Heart Watch Study: a Pragmatic Randomized Controlled Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Heart Watch Study: a Pragmatic Randomized Controlled Trial (NCT04468321), a randomized na study. The primary outcome is Change in Quality-of-life (Atrial Fibrillation Effect on QualiTy-of-life).

2. Clinical Question

3. Trial Design

The Heart Watch Study is registered as a randomized, controlled, parallel-group trial with no masking. The two listed interventions are the Apple Watch and the Withings Move. Randomization is relevant because, when implemented appropriately, it aims to make the groups comparable at the start of a trial with respect to both measured and unmeasured characteristics. The registry record does not describe the randomization sequence or allocation concealment, so those implementation details cannot be assessed from the record.

The study summary describes a prospective, multicenter trial intended to assess patient-reported and clinical outcomes at six months and up to one year. The registered primary endpoint is narrower: change in the AFEQT global score from baseline to six months. The registry status is completed. The start date is 2021-08-18 and the primary completion date is 2023-12-01.

The registry reports enrollment of 105, whereas the brief study summary describes a trial of 150 patients. These are different figures in the record; a reason for the difference is not reported in the registry record. The number contributing to the primary outcome is also smaller than the enrollment figure, and the record defines this outcome population by completion of the AFEQT questionnaire at six months. It does not explain why some participants are not represented in the primary outcome summaries.

Masking is listed as none. Participants and study staff therefore were not masked according to the registry entry, although the record does not detail who knew each participant's assignment at each stage. The follow-up, adherence procedures, and additional co-interventions are not reported in the registry record. These omissions matter when interpreting a pragmatic device comparison: the posted results summarize outcomes under the recorded study conditions, but do not establish how closely device use followed any intended pattern.

The study lists 20 outcome measures, including the primary AFEQT change and secondary quality-of-life domains and clinical outcomes. The record does not provide enough detail to reconstruct a complete prespecified analysis plan or determine whether every planned outcome was collected and analyzed.

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: Number of patients that completed the AFEQT questionnaire at 6 months

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 Change in Quality-of-life (Atrial Fibrillation Effect on QualiTy-of-life), measured over baseline and 6 months.

The primary endpoint is change in the Atrial Fibrillation Effect on QualiTy-of-life questionnaire global score from baseline to six months. AFEQT is a patient-reported measure focused on the effects of atrial fibrillation on quality of life. The registry describes the score as running from 0 to 100, with higher scores indicating a better outcome. The endpoint is expressed as a change score so that the six-month result reflects movement relative to baseline rather than only the score at follow-up.

Patient-reported outcomes can capture symptoms, daily activities, and treatment concerns that may not be represented by clinical events alone. A global score provides a broad summary, while the listed secondary AFEQT domains focus on particular aspects of quality of life. These are complementary ways to describe patients' reported experience; a global result does not show that every domain changed in the same way.

The endpoint is reported as a mean change with a 95% confidence interval in each group. The Apple Watch group mean change is 20.3, with a 95% confidence interval from 14.7 to 25.8. The Withings Move group mean change is 17.9, with a 95% confidence interval from 8.6 to 27.2. The registry defines the outcome population as patients who completed the questionnaire at six months. It does not report a direct between-group estimate for the difference in change, so the group summaries should not be mistaken for that comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a continuous primary outcome, summarized as a mean change with a 95% confidence interval. The following describes general statistical concepts relevant to reading those results; it does not imply that a particular unreported method was used.

What does a mean change represent?

For each participant with the required measurements, a change score represents the follow-up score relative to the baseline score. The mean change summarizes the average of those participant-level changes in the analyzed group. It is different from a comparison of follow-up scores alone, because it incorporates baseline measurements in defining the outcome. The AFEQT score is oriented so that higher values indicate a better outcome, according to the registry. The posted mean changes are 20.3 for Apple Watch and 17.9 for Withings Move.

A group mean does not describe every participant's experience. Some participants may have changes above the mean, some below it, and some may have no change or change in the opposite direction. The registry does not post individual-level outcome distributions in the registry record described for this analysis, so the amount of participant-to-participant variation cannot be assessed from the reported confidence intervals alone.

How should a confidence interval be read?

In general, a confidence interval expresses the statistical uncertainty of an estimated quantity under the assumptions of the analysis. The 95% intervals for the two group means are reported separately: 14.7 to 25.8 for Apple Watch and 8.6 to 27.2 for Withings Move. They indicate uncertainty in the estimated mean change for each group. They are not confidence intervals for a difference between groups, and their overlap or non-overlap is not a substitute for a direct comparison.

A confidence interval is also not a guarantee that the true value lies inside the particular interval, nor does it include all sources of uncertainty. In general, it reflects sampling variability under a specified statistical model; it cannot correct for missing outcomes, measurement problems, deviations from randomization, or selective reporting unless those issues are addressed in the analysis. The model and handling of these issues are not reported in the registry record.

Why is the between-group contrast important?

