University of Michigan Advanced Heart Failure Tele-Monitoring and Flexible Diuretic Project: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for University of Michigan Advanced Heart Failure Tele-Monitoring and Flexible Diuretic Project (NCT02344342), a randomized na study. The primary outcome is Number of Days Hospitalized or Dead in the 180 Day Follow up Period.

2. Clinical Question

3. Trial Design

MANIFOLD-HF was a randomized, unmasked, 2x2 factorial study sponsored by the University of Michigan. The registry lists enrollment as 51, the start as 2015-02, primary completion as 2015-09, and status as terminated. A reason for termination is not reported in the registry record.

The factorial design crossed two interventions: use of the Bosch Health Buddy Web management system and use of a flexible diuretic regimen based on specified weight ranges. This creates four treatment combinations, including combinations in which participants received both interventions, either intervention alone, or neither. The posted primary-outcome summaries are organized by those combinations.

Random allocation means assignment was described as randomized; the registry does not report the allocation procedure in further detail. Masking is listed as none. Further information about how randomization was implemented, whether allocation was concealed, or whether outcome assessment was masked is not reported in the registry record.

The record lists three secondary outcomes: the Self Care for Heart Failure Index Score, the Minnesota Living With Heart Failure Questionnaire, and days to hospitalization or death if it occurs within 180 days. Although the registry lists four outcome measures in total, analyses are not posted. The reason the study was terminated, the conduct of any interim analyses, and the trial’s statistical analysis plan are not reported in the registry record.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The subjects included in this measure each completed the 180 follow up period or died within the 180 follow up period. Subjects that had less than 180 days of follow up at the time of study closure are excluded from this analysis.

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: Group Number of subjects TM\_YES\_AND\_FD\_YES 13 TM\_YES\_AND\_FD\_NO 11 TM\_NO\_AND\_FD\_YES 14 TM\_NO\_AND\_FD\_NO 13

6. Primary Endpoint

The registry lists the primary outcome as Number of Days Hospitalized or Dead in the 180 Day Follow up Period, measured over 180 days (6 months).

The primary endpoint is the total combined number of days that participants in each arm were hospitalized or dead between day 1 and day 180. The registry describes this as a way to combine hospitalization and death in heart failure trials. It is a count-based measure: days hospitalized for participants are added to days dead for participants within each arm.

This kind of composite count can represent more than one clinically important event in a single outcome. It can also be difficult to interpret as a single effect, because its value may reflect different contributions from hospitalization and death. The posted endpoint is not a time-to-first-event measure; it is described as a total of days across participants in an arm.

The registry’s analysis population for this outcome includes participants who completed the 180-day follow-up or died within that period. Participants with less than 180 days of follow-up at study closure are excluded. How many participants were excluded, and how their follow-up was handled in other outcomes, is not reported in the registry record.

The arm summaries list values of 5, 23, and 0 for three treatment combinations. The combination with Health Buddy No and Flexible Diuretic Yes has no posted value. The record also notes that values are missing or not estimable for a two-arm comparison, so these arm-level summaries should not be read as a complete comparative estimate.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a factorial design ask?

In general, a factorial design studies more than one intervention within the same trial. In this trial, the factors were telemonitoring and the flexible diuretic regimen, each represented as present or absent in the treatment combinations. This arrangement can support separate questions about each factor and a question about whether the effect of one factor depends on the other.

That last question is an interaction question. In general, an interaction means that the contrast associated with one intervention differs according to whether the other intervention is present. It is not answered merely by observing that one treatment combination has a larger or smaller outcome value than another. Estimating a main effect or interaction requires a specified statistical comparison and appropriate uncertainty estimates. Those estimates are not reported in the registry record.

What does randomization contribute?

In general, random allocation is intended to make treatment groups comparable with respect to factors that could influence outcomes, on average across repeated trials. It does not guarantee that small groups will have similar characteristics, and it does not correct problems such as incomplete follow-up or an unsuitable analysis. The record says the study was randomized, but details of the allocation process are not reported in the registry record.

Masking is listed as none. In general, lack of masking can matter when participants or study staff know the assigned treatment, because knowledge may affect behavior, care, or the recording and assessment of outcomes. The registry does not provide details about outcome ascertainment or whether any safeguards were used.

How should the count outcome be analyzed?

The primary outcome is a count of combined days, summed within each arm. Count outcomes can have distributions that differ from the normal, symmetric pattern often assumed by simple mean-based methods. They may also show over-dispersion: in general, this means that observed variation is greater than a model’s basic count assumptions allow. Ignoring over-dispersion can make uncertainty estimates too small and statistical conclusions too confident.

