Dementia Symptom Management at Home Program: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dementia Symptom Management at Home Program (NCT03255967), a randomized na study. The primary outcome is Change in Quality of Life-Alzheimer's Disease Score.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group, single-masked, cluster-randomized study, described in the registry title as a multisite cluster RCT. The study summary identifies three study sites. The reported enrollment was 217, and the registry status is completed. The registry lists the condition as dementia and identifies the sponsor as NYU Langone Health. The phase is listed as not applicable.

The distinction between cluster and individual randomization matters. In a cluster trial, groups rather than individual participants are assigned to study conditions. Here, the care-team-level contrast is reflected in the descriptions: one group received care from teams that had received the DSM-H performance improvement program, while the control group received usual care from teams that had not received it. The specific unit randomized, and whether all sites contributed to both arms, are not reported in the registry record.

The program was intended for home healthcare registered nurses, occupational therapists and physical therapists. The study summary describes an aim to examine quality of life and healthcare utilization for people with dementia and their informal caregivers. The registry lists the primary endpoint as change in proxy-rated Quality of Life-Alzheimer's Disease score, with baseline and Day 60 as the assessment times. It also lists secondary outcomes covering symptoms, medication use, healthcare contacts and caregiver measures.

Individual dates and further operational details do not establish how outcomes were analyzed. The analysis plan, masking procedures beyond the single-masked designation, allocation concealment, and the handling of clustering are not reported in the registry record. Nor does the record summarize participant flow or the number included in each primary-outcome analysis. These omissions limit assessment of how the planned design was implemented.

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: The trial enrolled 215 Patient-Caregiver Dyads (215 patients with dementia (PWD) and 215 caregivers). 119 Dyads were randomized to the QI program care arm and 96 Dyads were randomized to the Control arm.

Overall Study

Reasons for not completing this period:

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-Alzheimer's Disease Score, measured over Baseline, Day 60.

The primary endpoint is change in the Quality of Life-Alzheimer's Disease score. It is a 13-item measure of health-related quality of life for people with Alzheimer's disease, assessed by a caregiver acting as proxy. Each item uses a 4-point Likert scale, from poor to excellent. The total score is the sum of responses and ranges from 13 to 52; higher scores indicate greater health-related quality of life. An increase indicates that the score rose during the observation period.

Quality of life is relevant to the study's stated focus on care for people with dementia in the home. A proxy assessment allows a caregiver to report on the person's quality of life, but it is not the same as a self-reported assessment by the person with dementia. The caregiver's perspective is therefore part of the measurement approach, not merely a way of recording an objective clinical event.

The registry specifies baseline and Day 60 as the assessment times. It does not report the observed scores, changes within either group, a between-group contrast, or its uncertainty. The record also does not state how the change score was constructed for analysis or how missing proxy assessments were handled. Without those results and methods, the endpoint definition can be described, but the effect of the program on this measure cannot be determined from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas are useful for understanding the design, even though the registry record does not report the analysis methods used for the primary endpoint. The distinctions below are general explanations, not claims about analyses conducted in this study.

Why clustering matters

In an individually randomized study, assignment is made separately for each participant. In a cluster-randomized study, assignment is made to groups, such as care teams or sites. Participants within the same group may share clinicians, practices, and local circumstances. Their outcomes can therefore be correlated. This similarity is called within-cluster correlation.

When outcomes are correlated within clusters, treating every participant as fully independent can make standard errors too small and confidence intervals too narrow. A suitable analysis typically accounts for the randomization unit and the within-cluster structure. The exact randomization unit, number of randomized clusters, correlation assumptions, and adjustment method are not reported in the registry record. The record cautions that a simple individual-level re-analysis would be misleading.

What an effect estimate and confidence interval would convey

An effect estimate describes the contrast between study conditions on an outcome, using a defined scale and analysis. A confidence interval describes uncertainty around that estimate under the assumptions of the statistical method. A narrower interval indicates greater precision than a wider interval, but neither an estimate nor its interval establishes clinical importance by itself. In a cluster trial, valid precision estimates must reflect clustering as well as the outcome distribution and study design.

No primary-outcome estimate or confidence interval is reported in the registry record. The specified endpoint is a continuous score, and the record identifies the mean as the summary measure and standard deviation as the dispersion measure. Those choices describe how the endpoint is characterized; they do not disclose the observed values or the method used to compare groups.

Change scores and comparisons

A change endpoint concerns how a measure differs between assessment times. Here, the registry specifies baseline and Day 60 and says that an increase in the score indicates increased health-related quality of life during the observation period. A study could compare average changes between groups or use another model that accounts for repeated measurements and clustering. The registry record does not say which approach was used. It also does not report whether baseline scores were adjusted for in the analysis.

Within-group change and a between-group contrast answer different questions. A group can change over time for reasons unrelated to the intervention, including changes that also occur under usual care. The randomized comparison is intended to assess the difference between conditions, subject to correct implementation and analysis. Since neither group changes nor a between-group result is reported, those comparisons cannot be interpreted from this record.

