Telehealth Management in HF Disparity Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Telehealth Management in HF Disparity Patients (NCT02196922), a randomized na study. The primary outcome is Hospitalizations.

2. Clinical Question

3. Trial Design

The registry record describes a randomized, parallel-group study with no masking. It lists the study as completed and gives enrollment as 104. The record names Northwell Health as sponsor and identifies chronic heart failure as the condition. The listed phase is NA. The start date is 2014-01-01 and primary completion is 2017-04-30.

Participants were assigned to either Telehealth Self Management (TSM) or Comprehensive Outpatient Management (COM). TSM was described as a weekly clinical telehealth visit combined with daily vital-sign self-monitoring. A monitor at the participant's residence connected to a provider station through a standard telephone line. COM was described as standard care at a Heart Failure clinic, including primary and cardiac care through Medicare reimbursement or sliding-scale or uncompensated care. COM participants also received weekly contact, intended to maintain comparable frequency of contact between groups.

The study description frames the intervention as follow-up after hospitalization for heart failure, with a bilingual clinician following participants for 3 months. It also describes plans for patient and caregiver input during the research and advice from a Community Advisory Board. These are details of the study description; the registry record does not report how consistently these processes occurred or how they changed intervention delivery.

The primary outcome is hospitalizations during a 90-day observation period, with the primary measure expressed as the participant mean. The listed timeframe is Baseline and Day 90. The registry record does not report the method used to generate the random sequence, allocation concealment, the basis for the enrollment target, or the original statistical design assumptions. It also does not report an analysis model beyond the posted outcome summary and comparison.

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: Intention to treat population (all participants who are randomized to either COM or TSM.

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 Hospitalizations, measured over Baseline and Day 90.

The primary endpoint is the number of hospitalizations during the 90-day observation period, summarized as a mean for each group. The registry record lists the timeframe as Baseline and Day 90 and identifies the analysis population as participants randomized to COM or TSM. The posted results give a mean of 0.55 hospitalizations for COM and 0.78 for TSM, with standard deviations of 0.9 and 1.3, respectively.

A mean provides a group-level summary of the number of hospitalizations per participant; it does not describe every participant's experience. The standard deviation indicates how much individual observations varied around each group mean. Hospitalization counts may be unevenly distributed, with many participants having few or none and some having more. A mean is still interpretable as an average count, but it does not show the full distribution or distinguish recurrent events from the experience of participants with no event.

Hospitalization is relevant to the study's focus on post-hospital heart-failure care because it records use of hospital services during follow-up. That relevance does not make the endpoint a complete measure of health, experience, or quality of life. Quality of Life is listed as a secondary outcome, but the registry record does not include a posted result for it.

7. Results Posted in the Registry

• Telehealth Self Management (TSM): mean 0.78 versus 0.55 on control; difference 0.23 (95% CI -0.21 to 0.67), standardized effect d = 0.21; z-test p = 0.307.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count summarized as a mean. The posted comparison treats the mean difference as TSM minus COM, so a positive estimate corresponds to a higher group mean in TSM, while a negative estimate corresponds to a lower group mean. The reported estimate is 0.23. This direction is a convention for expressing the comparison, not a judgment about whether a higher or lower count is clinically preferable.

What does a mean difference represent?

In general, a mean difference subtracts the average in one group from the average in another, using the stated order. It expresses the contrast in the original outcome units. Here, the units are participant mean hospitalizations during the 90-day observation period. The estimate of 0.23 is therefore a difference in mean hospitalization counts, not a difference in the number of participants hospitalized and not a percentage change.

The group means are 0.55 for COM and 0.78 for TSM. The posted standard deviations are 0.9 and 1.3. In general, the standard deviation describes the spread of individual values around a mean. It is not the uncertainty of the mean itself, and it should not be interpreted as a range containing all participants. The means and standard deviations give a compact summary, but they cannot show whether counts were concentrated at zero or spread across multiple values.

How should the confidence interval and p-value be read?

