Supporting Family Caregivers With Technology for Dementia Home Care: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Supporting Family Caregivers With Technology for Dementia Home Care (NCT02483520), a randomized na study. The primary outcome is Change in Behaviorally Coded Challenging Behaviors (Agitated, Disruptive, Resistive).

2. Clinical Question

The central statistical question is therefore narrower than whether expert guidance changed behavior. The record describes a randomized comparison and reports counts of care videos submitted in each arm, but it also states that the intended behavioral outcome could not be coded consistently after the intervention. Video-submission counts describe recorded material, not the coded frequency or change in challenging behavior. They cannot, by themselves, answer whether behavior changed.

The brief summary describes both groups as recording care situations during daily care for 3 months. It says FamTechCare was designed to use video context, including events preceding a behavior, so that experts could identify possible triggers and suggest individualized approaches. The comparison group received nurse feedback based on verbal communication, with video-based expert feedback delayed until the end of participation. The registry record does not report a completed, interpretable estimate of the primary behavioral contrast.

3. Trial Design

The registry identifies this as a randomized, unmasked, crossover trial of FamTechCare. Its status is Completed. The listed start is 2014-03, and primary completion is 2019-01-31. The sponsor is University of Kansas Medical Center. The record lists enrollment as 189. Separately, the brief summary describes caregiver-person-with-dementia dyads (N=88) assigned to intervention or control groups. These figures are both present in the record, but their relationship is not explained.

The intervention was behavioral rather than a medication. In the intervention arm, caregivers submitted videos and received weekly individualized feedback after dementia-care experts reviewed the recordings. The control and delayed-intervention arm received weekly nurse feedback based on verbal communication while participating; its videos were held for expert review until the end of the study. The summary describes recording during daily care for 3 months. It also says caregivers selected care situations to record and uploaded videos to a secure site.

The record labels the design crossover, while the brief description focuses on random assignment to intervention or control and describes delayed FamTechCare for the control group. The crossover sequence, timing of any transition between conditions, and any washout period are not reported in the registry record. The record also does not explain how the crossover designation relates to the two arm descriptions or to the primary outcome.

Masking is listed as none. The registry does not report whether behavioral coders, outcome assessors, or analysts were otherwise masked. Details of allocation concealment, the randomization procedure, and the planned analysis model are not reported in the registry record. Those omissions prevent a fuller appraisal of how assignment and the crossover structure were handled statistically.

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.

Before assignment: Caregiver-Person with Dementia (PWD) dyads were randomly assigned to the intervention (FamTechCare) group or the attention control group using a quarter-based blocking strategy with 1:1 allocation developed by the study statistician. Caregivers in multiple dyad homes were cluster randomized to the same group.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Over 3 months, caregivers recorded videos during challenging care situations in order to receive feedback on interventions. Behavioral coding of the videos was not possible due to lack of consistent content post intervention. Caregivers failed to record follow-up videos of the care situations when they were no longer challenging.

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 Behaviorally Coded Challenging Behaviors (Agitated, Disruptive, Resistive), measured over 3 months.

The primary outcome is titled “Change in Behaviorally Coded Challenging Behaviors (Agitated, Disruptive, Resistive).” Its unit is care videos submitted, its parameter is NUMBER, and its timeframe is 3 months. The intended measurement was behavioral coding of videos to identify changes in agitated, disruptive, or resistive behavior before and after the intervention. In principle, coding care recordings can turn observations of challenging behavior into an outcome that can be compared across study groups and time points.

That intended interpretation depends on having suitable, consistently collected recordings from the relevant periods. The registry record states that behavioral coding was not possible because post-intervention video content was not consistent. It further says caregivers failed to record follow-up videos of the care situations, but the explanation is truncated in the record. Consequently, the listed outcome values are counts of submitted care videos, not a demonstrated measure of change in coded behavior.

