MOVE OUT: A Partnership With Veterans Groups to Enhance Weight Management in the Veterans Health Administration (VHA): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for MOVE OUT: A Partnership With Veterans Groups to Enhance Weight Management in the Veterans Health Administration (VHA) (NCT01339390), a randomized na study. The primary outcome is Change in Weight (From Baseline).

2. Clinical Question

3. Trial Design

MOVE OUT was a randomized, parallel-group behavioral study. The registry describes the study as cluster-randomized; accordingly, the unit of randomization was not necessarily the individual participant. A simple re-analysis that treats all participants as independently randomized could misrepresent the uncertainty, so the sponsor’s own analyses are the appropriate results to interpret from the posted record.

The two interventions were MOVE OUT and MOVE!. MOVE OUT was intended to build on MOVE! with peer support and more convenient local access to educational sessions and exercise groups. The study rationale was that attendance at classes and exercise at fixed or inconvenient locations and times can be a barrier to participation. The registry describes MOVE! as part of routine care at the study center, with a clinical reminder used to prompt referral consideration by a primary care provider.

The registry lists an enrollment of 3169. The study start is recorded as 2012-04, and primary completion as 2015-03; the status is completed. The listed phase is not applicable. Masking is recorded as none. The sponsor is the VA Office of Research and Development.

The primary outcome was change in weight from baseline, with weight recorded in the medical record during a 6-month period around follow-up. The outcome timeframe names 12 and 24 months. The record describes using last observation carried forward for participants lost to follow-up. It does not report the number of clusters, how clusters were formed, the allocation procedure within clusters, or whether any design features changed during the study.

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: For the analysis, we included those lost to follow up by using the last observation carried forward to impute missing values

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 Weight (From Baseline), measured over Measured at 12 and 24 months.

The primary endpoint is change in weight from baseline, measured in pounds. Weight is a continuous outcome: it can take a range of values rather than being limited to categories such as improved or not improved. A mean summarizes the group’s average outcome, while its uncertainty reflects both the spread of participant measurements and the amount of information available.

The record states that weight was obtained from the medical record during a 6-month period around each follow-up point, and names 12 and 24 months as the measurement times. Using medical-record measurements provides a way to assess weight without relying only on participant recall. However, a window around follow-up means the recorded measurement may not have been taken on the exact nominal follow-up date. The registry does not describe how a measurement was selected when more than one was available in that window.

Weight is a direct measure of body mass and is relevant to a trial focused on weight management. It does not, by itself, capture every aspect of health or participation in a program. The registry’s primary-outcome description identifies change from baseline, whereas the reported analyses are labelled mean differences in final values. The record does not explain this distinction or specify how the two reported analyses map to the named follow-up times. Those details limit how specifically the endpoint results can be interpreted.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes the primary outcome as a continuous measure and lists the reported analyses as two-sided t-tests comparing mean final values. The statistical interpretation of these results depends on the group contrast, the uncertainty around the estimate, and the trial’s clustered randomization. The record does not provide enough detail to reconstruct the full analysis procedure.

What does a mean difference represent?

A mean difference expresses how far apart the average outcomes in two groups are on the outcome’s measurement scale. For weight, that scale is pounds. A mean difference is not the same as a difference for every participant: it summarizes the groups and does not describe the range of individual responses. The analysis entries label their parameter as a mean difference in final values, while the primary outcome is described as change from baseline. The registry does not explain whether the final-value comparisons were adjusted for baseline or why this label differs from the primary-outcome wording.

How should a confidence interval be read?

A confidence interval gives a range of estimates compatible with the observed data and the statistical model used. Its width reflects precision: a narrower interval indicates more precision, and a wider interval indicates greater uncertainty. A 95% confidence interval is a procedure that, under its assumptions, would cover the true parameter in 95% of repeated applications. It is not a statement that there is a 95% probability that a particular fixed parameter lies in the interval.