In general, a confidence interval describes the range of effect estimates that are reasonably compatible with the observed data under the assumptions of the calculation. It communicates uncertainty about the estimated difference, rather than the range of individual participant outcomes. The interval in this comparison includes zero and extends in both directions. Thus it is compatible with a negative mean difference as well as a positive one; the interval does not establish that the true difference is exactly zero.

The posted-summary recalculation reports an unadjusted p-value of 0.307. In general, a p-value describes how unusual the observed result, or a more extreme result, would be under a specified null model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The record does not report a full statistical analysis plan or adjusted model. The p-value should therefore be read as the unadjusted comparison reported from the posted summary statistics, not as a complete account of all analyses that may have been planned or conducted.

Why can count outcomes be challenging?

In general, count data are nonnegative and may have a skewed distribution. A mean and standard deviation can summarize them, but may not capture the pattern of event frequency, especially when many observations are low and a smaller number are high. Over-dispersion is a general concern when the observed variation in counts is greater than a simple model would expect. When present, it can affect standard errors and confidence intervals if the analysis model does not account for it. The registry record does not report a count-specific model, a dispersion assessment, or the underlying count distribution, so it is not possible to assess those features from the posted summary.

In general, repeated hospitalizations also raise a distinction between the number of events and the number of participants who experience at least one event. A participant mean count reflects all recorded events in the measure, whereas a participant-level yes/no measure would answer a different question. The posted endpoint is described as mean hospitalizations, and the registry record does not report a separate primary comparison of the proportion with any hospitalization.

What does the posted comparison establish?

The result is a comparison of group summaries for the stated 90-day outcome. Random allocation is part of the recorded design, but the registry record does not provide details about sequence generation or concealment. The posted result also does not establish that all randomized participants contributed complete outcome data, because missing-data counts and handling are not reported. An intention-to-treat population is named, but the record does not explain how that population was implemented in the analysis.

The results are not adjusted for other variables in the posted-summary recalculation. In general, adjustment can account for selected baseline characteristics or design features, but it does not automatically remove all bias and must be specified and interpreted in context. The registry record does not identify covariates, model diagnostics, or sensitivity analyses. Consequently, the reported mean difference and p-value should be read as the unadjusted summary comparison, with no claim that an unreported model was used.

10. Confidence Intervals and Precision

The mean difference is 0.23 hospitalizations, expressed as TSM minus COM. Its confidence interval extends across zero, with a negative lower endpoint and a positive upper endpoint. Because the interval includes values on both sides of no mean difference, the estimate is not precise enough to determine the direction of the underlying group difference from this comparison alone. The registry record's posted-summary recalculation also gives p = 0.307, which is consistent with substantial uncertainty under the unadjusted comparison.

Precision is different from whether an estimate is important. A confidence interval describes uncertainty around the estimated group contrast; it does not show the range of individual hospitalization counts or indicate whether any particular value would matter clinically. The record does not report a threshold for a clinically important difference, so the interval cannot be classified against such a threshold using the registry record alone.

The observed standard deviations are 0.9 in COM and 1.3 in TSM. These reflect variability in individual counts and help explain why group means alone are an incomplete account of the outcome. The registry record does not provide the raw distribution, the number with zero hospitalizations, or the number with recurrent events. It also does not report missing-data details or an adjusted estimate that might clarify the comparison.

The confidence interval and p-value should be understood as unadjusted recalculations from posted summary statistics. They summarize uncertainty under their calculation assumptions; they do not account for unknown analysis choices or unreported data features. In practical terms, the posted estimate points toward a higher mean in TSM, but its interval leaves room for a lower mean as well. The evidence is therefore inconclusive about the direction and size of the mean difference.

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. The record specifies a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The trial's own original sample-size justification and design assumptions are not reported in the registry record.

For this continuous outcome, the figures are participants per arm, not total events. The columns describe participants per arm for 80% power and participants per arm for 90% power. In general, power is the chance that a study would detect an effect of a specified size under the assumptions used, if that effect were present. The 90% power column plans for a greater chance of detecting the stated effect than the 80% power column, and therefore requires more participants per arm in the listed scenarios.