Over-dispersion and count outcomes

Some secondary outcomes are counts, such as outpatient visits and primary care contacts. In general, count data can vary more than a simple model assumes; this is called over-dispersion. If over-dispersion is present, a model that fails to account for it may understate uncertainty. The appropriate approach depends on how the counts were collected, their distribution, follow-up time, and clustering. Those details and the analysis methods for the listed count outcomes are not reported in the registry record.

Planning illustrations are not trial results

A planning calculation based on an observed effect is not the same as a prospective design justification, and it is not a measure of observed power. The registry's planning assumptions describe a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect expressed as a stated fraction of the observed effect. Such illustrations are conditional on their assumptions and do not demonstrate that the study had a particular probability of detecting an effect. The original sample-size justification is not reported in the registry record.

Power is a design concept tied to a specified effect and analysis assumptions. It should not be inferred after the fact from whether a result is statistically significant. In this study, no primary-outcome test or estimate is reported, and the cluster structure makes an individual-level calculation especially unsuitable for reconstructing the trial's precision. Planning numbers should therefore be read only as illustrations, not as observed evidence or as a substitute for the sponsor's analysis.

Multiplicity and interpretation

When a study evaluates multiple outcomes, tests, time points, or subgroup comparisons, the chance of at least one false-positive finding can increase if each is tested without adjustment. This general issue is known as multiplicity. The registry lists 23 outcome measures, but it does not report an analysis plan identifying which comparisons were confirmatory, whether multiplicity adjustments were planned, or which outcomes were analyzed. Without those details and results, isolated findings cannot be assessed in the context of a prespecified testing strategy.

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("QI Program Care", "Control")
n   <- c(119, 96)
m   <- c(12.32, 15.88)
sp  <- c(16.69, 22.77)
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  QI Program Care  119  12.320  16.690  9.290  15.350
CARM  Control  96  15.880  22.770  11.266  20.494

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 23 outcome measures, including the primary quality-of-life measure and secondary measures of symptoms, medication use, healthcare contacts, and caregiver outcomes. It reports no analyses in its analysis count. Results, test statistics, confidence intervals, and p-values for the primary and secondary outcomes are not reported in the registry record. The relationship between the registered outcomes and any prespecified statistical analysis plan is also not reported.

Multiplicity is a general concern when many outcomes or comparisons are evaluated. If multiple statistical tests are conducted, the chance of at least one false-positive result can rise unless the analysis plan accounts for that testing. The registry record does not report whether a hierarchy of outcomes, adjustment procedure, or other multiplicity strategy was used. Because no analyzed findings are presented, the impact of multiplicity on interpretation cannot be assessed.

Information about missing data and analysis populations is also limited. The record does not report the amount or pattern of missing primary-outcome measurements, reasons for missingness, imputation methods, or sensitivity analyses. It does not identify whether the analysis followed intention-to-treat principles, how participants were assigned to analysis groups, or who was included in each outcome analysis. Participant flow and exclusions from analysis are not reported in the registry record.

These gaps matter because missing outcomes and analysis choices can affect estimates and their uncertainty. In a cluster-randomized study, the handling of missing participants and clusters also needs to be understood alongside the clustering method. No assumptions about those procedures can be made from the registered endpoint descriptions alone.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hospitalization (19.8%), death related (10.4%), delirium (5.2%), caregiver severe depression (increase score on phq-9 [day 60]) (2.1%), caregiver suicide idealization (0.8%).

The registry reports serious and other adverse-event figures separately for people with dementia and caregivers in each study group. For QI Program Care, the serious-event figures are 26 among 119 people with dementia and 3 among 119 caregivers; the other-event figures are 21 among 119 people with dementia and 5 among 119 caregivers. For Control, the serious-event figures are 34 among 96 people with dementia and 4 among 96 caregivers; the other-event figures are 9 among 96 people with dementia and 1 among 96 caregivers. Deaths are listed as 6 for people with dementia and 0 for caregivers in QI Program Care, and 10 for people with dementia and 0 for caregivers in Control.

The registry also lists serious-event terms and percentages: Hospitalization, 19.8%; Death related, 10.4%; Delirium, 5.2%; Caregiver severe depression (Increase Score on PHQ-9 [Day 60]), 2.1%; and Caregiver suicide idealization, 0.8%. The denominator and method used to calculate these percentages are not reported in the registry record.

These counts should be read as descriptive registry entries, not as an adjusted comparison of safety between groups. The record does not report exposure time, event timing, statistical comparisons, or whether events were judged related to the program. The caregiver and people-with-dementia groups are distinct reporting categories, and the counts should not be combined. The figures alone do not establish causation or a difference in risk.

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

The registry record gives a clear description of the study question and the intended comparison. Home healthcare teams that received the DSM-H performance improvement program provided care to one group; another group received usual care from teams that had not received the program. The primary measure was caregiver-proxy-rated quality of life for people with dementia, assessed at baseline and Day 60. These details help identify the population, intervention, comparator, and outcome being studied.

