval from 3.53 to 11.41 and p < 0.001. The discrepancy means the sponsor’s analysis takes precedence, while the public record does not explain the divergence. Secondary outcomes, missing-data handling, multiplicity, model details, and analysis populations are incompletely described, limiting independent interpretation.