The arm entries list 784 for Intervention FamTechCare and 253 for Control and Delayed FamTechCare, with n values of 81 and 73, respectively. The record does not provide a dispersion measure or interval estimate for these entries. It also flags the outcome as not estimable for a two-arm comparison. The counts therefore should not be interpreted as a treatment-effect estimate or as evidence that challenging behaviors increased or decreased.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation begins by distinguishing the outcome the study intended to measure from the quantities the registry lists. The intended primary outcome is change in behaviorally coded agitated, disruptive, or resistive behavior. The unit attached to the outcome is care videos submitted, and the parameter is NUMBER. The arm entries give counts of 784 and 253. Because the registry says coding was not possible due to inconsistent post-intervention video content, those values should be read as video-submission counts rather than as coded behavior or change in behavior.

What does an estimable comparison require?

In general, a between-group treatment-effect estimate requires a defined outcome measured in a comparable way for the groups being contrasted. When the outcome is a change from before to after intervention, the relevant observations must also be available at the intended time points, and the analysis must specify how the measurements are combined. In this trial, the registry describes an intended before-and-after behavioral assessment but says consistent post-intervention content was unavailable for coding. It flags the outcome as not estimable for a two-arm comparison. No valid contrast for the primary behavioral measure is reported.

Video counts do not substitute for behavioral coding. A recording is an observation opportunity, not necessarily an occurrence of a challenging behavior. The number of submitted recordings can depend on recording and uploading practices as well as on the care situations captured. Since the record does not describe a coding result that can be compared across groups, the listed counts cannot establish a change in behavior or the direction or size of an intervention effect.

Why does the design label matter?

Randomized allocation is listed, and the trial model is listed as crossover. In general, a crossover design compares conditions within participants across periods, so period order and the timing of transitions are relevant to analysis. A delayed-intervention comparison may also involve different exposure timing. The brief summary describes random assignment to intervention or control groups and later expert video feedback for the control group, but the registry does not report the crossover sequence, periods, transition timing, or washout details. It also does not explain how those design features were incorporated into an analysis. The design label alone is not enough to reconstruct the statistical comparison.

Masking is listed as none. In general, lack of masking can affect reporting, measurement, or interpretation, depending on who knows group assignment and how an outcome is assessed. The registry does not report whether coders or analysts were masked, and the primary outcome could not be coded consistently in any case. No further conclusion about bias from masking can be drawn from the record.

How should uncertainty be understood?

In general, a confidence interval describes the range of effect sizes compatible with the data and the analysis, under its assumptions. A narrow interval indicates greater precision than a wide interval, while an interval spanning effects in different directions may leave the direction uncertain. For the primary behavioral comparison, no confidence interval is reported, and the outcome is marked not estimable. There is therefore no posted interval with which to judge the precision of a behavioral effect estimate.

In general, a p-value evaluates how compatible data are with a specified null model under the chosen analysis; it does not measure the probability that a treatment works or the size or importance of an effect. No p-value for an estimable primary behavioral comparison is reported in the registry record. Without a valid endpoint comparison, a test statistic would not resolve the underlying measurement problem.

The record notes 17 outcome measures and 13 analyses, and lists two secondary outcomes concerning treatment cost and an incremental cost-effectiveness ratio. In general, examining multiple outcomes or analyses raises questions about which comparisons were prespecified and how false-positive findings were managed. The multiplicity strategy, outcome hierarchy beyond the primary designation, and statistical analysis plan are not reported in the registry record. The counts of outcomes and analyses alone do not establish whether adjustment was planned or performed.

What do the planning calculations represent?

In general, a sample-size calculation is a planning exercise based on assumptions about the effect to detect, variability, allocation, significance level, and analysis. It is distinct from an observed result. The planning illustration associated with this study is described as using the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation; its target effect is a stated fraction of the observed effect. These are planning assumptions, not evidence that the trial achieved a particular power or that an effect was observed for the intended primary endpoint. The record does not report the trial’s own sample-size justification.

More generally, an analysis of count outcomes may need to account for how variable counts are and whether their variability is greater than a simple model assumes. Over-dispersion means that observed variability exceeds the model’s expected variability; ignoring it can make uncertainty appear smaller than it is. This is a general statistical consideration, not a reported feature of this trial’s analysis. The registry does not report a count model, dispersion estimate, or model-based comparison for the video counts.