In a randomized parallel-group trial, the central comparative question is generally addressed by estimating an outcome contrast between randomized groups. For a change-score endpoint, that could be a difference in mean change, with an uncertainty interval and an inferential test, subject to the chosen analysis plan. Such a contrast directly addresses whether the groups' changes differ. The registry record posts group means and confidence intervals, but does not report the primary between-group contrast or its test. No new contrast should be inferred from the two means in place of that missing analysis.

The study is randomized, which supports a comparative interpretation when analysis preserves the randomized groups and outcome information is adequately handled. However, the registry does not state whether the primary analysis followed intention-to-treat principles, whether it adjusted for baseline score or other variables, or whether a different method was used. Those methods are not reported in the registry record, so the group summaries cannot be used to infer them.

What does the secondary comparison add?

The secondary anticoagulant outcome is recorded as a percentage at six months. The posted percentages are 98.3% for Apple Watch and 91.6% for Withings Move. A recalculated, unadjusted risk difference comparing Withings Move with Apple Watch is -0.067, with a 95% confidence interval from -0.182 to 0.049 and p = 0.137. This is a simple unadjusted re-calculation from the posted summary statistics, not a reported adjusted trial analysis.

In general, a risk difference expresses the contrast between two proportions on an absolute scale, while a confidence interval represents uncertainty in that contrast. The interval for this recalculated difference includes zero, so these data do not provide clear evidence of a difference under the corresponding unadjusted test. This finding should not be interpreted as proof that the groups are equivalent. The registry does not report an adjusted analysis for this outcome, and the secondary comparison has to be considered alongside the trial's multiple outcomes and incomplete analysis details.

How should planning illustrations be distinguished from trial results?

The sample-size planning illustrations are not observed power and are not a record of the trial's original design justification. They are computed from the observed effect using a two-sided significance level of 0.05, equal allocation, and a normal approximation; the illustrations consider detecting 100%, 75%, or 50% of the observed effect. Such calculations are hypothetical planning aids: they describe sample sizes under stated assumptions, not the probability that the completed trial detected an effect, and not the sample size used in this trial. The original sample-size justification is not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 20 outcome measures, including the primary AFEQT change and multiple secondary outcomes. When a study evaluates many outcomes, the chance of at least one apparently unusual result can increase if each is assessed independently. In general, multiplicity can be addressed through prespecifying a hierarchy, adjusting inferential thresholds, or treating secondary analyses as exploratory. The registry record does not report a multiplicity strategy, so it is not possible to determine whether or how the family of outcomes was accounted for.

The record identifies the primary outcome population as patients who completed the AFEQT questionnaire at six months. The group sample sizes reported for the primary outcome are 60 for Apple Watch and 25 for Withings Move. Enrollment is recorded as 105, and the brief summary describes 150 patients. These figures refer to different record entries and should not be combined or treated as interchangeable. The reason for the difference between the enrollment and summary figures is not reported in the registry record.

The record does not describe missing-data reasons, the amount of missing baseline or follow-up information, an imputation method, or sensitivity analyses for missing outcomes. It also does not state whether the primary analysis used all randomized participants, only participants with observed questionnaires, or another analysis population. These details are important because an analysis restricted to people with completed questionnaires can be affected if completion is associated with outcomes or group assignment. Whether that occurred, and in which direction it might influence the results, cannot be determined from the registry record.

The registry does not describe a statistical analysis plan for the primary endpoint, including adjustment for baseline score, covariates, or other design features. The posted group summaries alone do not establish which analysis was planned or performed. A clear account of the analysis population, missing-data approach, and multiplicity handling would be needed to assess the inferential framework fully. In their absence, interpretation should remain focused on the posted descriptive estimates and their stated uncertainty.

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.

The registry reports serious adverse events in the groups listed as at risk: 0 of 70 for Apple Watch and 0 of 35 for Withings Move. It also reports 0 other adverse events and 0 deaths in each group. These are event counts recorded in the registry; they are not proof that no events occurred outside the recorded categories or observation process.

The safety denominators differ from the primary endpoint sample sizes, which are based on completion of the AFEQT questionnaire. They should not be substituted for one another. The registry does not explain the difference in group denominators, the period over which adverse events were collected, or how events were defined and assessed. The listed serious adverse event terms are empty, so no specific serious event type is enumerated.

In general, zero recorded events in a study does not establish that the underlying risk is zero. Rare events may not appear in a study, and interpretation depends on follow-up, event ascertainment, and the number of participants observed. The registry provides no comparative safety analysis or precision estimate for these counts. They are therefore a descriptive record of reported events, not a definitive estimate of relative safety between devices.

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

The primary results show mean AFEQT improvement from baseline in both randomized groups among the participants represented in the reported outcome summaries. The mean change is 20.3 for Apple Watch and 17.9 for Withings Move. The separate confidence intervals describe uncertainty around those group means. Because the registry does not report a direct comparison of mean changes, the summaries do not establish the size or precision of a difference between the devices.

This distinction is important: a within-group change asks how the average score shifted from baseline in that group, while a between-group contrast asks whether the change differed across randomized assignments. Participants' scores can change over time for many reasons, and the trial comparison is intended to distinguish the relative group outcomes. The posted group estimates alone cannot answer that comparative question with the same precision as a reported contrast and its confidence interval.