What the record does not provide is an analyzed result for that primary outcome. There is no reported group difference, estimate of change, confidence interval, or significance test. Therefore, no conclusion about whether scores were higher, lower, or similar between groups can be drawn from the posted registry information. A missing result is not evidence of no effect; it is an absence of reported comparative evidence.

The same distinction applies to the listed secondary outcomes. Their inclusion in the registry indicates that they were designated as outcomes, but it does not establish what the study found. The safety counts are available as descriptive figures, but the record does not provide the analytic context needed to interpret them as causal contrasts between study conditions.

General statistical interpretation also depends on design and methods. Because this was cluster-randomized, outcomes from people receiving care through the same cluster may be related. A valid comparison needs to reflect the randomization and account for that dependence. The registry record cautions against a simple individual-level re-analysis, while the detailed sponsor analysis and its estimates are not reported in the record.

The record can thus support statements about the registered design, endpoint definitions, outcome list, and recorded adverse-event counts. It cannot support a claim that DSM-H did or did not change quality of life, symptoms, healthcare use, or caregiver outcomes. Nor can it establish clinical importance or precision without estimates and uncertainty measures. Those conclusions require analyzed results and enough information to evaluate the analysis.

14. Why This Trial Matters Statistically

This trial is statistically notable because the question is evaluated through a cluster-randomized design rather than a simple individually randomized comparison. In home healthcare, care is delivered by teams, and a team-level program can influence the care received by multiple people. That structure can make outcomes within a cluster more alike. Statistical analysis must therefore respect the way assignment occurred, or uncertainty may be misrepresented.

The primary endpoint also illustrates the importance of specifying who reports an outcome. Quality of Life-Alzheimer's Disease is assessed by caregiver proxy in this study. That gives the endpoint a defined perspective and makes caregiver experience relevant to measurement, while distinguishing it from a direct report by the person with dementia. A clear scale and assessment schedule do not, on their own, show whether the intervention changed the measure.

The record also demonstrates why outcome registration and outcome reporting serve different purposes. It lists a broad set of measures, from quality of life to medication use and caregiver burden, but does not report analyzed comparisons. Without estimates and uncertainty, readers cannot judge the direction, size, or precision of the findings. Reporting the design and outcome definitions is useful, but it cannot substitute for results.

Finally, the safety information illustrates the limits of raw counts. Counts separated by participant type and study group provide descriptive information, yet they do not account for exposure, cluster structure, or other features needed for a comparative safety analysis. The registry record makes the study's statistical structure important to interpretation while leaving key analysis details and outcome findings unreported.

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 for interpreting this trial from the registry record is the absence of analyzed primary and secondary outcome results. The record defines the primary endpoint and lists other outcomes, but it does not provide estimates, uncertainty intervals, or test results. As a result, the study's findings and precision cannot be evaluated from the registered information.

The cluster design raises further questions that the record does not answer. The specific unit of randomization, number of clusters, allocation procedures, and cluster-level analysis methods are not reported in the registry record. The summary identifies three study sites, but that does not establish the number or type of randomized clusters. The record itself cautions that simple individual-level re-analysis would be misleading.

Important implementation and population details are also not reported, including eligibility criteria, baseline participant characteristics, participant flow, and the analysis population for each outcome. The record does not describe adherence to the program, variation in usual care, or how masking was implemented beyond the single-masked designation. These omissions limit assessment of how closely the study conduct matched its registered design.

For the primary endpoint, the registry specifies caregiver proxy assessment at baseline and Day 60 but does not report missingness, the method used to analyze change, or how clustering and baseline measures were handled. For secondary outcomes, outcome-specific analysis methods are not reported. The statistical plan, multiplicity strategy, and sensitivity analyses are also not reported in the registry record.

Finally, adverse-event entries are not accompanied by definitions of the percentage denominators, exposure duration, timing, or causal attribution. They are descriptive counts and terms rather than a reported adjusted safety comparison. These limitations do not establish that the study was conducted inadequately; they define what cannot be assessed from the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Dementia Symptom Management at Home Program — NCT03255967.
• Registry record last updated 2024-01-30; status last verified 2024-01.
• Results first posted 2024-01-30.
• 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

The registry record describes NCT03255967, Dementia Symptom Management at Home Program, as a completed, randomized, parallel-group, single-masked, multisite cluster RCT with enrollment of 217. Its summary identifies three study sites and describes the intervention as a multicomponent quality improvement program for home healthcare registered nurses, occupational therapists and physical therapists. The comparator was usual care from teams that had not received the program.

The primary endpoint was change in caregiver-proxy-rated Quality of Life-Alzheimer's Disease score from baseline to Day 60. The registry lists 23 outcome measures, but reports no analyses in its analysis count and does not provide analyzed outcome results. It includes serious and other adverse-event figures by group and participant type, along with listed serious-event terms and percentages. The detailed statistical analysis methods, missing-data procedures, multiplicity approach, and analysis populations are not reported in the registry record. Consequently, the record describes the study question and selected safety information but does not permit assessment of the primary treatment contrast.