Finally, the different enrollment and arm figures should be kept separate. The registry reports enrollment of 189; the brief summary describes dyads (N=88); the primary-outcome entries list n values of 81 and 73; and the adverse-event entries use at-risk counts of 92 and 97. The record does not explain how these figures relate to one another. Statistical interpretation should use the denominator attached to each particular entry rather than assume that the figures describe the same analysis population.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 17 outcome measures and 13 analyses. It identifies the behaviorally coded challenging-behavior measure as primary and lists two secondary outcomes concerning FamTechCare treatment cost and an incremental cost-effectiveness ratio. The record does not report a multiplicity strategy, a detailed outcome hierarchy, or which analyses were prespecified. In general, when many outcomes or comparisons are examined, the chance of observing an apparently notable result by chance can increase; a prespecified plan or adjustment approach helps readers understand how that issue was handled. No such approach is reported in the registry record.

Missing outcome data are especially important for the primary endpoint. The record states that behavioral coding was not possible because post-intervention video content was not consistent, and that caregivers failed to record follow-up videos of the care situations. The registry entry’s explanation is truncated, so the full account of the recording problem is not available. It does not report how many follow-up recordings were missing, whether missingness differed by group, or whether any imputation, weighting, sensitivity analysis, or alternative endpoint was used. These details are not reported in the registry record.

The populations used for the different reported entries are also unclear. Enrollment is listed as 189, the brief summary describes dyads (N=88), arm entries for the primary outcome list n values of 81 and 73, and adverse-event entries give at-risk counts of 92 and 97. The registry does not reconcile these numbers or identify analysis populations such as intention-to-treat or per-protocol populations. It does not report how crossover status, delayed feedback, or participants with incomplete video data were handled in the analysis. Consequently, readers cannot determine from the record which participants contributed to a given outcome analysis or how missing data affected its interpretation.

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 5% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: attempted suicide (1.0%).

The registry reports serious adverse-event entries separately from the primary behavioral outcome. For Intervention FamTechCare, the serious-event entry lists 92 at risk and zero affected. For Control and Delayed FamTechCare, it lists 97 at risk and one affected. The recorded serious-event term is attempted suicide, listed as 1.0% for that term. The record lists zero other adverse events and zero deaths in each group.

These entries are descriptive counts. They indicate what the registry reports for the stated risk groups; they do not show whether an event was caused by the intervention, whether group differences are statistically distinguishable, or whether the groups had comparable observation periods. The record does not report a safety-effect estimate, confidence interval, or hypothesis test. It also does not explain how these at-risk denominators relate to the outcome arm entries or the enrollment figures.

Zero recorded events in a group should not be read as proof that an event cannot occur or that there is no risk. Likewise, a single recorded event does not establish a causal relationship. The appropriate reading is limited to the adverse-event counts and term recorded, with the denominator and reporting context kept in view.

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

The posted record documents a randomized comparison of FamTechCare with a nurse-feedback control that later received delayed video-based expert feedback. The intervention was designed to connect family caregivers with dementia-care experts using submitted recordings of challenging care situations. The intended primary outcome was change in agitated, disruptive, or resistive behaviors identified through behavioral coding of videos over 3 months.

What the record does not establish is whether the intervention changed those behaviors. It says coding was not possible because the post-intervention video content was not consistent and marks the primary outcome as not estimable for a two-arm comparison. The arm values of 784 and 253 are counts associated with care videos submitted, not coded behavior outcomes. Without a valid, comparable behavioral measure, those figures cannot be treated as an effect estimate.

The record also does not provide an interval or inferential test for a primary behavioral comparison. No conclusion about the magnitude or precision of a behavioral effect can therefore be drawn from the posted values. The mismatch between the registry enrollment figure, the dyad count in the brief summary, and the outcome and safety denominators further limits reconstruction of the analysis populations. The record does not explain those differences.

These limitations do not erase what the record describes about the intervention and data-collection approach. They do constrain the statistical claim: the study record reports video submission and adverse-event entries, but it does not report a completed, interpretable estimate for the primary coded-behavior outcome. That is a statement about the available registry result, not a medical recommendation or a conclusion that the intervention is effective or ineffective.