The rows use planning for 100%, 75%, and 50% of the observed effect. The figures are 356 and 477 participants per arm for planning at 100% of the observed effect; 633 and 847 participants per arm for planning at 75% of the observed effect; and 1,423 and 1,905 participants per arm for planning at 50% of the observed effect, respectively for 80% and 90% power. Smaller effects are harder to distinguish from random variation, so planning to detect a smaller effect requires more participants.

These are planning illustrations, not observed power. They are sensitive to the observed effect used as the starting point and to the stated assumptions. They do not establish that the original study was designed with these targets, and they do not revise or replace the observed outcome results.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.230; pooled standard deviation σ = 1.095.
2. n = 2 × (1.96 + 0.842)² × 1.095² / 0.230² = 355.6, rounded up to 356 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 <- 0.23                      # observed difference in means
sigma <- 1.09463934431284                     # 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(2196922)
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 = 356 (80% power), 477 (90% power)
 75% of observed effect: n per arm = 633 (80% power), 847 (90% power)
 50% of observed effect: n per arm = 1,423 (80% power), 1,905 (90% power)
power.t.test: 356.5 
Simulated power at n = 356 per arm: 0.799 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists three outcome measures, identifies Hospitalizations as the primary outcome, and lists Quality of Life as a secondary outcome. It reports one secondary outcome but does not include a posted result for it. The record does not report a complete outcome hierarchy, a prespecified multiplicity strategy, or whether adjustments were planned for testing more than one outcome or comparison.

Multiplicity is a general issue when several outcomes, time points, groups, or analyses are tested. The more tests conducted, the greater the chance that at least one result may appear unusual by chance if each is interpreted without considering the full set of tests. Methods for addressing multiplicity depend on the study questions and analysis plan. Since that plan is not reported in the registry record, it is not possible to determine whether or how multiplicity was handled. The posted p-value is described as unadjusted and should not be read as a multiplicity-adjusted result.

The primary outcome timeframe is listed as Baseline and Day 90, and the outcome description refers to hospitalizations during the 90-day observation period. The registry record does not report whether baseline values were used as covariates, how follow-up time was handled, or how any repeated measurements were analyzed. It does not report a model specification, analysis diagnostics, or sensitivity analyses.

The stated population is intention to treat, defined in the record as all participants randomized to either COM or TSM. However, the registry record does not report outcome completeness, the number of missing observations, reasons for missingness, or the method used to handle missing data. Naming an intention-to-treat population does not by itself explain how missing outcomes were addressed. The record also does not report deviations from assigned treatment or how such deviations were handled in the analysis.

These missing analysis details limit assessment of the robustness of the reported comparison. They do not show that the analysis was necessarily inappropriate; they mean the registry record does not provide enough information to evaluate these aspects.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The serious adverse event figures list 0 affected participants in COM and 0 in TSM. The numbers at risk are 58 for COM and 46 for TSM. The record lists 1 death in each group. It reports 0 affected participants for other adverse events in each group as well.

These entries are counts as recorded, not rates adjusted for follow-up time. The registry record does not report event descriptions or serious adverse event terms; the serious-event terms list is empty. It also does not report whether deaths were considered related to study participation or an intervention, nor does it provide timing or cause of death. The listed death counts should not be interpreted as evidence of causation.

Zero affected participants recorded for adverse events does not establish that no events occurred outside the registry's reporting definitions or observation process. The record does not provide additional detail about safety monitoring, ascertainment, or completeness. These figures describe what is posted and should not be used alone to compare safety between groups or draw a broader safety conclusion.

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

The posted results show that the mean number of hospitalizations during the 90-day observation period was 0.55 in COM and 0.78 in TSM. The unadjusted mean difference, TSM minus COM, is 0.23. The confidence interval includes zero and extends in both directions, and the posted-summary recalculation reports p = 0.307. The estimate is therefore uncertain: the data do not resolve whether the underlying mean was higher or lower in TSM compared with COM.

This does not establish that the groups had identical outcomes. A nonzero estimate was observed, but the uncertainty is broad enough to include both negative and positive differences. Conversely, the numerical direction of the point estimate alone does not establish that TSM caused a higher mean. The registry record does not report enough analysis detail to evaluate adjusted comparisons, missing outcome handling, or the distribution of counts.

