Socially Assistive Robots for Interaction With Older Adults With Dementia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Socially Assistive Robots for Interaction With Older Adults With Dementia (NCT04384978), a randomized na study. The primary outcome is Change in Cognitive/Memory Ability.

2. Clinical Question

The study is registered as NCT04384978 under the title Socially Assistive Robots for Interaction With Older Adults With Dementia. Its official title describes development of serious games and physical activities instructed by a social robot for improving well-being among elderly individuals with Alzheimer disease and dementia. The primary outcome registered for results is cognitive or memory ability, rather than a direct measure of social or emotional well-being. The brief study summary says the investigators intended to measure social and emotional well-being as well as cognitive or memory functions, but no secondary outcomes are posted.

The clinical comparison is specifically between games administered with robot interaction and solitary games without robot interaction. Because the comparator also involves games, the question is not simply whether taking part in an activity differs from doing nothing. The registry record does not report enough outcome information to establish whether the groups differed in cognitive or memory ability.

3. Trial Design

NCT04384978 is recorded as a completed, randomized, parallel-group behavioral study with no masking. The record lists DreamFace Technologies, LLC as sponsor. The study start date is 2020-01-12, and the primary completion date is 2022-03-06. The phase is listed as NA. These registration details describe the study structure but do not disclose how the random sequence was generated, how assignment was concealed, or how completion was defined.

The Active Group was assigned serious games with Ryan, the social robot. The registry description says participants played games with and administered by Ryan 2-3 times a week and 30-minutes per day. The Control Group played solitary games and activities, including word puzzles, without interaction with Ryan. The record does not report whether the activities, session duration, staff contact, or other aspects were otherwise matched between groups.

The brief summary describes a 12-month pilot and says a total of 80 senior subjects would be assigned to two groups. The registered enrollment is 22. The primary-outcome description also states that participants did not complete the study. These registry entries indicate that the posted outcome record should not be read as a complete account of the originally described study plan or as evidence that all planned follow-up was obtained. The registry does not report why enrollment differed from the summary's stated total, how many participants completed each study visit, or why participants did not complete the study.

Randomization can support a fair comparison when allocation is implemented as intended and outcome information is available for the groups. It does not by itself resolve missing outcomes or establish that the analyzed participants represent everyone assigned. With no control-group outcome and no reported comparative analysis, the randomized design cannot yield an interpretable estimate of the difference between the assigned interventions from the posted results.

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 participants did not complete the study.

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 Cognitive/Memory Ability, measured over assessed at baseline (week 1), week 4, and 8. Week 8 reported..

The primary outcome is titled Change in Cognitive/Memory Ability. It uses the Saint Louis University Mental Status test, abbreviated SLUMS in the registry description. The description characterizes SLUMS as a screening method for Alzheimer disease and other dementia. It is a score on a scale, with values ranging from 0 to 30. In a person with a high school education, scores of 27 to 30 are considered normal, scores between 21 and 26 suggest a mild neurocognitive disorder, and scores between 0 and 20 indicate dementia.

A cognitive screening score is used to characterize performance on the assessment. A change outcome is intended to describe how that score moves from an earlier assessment to a later one. The record specifies assessment at baseline (week 1), week 4, and week 8, and says week 8 is reported. The outcome is recorded as a mean with standard deviation.

Care is needed in interpreting the posted number: the registered outcome is described as a change, but the record does not clarify the calculation used to define that change or provide the baseline result. It reports a mean of 23.75 and standard deviation of 3.45 for the Active Group, with 18 participants. The Control Group has no posted value. The Active Group mean therefore cannot be treated as a comparative change estimate, and the listed score thresholds should not be applied to it as though it were an individual participant's raw SLUMS score. The registry does not report additional details about scoring, assessor procedures, or the clinical interpretation of a given amount of change.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The record identifies a randomized parallel-group design and reports a continuous primary outcome summarized as a mean and standard deviation. Those features describe the intended comparison and the kind of outcome summary. They do not show which statistical model was used, how the registered change outcome was calculated, or how uncertainty in a group difference was estimated. No analysis is posted.

What does a mean describe?

General: a mean is the arithmetic average of the observations included in a summary. It can give a compact description of a group, but it does not reveal the distribution of individual values or how the average changed from baseline unless the outcome definition and the relevant baseline data are available. In this record, the Active Group mean is 23.75, based on 18 participants. The registry also gives a standard deviation of 3.45. Because the outcome is titled as a change but the scoring details and baseline mean are not reported, the meaning of the mean cannot be resolved more precisely from the registry record.

General: a standard deviation describes how spread out observations are around their mean. It is a description of participant-level variability, not an estimate of the uncertainty of the mean itself and not a measure of the difference between treatment groups. The posted standard deviation belongs to the Active Group result; there is no corresponding Control Group summary with which to compare spread or average outcome.

Why is the group comparison unavailable?

A randomized trial's main treatment-effect question is ordinarily expressed as a contrast between groups, such as a difference in average outcome. General: estimating that contrast requires outcome information for both groups and a defined analysis population and time point. Here the Control Group has no participants or outcome value recorded for the primary endpoint, and the registry lists no analyses. Thus a two-group effect estimate cannot be recovered from the posted endpoint summaries. The record explicitly indicates that values are missing or not estimable for a two-arm comparison of this continuous endpoint.