The appropriate analysis depends on the outcome’s distribution, the trial design, the estimand of interest, and the way follow-up and death are incorporated. Because the registry defines the measure as a combined total across participants, a reader would need analysis details to know whether the reported arm values are totals, how exposure or unequal follow-up was addressed, and how the factorial structure was reflected in the model. The registry does not report those details or a statistical analysis plan.

General methods for count outcomes include models that account for dispersion and methods that compare distributions or rates under clearly stated assumptions. These are general possibilities, not a description of the method used in MANIFOLD-HF. The registry lists zero analyses and does not identify a fitted model, an adjustment set, an estimate of effect, or an uncertainty interval.

What is a comparison estimate?

In general, a comparison estimate quantifies the contrast between groups for a stated outcome and target population. A confidence interval describes the range of values compatible with the data and statistical model under the method’s assumptions. It communicates precision as well as direction and magnitude. Neither a point estimate nor its interval establishes that a difference is causal without a sound design and analysis, and an interval’s interpretation depends on the statistical procedure used.

Here the record does not provide an estimable two-arm comparison for the primary count outcome. It therefore does not support a recalculated comparison from posted summary statistics. No recalculated estimate should be substituted for the sponsor’s analysis when the registry’s own analysis is available; in this record, analyses are not posted and no discrepancy is identified.

What can a planning calculation tell readers?

In general, a sample-size calculation describes the number of participants that a future study might require under stated assumptions about the outcome, effect to detect, allocation, significance level, and power. It is a planning illustration, not a measure of the probability that a completed study has reached a correct conclusion. An observed effect can be unstable, especially when based on sparse groups, so using it for planning requires caution.

The registry describes planning assumptions in general terms, including a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a fraction of the observed effect. It also distinguishes event totals for time-to-event outcomes from participants per arm for other outcomes. However, for this primary outcome, the record says values are missing or not estimable for a two-arm comparison. There is consequently no supported two-arm planning illustration based on an observed comparison for this endpoint. Such a calculation would not be observed power and should not be presented as a property of the completed trial.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists three secondary outcomes alongside the primary combined-days endpoint. These are the Self Care for Heart Failure Index Score, the Minnesota Living With Heart Failure Questionnaire, and days to hospitalization or death if it occurs within 180 days. Analyses are not posted, and the registry does not report estimates for these outcomes.

When a study examines several outcomes, treatment combinations, or contrasts, the chance of obtaining at least one apparently notable result can increase if many statistical tests are performed. In general, multiplicity procedures or a prespecified hierarchy can address this issue. The registry does not report whether outcomes were prioritized beyond identifying the primary outcome, how many comparisons were planned, or whether any adjustment for multiple testing was intended or used.

The primary outcome’s analysis population is described as participants who completed 180 days of follow-up or died within that period. Participants with less than 180 days of follow-up at study closure are excluded from this analysis. The number of participants excluded, reasons for incomplete follow-up, and any methods for handling missing outcome information are not reported in the registry record. How analysis populations were defined for the secondary outcomes is also not reported.

These omissions matter for interpretation. Excluding participants with incomplete follow-up may produce a different analyzed population from the randomized population, while a complete-case summary alone does not show whether missingness could have affected the results. The record does not provide enough information to assess that possibility or to reconstruct an intention-to-treat analysis.

11. 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 (38.5%), ventricular tachycardia/ventricular fibrillation (9.1%), congestive heart failure, (7.7%), occipital stroke (7.7%), death (7.1%).

The adverse-event record gives serious-event counts and at-risk denominators for each treatment combination. In the order Health Buddy Yes and Flexible Diuretic Yes, Health Buddy Yes and Flexible Diuretic No, Health Buddy No and Flexible Diuretic Yes, and Health Buddy No and Flexible Diuretic No, serious events affected 5 of 13, 4 of 11, 7 of 14, and 6 of 13 participants, respectively. Other events affected 0 in each group. The record lists deaths as 1, 1, 2, and 1 for those groups.

These are descriptive safety counts. They do not establish that one combination caused more events than another, and the registry does not post a comparative safety analysis or uncertainty estimates. The death counts in the adverse-event record should not be treated as the primary endpoint’s combined-days values; they are presented as a separate safety item.

The registry also lists serious-event terms with percentages: Hospitalization, 38.5%; ventricular tachycardia/ventricular fibrillation, 9.1%; Congestive heart failure, 7.7%; Occipital stroke, 7.7%; and Death, 7.1%. The denominator and calculation basis for these percentages are not reported in the registry record. These term percentages should therefore not be interpreted as group-specific risks or compared directly with the at-risk counts.