The result concerns one outcome over the stated observation period. It does not establish the effect of either care approach on quality of life, longer-term hospitalization, mortality, or other outcomes. Quality of Life is listed as a secondary outcome, but its result is not reported in the registry record. The posted safety counts also do not provide enough context to support a comparative safety conclusion.

In general, a randomized comparison can help assess differences between assigned groups, but interpretation still depends on how the trial was conducted, how outcomes were measured, and how missing data and analyses were handled. The registry record describes randomization but does not report all of those details. The result should therefore be read as a limited, imprecise comparison of the posted group means, not as a verdict that one approach works or does not work.

16. Why This Trial Matters Statistically

This trial addresses a statistical challenge that is common in clinical research: a clinically relevant outcome may be recorded as a count, while the registry reports only group means and standard deviations. The mean difference is easy to state in the outcome's original units, but the summary does not show how many participants had no hospitalizations or how many experienced repeated events. That distribution can matter when interpreting count outcomes.

The uncertainty is central to the result. The estimated difference is 0.23, but its confidence interval crosses zero, and the posted-summary recalculation gives p = 0.307. The point estimate alone would give an incomplete account; the interval shows why the data do not settle the direction of the group difference. This illustrates the general value of reporting estimates alongside uncertainty rather than relying on a single test result.

The study also has features relevant to interpretation of trial comparisons: randomized assignment, parallel groups, weekly contact in both arms, and a 90-day outcome period. At the same time, the registry record does not report key analysis details, including missing-data handling, the count-outcome model, and multiplicity plans. Those omissions make it harder to judge how fully the posted summaries represent the intended analysis.

The planning illustrations show how sample-size needs depend on the effect targeted and the assumed power. They are not a retrospective measure of the trial's success and do not explain the original enrollment target, which is not reported in the registry record. Together, the outcome summary, uncertainty, and missing analysis details provide a useful case for separating observed estimates from conclusions that the available record cannot support.

17. Limitations

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

The registry record provides a limited account of the analysis. It reports group means, standard deviations, a mean difference, confidence interval, and unadjusted p-value, but does not report a detailed statistical analysis plan or the model used for the hospitalization count. The distribution of counts is not reported, so the frequency of zero events and recurrent hospitalizations cannot be assessed.

The record identifies an intention-to-treat population but does not report how many randomized participants had outcome data, whether any outcomes were missing, or how missing data were handled. It also does not report treatment deviations, adherence, or whether participants received additional care during follow-up. These details could affect interpretation of the group comparison, but their absence does not establish that such issues occurred.

Although the record describes randomization, it does not report how the sequence was generated or whether allocation was concealed. It does not report baseline characteristics or an adjusted estimate, so the extent to which groups differed at baseline cannot be assessed from the record's posted results. The original sample-size assumptions and justification are also not reported; the listed sample-size figures are planning illustrations based on the observed effect.

The study description focuses on community-dwelling patients in health disparity communities and follow-up after hospitalization, but the registry record does not report enough eligibility detail to judge how broadly the results apply. The primary result covers a 90-day observation period. A Quality of Life secondary outcome is listed without a result, and the record does not provide the broader outcome information needed to interpret other aspects of health or care.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Telehealth Management in HF Disparity Patients — NCT02196922.
• Registry record last updated 2017-11-17; status last verified 2017-10.
• Results first posted 2017-11-17.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes a completed, randomized, parallel-group study with no masking, sponsored by Northwell Health. It lists chronic heart failure as the condition and enrollment as 104. The study compared Telehealth Self Management with Comprehensive Outpatient Management, with weekly contact in both groups and a primary outcome of hospitalizations during a 90-day observation period.

The posted mean was 0.55 in COM and 0.78 in TSM. The TSM-minus-COM mean difference was 0.23; the confidence interval crossed zero, and the posted-summary unadjusted recalculation reported p = 0.307. The record lists Quality of Life as a secondary outcome but does not report its result. It also does not report detailed analysis methods, missing-data handling, or the original sample-size justification.