For a group difference, zero is the value corresponding to no difference. Both reported 95% intervals include zero. Thus, the intervals do not rule out no difference, and they also do not establish that the groups are equivalent. The confidence limits also show why a non-significant result should not be reduced to the claim that the interventions had exactly the same outcome. The registry does not report a clinically important difference threshold against which to judge the range of plausible effects.

What does the p-value add?

The two-sided p-value describes how unusual data at least as inconsistent with a no-difference model would be if that model and the test assumptions were correct. Smaller values indicate greater incompatibility with that model, but p-values do not measure effect size, clinical importance, or the probability that a treatment works. The reported p-values are 0.494 and 0.306. They are considered together with the confidence intervals, not as replacements for them.

Two-sided testing considers departures in either direction. This is pertinent because the result entries do not specify which group was subtracted from which. Without that contrast definition, the direction of a positive or negative estimate cannot be translated reliably into a statement about which group’s average was higher.

Why is clustering important?

In a cluster-randomized study, groups or settings, rather than necessarily individual participants, are assigned to study arms. Outcomes among people in the same cluster can be correlated because they share a location, staff, practices, or other experiences. When this correlation is ignored, the analysis can treat the data as containing more independent information than they do and can produce confidence intervals that are too narrow. General statistical practice accounts for clustering in the model or variance calculation. The registry record does not report the number of clusters, the estimated within-cluster correlation, or enough information to reproduce the sponsor’s approach.

The record specifically indicates that a simple individual-level re-analysis would be misleading and that only the sponsor’s own analysis is shown. This is an important design constraint: the reported analysis should not be replaced by a basic comparison that assumes every participant was independently randomized. Details of the sponsor’s cluster-adjusted model are not reported in the registry record.

What does the reported outcome timeframe mean?

The primary outcome description names measurements at 12 and 24 months. It also says weight was taken from the medical record during a 6-month period around follow-up. A window can accommodate routine clinical measurement schedules, but it can also mean that measurements are not exactly concurrent across participants. The registry does not say how measurements within the window were selected or how the two reported analyses correspond to the follow-up times.

The analysis entries identify final values, while the outcome is titled change from baseline. These can refer to different analytical approaches: comparing final measurements is not necessarily the same as comparing changes or adjusting final measurements for baseline. The record does not report the model specification or clarify which approach was used beyond the label attached to the entries. Consequently, the interpretation should stay close to the reported final-value comparisons.

How does missing-data imputation affect interpretation?

The registry says that participants lost to follow-up were included using last observation carried forward. This procedure substitutes a participant’s most recently observed value for a missing later measurement. It can retain participants in an analysis, but it assumes that the last recorded value is an appropriate stand-in for the missing value. That assumption may not hold if weight changes after the last observation. The record does not give the amount or pattern of missingness, or report sensitivity analyses using alternative assumptions.

These methods explanations are general statistical context. The registry record does not provide a full statistical analysis plan, detailed model formula, or enough information to independently reproduce the sponsor’s cluster-aware analyses.

10. Analyses Reported by the Sponsor

The registry’s own results are the appropriate basis for reading this trial because it identifies a cluster-randomized design and warns that a simple individual-level re-analysis would be misleading. The reported sponsor analyses use two-sided t-tests and give two mean differences in final values, with 95% confidence intervals and p-values. A separate recalculation based on individual-level summary statistics is not presented: the record explicitly indicates that only the sponsor’s own analysis is shown.

This distinction is important. A recalculation that ignored clustering could give misleading uncertainty, even if its arithmetic were straightforward. The record does not supply the cluster-level information or model details needed to check the sponsor’s procedure independently. The sponsor’s reported analysis therefore takes precedence over any hypothetical individual-level calculation.