General: an estimate is more informative when paired with an uncertainty interval, because the interval conveys how much precision the data and analysis support. No confidence interval is reported. A p-value is also absent. Neither can be inferred solely from the Active Group mean and standard deviation; doing so would require information and assumptions not reported in the registry record.

What does randomization contribute?

General: random assignment is intended to make the groups comparable, on average, at the point of assignment, including with respect to factors that are measured and factors that are not measured. That design benefit depends on retaining and analyzing outcome information appropriately. Randomization does not fill in an absent control outcome or eliminate the need to describe missing data. The registry does not report the sequence-generation method, allocation concealment, a prespecified analysis plan, or how assigned participants were handled in analysis.

The study is described as unmasked. The registry does not specify who knew the assignments, whether outcome assessors were aware of group assignment, or what safeguards were used against measurement or reporting differences. These details matter particularly for an outcome collected through an assessment, but no further procedures are reported.

How should the planning illustrations be understood?

General: sample-size planning translates a chosen effect and statistical assumptions into a target number of participants. Such planning depends on the target effect, significance level, allocation, variability assumptions, and analysis method. Any figures based on an observed effect are illustrations under their stated assumptions; they are not observed trial results and are not observed power. The record describes a planning approach using a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects expressed as fractions of an observed effect. However, it also indicates that values are missing or not estimable for a two-arm comparison. No usable comparison effect is reported for this endpoint, so there is no posted basis for interpreting a planning illustration as a result of this study. The registry does not report the original sample-size justification or the assumptions used to plan the trial itself.

General: over-dispersion matters when observed variation is greater than a statistical model assumes, because standard errors and uncertainty intervals can then be too small if the extra variation is ignored. This outcome is continuous and the registry gives a standard deviation, but it does not report a fitted model or evidence of over-dispersion. Count-outcome methods that explicitly model over-dispersion are not described for this endpoint and cannot be inferred from the record.

Overall, the available statistical information supports a descriptive reading of the Active Group summary only. There is no registered analysis result from which to assess adjustment, handling of repeated assessments, missing outcomes, or an estimated treatment contrast.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Active Group")
n   <- c(18)
m   <- c(23.75)
sp  <- c(3.45)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Active Group  18  23.750  3.450  22.034  25.466

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and no secondary outcomes. It reports no analyses. With no posted analysis results, the record does not establish which outcome or time point was treated as the primary inferential test, whether multiple comparisons were planned, or how any multiplicity would have been addressed. No statistical analysis plan or prespecified hierarchy of outcomes is reported. The fact that one primary outcome is listed is not enough to establish that only one analysis was conducted.

The registered assessment schedule includes baseline (week 1), week 4, and week 8, while the timeframe says week 8 was reported. The record does not state whether week 4 was analyzed, whether each visit was considered a separate comparison, or how repeated observations within a participant were modeled. No adjustment for baseline outcome is described.

The outcome description states that participants did not complete the study. The registry does not explain the extent or reasons for missing outcome data, whether missingness occurred before or after assignment, or how incomplete observations were handled. It also does not report whether analyses followed the assigned groups regardless of adherence, or whether another analysis population was used. Those details cannot be inferred from the randomized designation.

The primary outcome gives an Active Group result for 18 participants, but no Control Group outcome. The record does not indicate whether the number represents everyone assigned to that group, everyone assessed at week 8, or another set of participants. It therefore cannot support conclusions about attrition patterns or the representativeness of the reported mean. A missing result is not evidence of no effect, and the absence of secondary outcomes does not establish that no other outcomes were measured outside the registry record.

12. 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 adverse-event record lists 4 Control Group participants at risk and 18 Active Group participants at risk. It reports 0 serious events among those at risk in each group, 0 other events among those at risk in each group, and 0 deaths in each group. These are the posted event counts; they should be read alongside the different group sizes recorded for safety monitoring.

General: an at-risk count identifies the participants included in a particular safety summary, while an affected count describes those recorded as having an event. The registry does not explain how these safety populations were defined, how long participants were followed for safety, or how events were solicited and classified. It also does not provide serious-event terms. The absence of recorded events in these summaries does not establish that no events occurred outside the recorded monitoring or reporting conditions, and it does not permit a reliable comparison of safety between groups.

The 4 Control Group participants at risk for adverse events do not match the Control Group's 0 participants recorded for the primary outcome. The record does not explain this difference. Safety counts and primary-outcome counts represent distinct posted information and should not be treated as interchangeable measures of study completion or analysis population.

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

The record documents a randomized comparison of serious games administered with a social robot against solitary games without robot interaction, and it reports one primary-outcome summary for the Active Group. That summary is a mean of 23.75 with standard deviation 3.45 among 18 participants. The record also reports no serious adverse events, other adverse events, or deaths among the listed at-risk participants in either group.