More generally, an intervention can be evaluated only to the extent that the intended outcome is measured consistently and analyzed in a way that matches the design. Here, the registry’s account of incomplete or inconsistent follow-up video content is central to interpretation. Further details about missing recordings, outcome coding, analysis populations, crossover handling, and statistical methods are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically informative because its record illustrates the difference between collecting data and measuring the outcome that a study set out to assess. FamTechCare used video submissions to support expert feedback, and the primary outcome was intended to quantify changes in coded challenging behavior. The registry says the behavioral coding could not be completed consistently after the intervention. That limitation affects whether the central randomized comparison can be estimated, regardless of how many recordings were submitted.

The posted video counts are relevant to understanding data collection, but they are not interchangeable with the behavioral endpoint. Treating submission volume as though it were a coded behavior result would change the meaning of the outcome. The record’s explicit designation of the primary result as not estimable for a two-arm comparison is therefore important: it distinguishes an unavailable effect estimate from a null result or a precise estimate showing little change.

The record also gives readers a reason to examine design and denominators carefully. It identifies a crossover model but describes random assignment and delayed expert feedback without reporting the sequence or analysis details. Enrollment, dyad, outcome, and safety figures are all listed, but their relationship is not explained. These features show why an outcome count alone cannot reveal which population was analyzed or whether the reported comparison matches the trial design.

Finally, the safety entries can be read descriptively, while the lack of a reported inferential safety analysis limits broader comparison. The key statistical lesson is not a claim about benefit or harm. It is that the validity and interpretability of an effect estimate depend on outcome completeness, consistent measurement, clear denominators, and an analysis aligned with the design.

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 the registry’s statement that the intended behavioral outcome could not be coded because post-intervention video content was not consistent. The description also says caregivers failed to record follow-up videos of the care situations, but the sentence is incomplete. The record does not provide enough detail to establish how extensive the missing recordings were, why they were missing, or whether the problem differed between groups.

The primary-outcome entries report counts of videos submitted, while the outcome title concerns change in coded behavior. The record marks a two-arm comparison as not estimable and gives no confidence interval or p-value for the intended behavioral effect. A video count cannot resolve the absence of a usable coded outcome. The posted counts should therefore be interpreted as recording activity rather than behavioral change.

There are unresolved differences among the enrollment and analysis figures: enrollment is listed as 189, the brief summary describes dyads (N=88), the outcome arm entries list n values of 81 and 73, and safety entries list 92 and 97 at risk. The registry does not reconcile these figures or identify the analysis population associated with each. This makes it difficult to determine how the reported entries map to randomized participants and follow-up.

The record labels the model crossover but does not report the sequence, timing of conditions, washout, or statistical handling of periods. Masking is listed as none, while masking of outcome coders and analysts is not reported. The statistical analysis plan, missing-data methods, multiplicity approach, and trial-specific sample-size justification are also not reported in the registry record. These gaps limit assessment of the intended analysis as well as interpretation of the posted descriptive results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Supporting Family Caregivers With Technology for Dementia Home Care — NCT02483520.
• Registry record last updated 2020-11-13; status last verified 2020-10.
• Results first posted 2020-11-13.
• 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

NCT02483520, “Supporting Family Caregivers With Technology for Dementia Home Care,” is listed as a completed randomized, unmasked crossover trial sponsored by University of Kansas Medical Center. The registry lists enrollment as 189 and primary completion as 2019-01-31. The brief summary describes caregiver-person-with-dementia dyads assigned to FamTechCare or control and video recording during challenging care situations over 3 months.

FamTechCare involved weekly individualized dementia-expert feedback based on submitted videos. The control and delayed-intervention group received nurse feedback based on verbal communication, with video-based expert feedback delayed until the end of participation. The primary outcome concerned change in coded agitated, disruptive, or resistive behavior. The record says coding was not possible because post-intervention video content was not consistent and marks the two-arm outcome comparison as not estimable.

The record lists video-submission counts of 784 and 253 for the two arms. It also reports serious adverse-event entries of zero among 92 at risk in the intervention group and one among 97 at risk in the control and delayed group; other events and deaths are listed as zero. The relationship among enrollment, dyad, outcome, and safety denominators is not reported in the registry record.