The two reported intervals both include zero, and the two p-values are 0.494 and 0.306. These findings do not establish equivalence or prove that there was no difference. The record also does not clearly map the two entries to the 12- and 24-month follow-ups, define the subtraction order, or reconcile the final-value analysis label with the change-from-baseline endpoint description. A precise account should retain these qualifications rather than infer a direction or assign an analysis to a particular time point.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and two reported analyses. The outcome timeframe names 12 and 24 months, but the two analysis entries are not explicitly identified as belonging to those separate time points. If both analyses represent testing at different follow-ups, considering multiple tests can affect the chance of at least one result appearing unusual by chance. The registry record does not report whether multiplicity was addressed, whether one time point was designated as primary for hypothesis testing, or whether any adjustment was planned. It is therefore not possible to determine how the two tests fit into a prespecified testing hierarchy.

The record states that participants lost to follow-up were included using last observation carried forward. This is an imputation method, not a direct observation of a missing outcome. It assumes the last recorded weight is an adequate substitute for the missing follow-up value. That assumption may be questionable when the outcome can change over time. The record does not report the extent of missing follow-up, the reasons for missingness, the number of imputed values, or sensitivity analyses that test alternative assumptions.

The analysis population is described only as including those lost to follow-up through that imputation approach. The registry does not report whether all randomized participants were included, whether any participants were excluded after randomization, or how the analysis population was defined beyond the imputation statement. It does not report a per-protocol analysis or other population-specific results.

The record also does not report an interim-analysis plan, stopping rules, or a complete statistical analysis plan. These items should not be inferred from the reported results. Together with the cluster-randomized design and the absence of detail on cluster-adjusted modeling, the missing information limits independent assessment of the analysis and of how uncertainty was handled.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: death at 24 months (1.6%), death at 12 months (0.4%).

The registry reports serious adverse events for the two study arms. For MOVE OUT, the record lists 32 serious adverse events among 1571 at risk; for MOVE!, it lists 29 among 1598 at risk. These are the registry’s reported counts, not a comparison adjusted for cluster assignment, follow-up, or participant characteristics. The record does not report a statistical comparison of these counts, so they should not be read as evidence of a difference between arms.

The serious-event terms listed include Death at 24 months at 1.6% and Death at 12 months at 0.4%. The record does not make clear in this listing which denominator or group each percentage uses, or how the two time-specific entries relate to the serious-event counts. Death fields for the arms are recorded as unavailable. The registry also lists no other adverse-event counts in the categories shown, but it does not provide details that would establish how completely or uniformly adverse events were collected.

These figures describe what is recorded in the registry, not the cause of an event or whether it was related to an intervention. The record does not report a between-group safety analysis or enough detail to interpret the percentages as risks over a clearly specified denominator. They should be read descriptively and in light of those reporting limits.

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

The weight results do not show a clearly resolved difference between MOVE OUT and MOVE! In the reported mean final values. Both confidence intervals include zero, and the reported p-values are 0.494 and 0.306. These results are compatible with uncertainty about the group difference. They do not prove that the interventions have identical effects, and they do not show that any possible difference is too small to matter.

The distinction between lack of clear evidence for a difference and evidence of equivalence is important. Equivalence requires a prespecified range of differences considered sufficiently small, with analysis designed to assess whether the result falls within that range. The registry does not report an equivalence or non-inferiority design or a margin for such a judgment. Therefore, the results cannot be interpreted as demonstrating equivalence.

The record also leaves important details unresolved. It does not map each analysis explicitly to the named 12- and 24-month follow-ups, define the order of the group contrast, or explain why results labelled as final values accompany an endpoint described as change from baseline. It states that last observation carried forward was used for missing follow-up, but does not describe the amount of missing data or sensitivity analyses. Finally, the cluster-randomized design makes the sponsor’s own analysis more appropriate than a simple individual-level recalculation, while the detailed cluster-adjusted model is not reported.

These statistical findings concern the recorded weight endpoint and the analyses described in the registry. They do not establish effects on other health outcomes, participation, or quality of life; those results are not reported in the registry record. They also do not establish that either program works or fails. The defensible conclusion is narrower: the posted statistical results leave substantial uncertainty about the size and direction of any difference in final weight values between the randomized groups.