Those observations do not establish whether robot-assisted games changed cognitive or memory ability, whether the change differed from that under solitary games, or whether either approach had an effect on well-being. The primary outcome has no Control Group result and no reported analysis. The result label refers to change, but the record does not explain the calculation or provide a baseline value. It is therefore not possible to interpret the posted Active Group mean as a verified change from baseline or as a comparative effect.

Nor does the missing comparison show that there was no difference between groups. Equivalence or absence of an important difference requires a defined comparison and sufficiently precise data; neither is reported. Likewise, the lack of a p-value is not evidence of a null result. The registry record does not provide a confidence interval or other measure of comparative precision.

The listed SLUMS score thresholds offer context for the screening measure, but they do not resolve what the reported mean represents. The thresholds concern scores on the test, while the registered outcome is framed as change. Without a clear definition of the reported score, individual-level information, and a comparison group result, it would be inappropriate to assign a clinical category or infer a clinically meaningful change from the group mean.

The most defensible reading is narrow: a week 8 primary-outcome summary is posted for the Active Group, while the control outcome and the between-group analysis are unavailable in the registry record. The findings therefore describe incomplete outcome reporting rather than an interpretable test of comparative effectiveness. The safety summaries likewise report zero recorded events within the stated at-risk groups, but do not provide the monitoring detail needed for a broad safety conclusion.

14. Why This Trial Matters Statistically

This trial illustrates an important distinction between a randomized design and an analyzable randomized comparison. Random assignment is a valuable design feature, but its purpose is to support a comparison. When one group has no primary-outcome result and no analysis is posted, the central statistical question cannot be answered from the outcome record. A group mean by itself does not substitute for a treatment contrast.

The endpoint also highlights why outcome definitions matter. The registry labels the primary measure as change in cognitive or memory ability, identifies the SLUMS test and its scoring range, and specifies several assessment times. Yet it does not clarify how the posted mean relates to baseline or how change was calculated. A statistic can be numerically precise as a recorded summary while remaining difficult to interpret because its estimand and construction are unclear.

Recruitment and completion information are also relevant to statistical interpretation. The brief summary describes a target of 80 senior subjects, while the enrollment is 22, and the outcome description says participants did not complete the study. These registry entries raise questions about study completion and data availability, but the registry does not report the reasons or provide enough detail to answer them. This is a reminder that sample size alone does not explain the reliability of a result; the number and pattern of observed outcomes, the analysis population, and the comparison are essential.

For clinical researchers and biostatisticians, the record is a useful example of why results reporting should connect protocol-defined outcomes, assessment timing, participant flow, missing-data handling, and comparative estimates. Without these elements, readers cannot distinguish a descriptive observation from evidence about a randomized intervention contrast. Clear reporting of uncertainty would also be needed to assess precision rather than relying on a point estimate alone.

15. 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 principal limitation is the absence of a Control Group primary-outcome value. Although the study is described as randomized and parallel-group, the posted endpoint information cannot support an estimate of the difference between the assigned interventions. No analyses are reported, and no confidence interval or p-value is available.

The outcome definition is incomplete for interpretation. The registry names a change outcome and reports that week 8 was reported after assessments at baseline (week 1), week 4, and week 8. It does not state the formula for change, provide a baseline summary, or explain whether the posted mean is a change score or a week 8 score. This makes the scale context harder to apply to the result.

The primary-outcome description states that participants did not complete the study, but the record does not explain what completion means in this context, how many participants completed each assessment, or why observations are missing. The Control Group is recorded with 0 participants for the endpoint, while safety information lists 4 Control Group participants at risk. The record does not reconcile those different counts.

Details important for evaluating trial conduct and analysis are not reported in the registry record, including the randomization procedure, allocation concealment, masking of outcome assessors, analysis population, model, baseline adjustment, missing-data approach, and original sample-size justification. The brief summary's stated total of 80 differs from the registered enrollment of 22, but the reason for the difference is not given.

Safety interpretation is also limited by the lack of event definitions, monitoring duration, and reporting procedures. Finally, there are no secondary outcomes posted, so the registry results do not address the summary's stated interests in social or emotional well-being. These limitations restrict interpretation to the registry entries recorded and do not establish why the information is absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Socially Assistive Robots for Interaction With Older Adults With Dementia — NCT04384978.
• Registry record last updated 2024-07-23; status last verified 2024-06.
• Results first posted 2024-07-23.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT04384978 is a completed randomized, unmasked, parallel-group behavioral study sponsored by DreamFace Technologies, LLC. It compares serious games with the social robot Ryan against solitary games and activities without robot interaction. Registered enrollment is 22; the brief summary describes a 12-month pilot with a stated total of 80 senior subjects.

The primary outcome is change in cognitive or memory ability measured with the SLUMS test, with assessments at baseline (week 1), week 4, and week 8. The Active Group result is mean 23.75, standard deviation 3.45, among 18 participants. The Control Group has no primary-outcome participants or result, and no analyses are reported. The outcome description states that participants did not complete the study.

For safety, the record lists 4 Control Group and 18 Active Group participants at risk, with 0 serious events, 0 other events, and 0 deaths in each group. No secondary outcomes are posted. The registry record does not report a comparative effect estimate, confidence interval, p-value, or explanation of missing outcome information.