The results also do not establish how important the observed changes are to patients. The registry specifies the score direction but does not report a threshold for clinically meaningful change. Nor do the reported means show the range of individual experiences or whether all participants improved. Averages summarize groups and can conceal variation.

The secondary and safety information has separate limitations. The secondary anticoagulant comparison is an unadjusted recalculation, and the record does not provide a reported analysis plan for interpreting multiple outcomes. The adverse-event counts are descriptive, with no comparative safety estimate or detailed event-collection information. Neither set of results should be taken as proof of equivalence or proof of absence of risk.

Overall, the registry supports a description of the recorded group-level AFEQT changes and available secondary and safety entries. It does not support a definitive comparative claim about device effectiveness, nor does it provide enough analysis detail to resolve missing-data, multiplicity, or clinical-importance questions. No treatment advice follows from these registry summaries.

13. Why This Trial Matters Statistically

This trial raises an important statistical distinction between improvement within each group and a comparative effect between groups. In randomized studies, the central question is usually about the contrast associated with assignment, not whether each group changed from its own baseline. Reporting group means and their separate confidence intervals is useful descriptive information, but it does not replace a direct estimate of the primary between-group difference and its uncertainty.

The outcome is patient-reported and relevant to quality of life, while the interventions differ in their listed capabilities: one device includes irregular rhythm detection and ECG capabilities, and the other provides activity tracking. The result therefore concerns a comparison of assignment to these device strategies in the trial context, not an isolated test of any single device feature. The registry does not report adherence or device-use information that would clarify exposure to those capabilities.

The difference between the registry's enrollment figure and the number in its brief study summary, together with the smaller primary outcome populations, makes transparent reporting of participant flow and analysis populations especially important. Missing outcome information can affect both precision and validity. Clear reporting of a prespecified primary analysis, missing-data handling, and secondary-outcome strategy would help readers distinguish planned inference from descriptive or exploratory results.

Finally, the secondary outcomes and adverse-event counts illustrate why evidence needs to be read in context. Multiple outcomes can broaden understanding, but unadjusted secondary findings are less conclusive; zero recorded events do not establish zero risk. The registry's record is valuable as a source of outcome summaries, yet its incomplete comparative analysis details limit how far those summaries can answer the primary statistical question.

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.

A central limitation is that the registry reports group-level means and confidence intervals for the primary endpoint but not a direct between-group estimate, confidence interval, or test. As a result, the comparative question is not answered with a reported primary contrast. The separate intervals cannot be used as though they were an interval for the difference.

The primary endpoint is based on patients who completed the AFEQT questionnaire at six months. The registry does not give a participant-flow explanation, reasons for missing questionnaires, or methods for handling missing data. The outcome sample sizes are 60 and 25, while enrollment is recorded as 105 and the brief summary describes 150 patients. The reason for the mismatch between enrollment and summary is not reported in the registry record. Without further information, it is difficult to judge how representative the outcome population is of those enrolled.

The record does not provide a detailed statistical analysis plan, including the planned primary model, analysis population, baseline adjustment, or handling of multiple outcomes. It also does not report a threshold for meaningful AFEQT change. These omissions limit assessment of both the inferential approach and the practical importance of the score changes.

The trial is unmasked according to the registry, and the record does not describe adherence, device-use intensity, or participant and staff procedures. For a device comparison involving patient-reported outcomes, these details could inform interpretation, but their effects cannot be inferred. Allocation concealment is also not described.

Secondary reporting is incomplete for assessing a consistent pattern across outcomes, and the available recalculated comparison is unadjusted. Safety counts are descriptive, with limited information on event definitions, collection periods, and comparative precision. Finally, the registry's planned summary and enrollment entries differ, and no explanation is recorded. These limitations do not negate the posted summaries, but they constrain conclusions beyond those descriptive data.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Heart Watch Study: a Pragmatic Randomized Controlled Trial — NCT04468321.
• Registry record last updated 2025-04-18; status last verified 2025-04.
• Results first posted 2024-12-20.
• 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

The Heart Watch Study is a completed, randomized, controlled, parallel-group trial with no masking. It compared an Apple Watch Series 6 with Irregular Rhythm Detection and ECG capabilities against a Withings Move with activity tracking. The registry lists atrial fibrillation and cardioversion as conditions, records enrollment of 105, and gives dates of 2021-08-18 for study start and 2023-12-01 for primary completion. The brief summary describes a trial of 150 patients, a difference from the enrollment entry that is not explained in the registry record.

The primary outcome is change in AFEQT global score from baseline to six months, with higher scores indicating a better outcome. The reported mean change is 20.3 for Apple Watch and 17.9 for Withings Move, with group-specific 95% confidence intervals. A direct between-group primary estimate is not reported. Secondary quality-of-life and clinical outcomes are listed, and the record includes an anticoagulant-use comparison that is available as an unadjusted recalculation. Safety entries report no serious adverse events, other adverse events, or deaths in the listed groups. Missing-data methods, analysis populations, and a detailed statistical plan are not reported in the registry record.