14. Why This Trial Matters Statistically

This study is statistically interesting because a pragmatic behavioral intervention is being compared with a routine-care program, while the design is described as cluster-randomized. That combination makes the assignment unit central to interpretation. If groups or settings are randomized, observations from participants sharing a cluster may be correlated. The analysis must reflect that structure to describe uncertainty appropriately; treating every person as independently assigned can produce misleading precision.

The trial also illustrates why endpoint definitions and analysis labels need to align. The primary outcome is described as change from baseline, but the posted tests are labelled mean differences in final values. The registry names two follow-up times without explicitly matching each result to a time point. These differences in wording do not invalidate the reported analyses, but they make it harder for readers to know precisely what estimate each result represents.

Missing outcomes create another interpretive issue. The record states that last observation carried forward was used for those lost to follow-up. This approach makes explicit that missing values were imputed, but the resulting estimate depends on whether carrying an earlier measurement forward is a reasonable assumption. Without information on missingness and sensitivity analyses, readers cannot assess how much that choice influences the findings.

The posted confidence intervals include zero, and the p-values do not provide strong evidence against a no-difference model under common testing conventions. At the same time, that does not establish that the programs are equivalent. The absence of an equivalence margin and the unresolved details about analysis time points underscore why statistical conclusions should be matched to the question actually studied. The trial’s record is a useful example of the value of reporting cluster methods, endpoint definitions, missing-data procedures, and the mapping between analyses and follow-up assessments.

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.

Several limitations specific to this trial follow from the registry record. First, the design is identified as cluster-randomized, but the number and nature of clusters and the handling of clustering in the sponsor’s statistical model are not reported. The record cautions against a simple individual-level re-analysis, yet does not give the detail needed to independently assess the reported variance estimates.

Second, the primary endpoint is described as change from baseline, while the two result entries are described as mean differences in final values. The registry does not explain the relationship between these descriptions, report whether baseline adjustment was used, or specify the subtraction order. The two results are also not explicitly linked to the outcome’s 12- and 24-month timeframes.

Third, the record reports last observation carried forward for participants lost to follow-up, but does not state how much follow-up was missing or provide sensitivity analyses. The method relies on the assumption that the last available observation is a suitable replacement for the missing measurement. The available record does not allow that assumption to be assessed.

Finally, the registry does not report a complete statistical analysis plan, multiplicity handling, interim analyses, or an equivalence margin. It does not provide detailed participant characteristics or sufficient adverse-event denominator information for interpreting the listed death percentages. The descriptions of intervention arms are incomplete in the record, limiting assessment of the exact intervention content and delivery. These omissions constrain interpretation; they do not establish that the study lacked procedures that are simply not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: MOVE OUT: A Partnership With Veterans Groups to Enhance Weight Management in the Veterans Health Administration (VHA) — NCT01339390.
• Registry record last updated 2016-04-28; status last verified 2016-03.
• Results first posted 2016-04-28.
• 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

MOVE OUT was a completed, randomized, parallel-group behavioral study sponsored by the VA Office of Research and Development. It compared MOVE OUT with MOVE! and enrolled 3169. The study start is recorded as 2012-04 and primary completion as 2015-03. Masking is listed as none, and the phase is not applicable.

The primary outcome was change in weight from baseline, measured in pounds at 12 and 24 months using medical-record measurements within a 6-month period around follow-up. The registry reports two two-sided t-test analyses of mean final values. Both confidence intervals include zero; the reported p-values are 0.494 and 0.306. The entries are not explicitly assigned to a follow-up time.

The record describes the study as cluster-randomized and states that a simple individual-level re-analysis would be misleading. It reports last observation carried forward for participants lost to follow-up. Serious adverse-event counts are listed for both arms, while the record does not provide a clear interpretation of the denominators for the listed death percentages. Important analysis details, including the cluster-adjusted model, are not reported in the registry record.