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

The record documents a randomized, unmasked factorial trial of home telemonitoring and a flexible diuretic regimen, with a primary outcome that combines days hospitalized and days dead during a 180-day period. It also documents four treatment combinations, an enrollment entry of 51, a terminated status, and descriptive primary-outcome and safety entries.

What the record does not provide is a completed statistical comparison of the treatment combinations. It lists outcome values for three combinations and no value for the fourth, and explicitly says a two-arm comparison is missing or not estimable. There are no posted analysis results, confidence intervals, or reported estimates of the separate effects of telemonitoring and flexible diuretic assignment. No conclusion about the direction, size, or precision of a treatment effect can be drawn from the available primary-outcome summaries.

The endpoint combines two types of burden into one count. A given arm total could reflect different contributions from hospitalization and death, and the posted data do not break the combined measure into those components. The outcome’s restriction to participants who completed follow-up or died, with those having less than 180 days at closure excluded, also limits what can be inferred about the randomized population as a whole.

The safety entries describe observed counts, not comparative causal effects. The percentages attached to event terms do not have a reported denominator in the registry record. The absence of posted analyses means the safety information cannot be used to quantify uncertainty or establish differences between treatment combinations.

Overall, the record describes the study’s question and some observed entries, but it is not a basis for declaring that an intervention did or did not change outcomes. The termination reason, detailed analysis methods, and handling of missing data are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable because it combines two interventions in a factorial design while defining its primary outcome as a total count that combines hospitalization and death. Each feature raises an interpretive question. The factorial structure calls for clarity about separate intervention effects and possible interaction; the combined count calls for clarity about how different event contributions are represented and analyzed.

Factorial designs can make efficient use of a study when the effects of multiple interventions are of interest. But the ability to draw conclusions depends on the observed data and on analyses that reflect the design. Here the posted primary-outcome record does not support a two-arm comparison, and no factorial estimates or interaction analysis are posted. The four combinations therefore describe the design, not a complete set of inferential findings.

The follow-up rule is also important: the primary analysis excludes participants with less than 180 days of follow-up at study closure. Readers need information about who was excluded and how missing outcomes were handled to judge how this restriction might affect interpretation. Those details are not reported in the registry record.

Finally, the status is terminated and the registry does not report why. That fact does not by itself explain the incomplete comparative results or establish that the study’s conduct or data were compromised. It does mean that readers should distinguish the trial’s stated objective and posted descriptive entries from evidence of a completed treatment comparison.

14. 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-outcome summaries are incomplete for a four-combination factorial design: one combination has no posted outcome value, and the registry states that a two-arm comparison is missing or not estimable. The listed participant counts for the primary outcome are 2, 2, 0, and 3, so the summaries offer very limited information for comparing combinations. The record does not include uncertainty estimates or analyses that could clarify how stable the observed values are.

The registry’s enrollment entry is 51, but the primary endpoint uses a restricted population and the arm summaries list smaller counts. The number excluded for having less than 180 days of follow-up at study closure is not reported. The record also does not report missing-data methods, analysis populations for secondary outcomes, or whether the primary outcome was analyzed according to randomized assignment.

The study was unmasked, and the registry does not describe outcome ascertainment or protections against possible effects of treatment knowledge. The statistical analysis plan, allocation details, interim analyses, multiplicity procedures, and reason for termination are not reported in the registry record. These gaps prevent assessment of whether the posted summaries reflect prespecified methods or how decisions about comparisons were made.

Safety information is also limited to descriptive counts and event terms. The basis for the listed event-term percentages is not reported, and there are no comparative safety estimates. Taken together, the record is useful for identifying the design and outcome definitions, but does not provide enough analysis detail to support a precise assessment of treatment effects.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: University of Michigan Advanced Heart Failure Tele-Monitoring and Flexible Diuretic Project — NCT02344342.
• Registry record last updated 2017-05-02; status last verified 2017-03.
• Results first posted 2017-05-02.
• 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.

18. Record Summary

MANIFOLD-HF was a University of Michigan randomized, unmasked 2x2 factorial study of telemonitoring and a flexible diuretic regimen for high-risk patients with heart failure. The registry lists enrollment as 51, start as 2015-02, primary completion as 2015-09, and status as terminated.

The primary outcome was the combined number of days hospitalized or dead during 180 days. The registry posts values for three treatment combinations, no value for the fourth, and indicates that a two-arm comparison is missing or not estimable. It lists three secondary outcomes but no analyses. The reason for termination, statistical analysis plan, missing-data methods, and detailed analysis populations are not reported in the registry record.