This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. View the MOVe-OUT registry record.
1. Trial at a Glance
MOVe-OUT was a randomized, parallel-group, unmasked trial conducted in the Veterans Health Administration. It compared MOVE OUT, a behavioral intervention, with MOVE!, also a behavioral intervention, with change in weight from baseline measured at 12 and 24 months as the registered primary endpoint.
| Feature | MOVe-OUT |
|---|---|
| Trial name | MOVe-OUT |
| ClinicalTrials.gov identifier | NCT01339390 |
| Therapeutic area | Infectious Disease |
| Conditions | Obesity; Overweight; Hypertension; Diabetes Mellitus |
| Phase | NA |
| Status | Completed |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 3,169 |
| Interventions | MOVE OUT (behavioral); MOVE! (behavioral) |
| Lead sponsor | VA Office of Research and Development |
| Sponsor type | FED |
2. Clinical Question
The registered trial question can be expressed as a randomized comparison of two behavioral approaches to weight management in the Veterans Health Administration. The primary outcome was change in weight from baseline, measured at 12 and 24 months.
Population
Eligible patients enrolled through the Veterans Health Administration, in a trial addressing obesity, overweight, hypertension, and diabetes mellitus.
Intervention
MOVE OUT, described in the registry as a behavioral intervention.
Comparator
MOVE!, also described in the registry as a behavioral intervention.
Primary question
Does MOVE OUT produce a different change in weight from baseline than MOVE! at the registered 12- and 24-month assessments?
3. Trial Design
MOVE OUT behavioral intervention
- Behavioral intervention
- Randomized comparison arm
- Serious adverse events: 32/1571 affected/at risk
MOVE! behavioral intervention
- Behavioral intervention
- Randomized comparison arm
- Serious adverse events: 29/1598 affected/at risk
4. Enrollment, Follow-Up, and Trial Context
The trial began in April 2012 and had a primary completion date of March 2015. A total of 3,169 participants were enrolled across the two randomized arms.
| Trial characteristic | Registry information |
|---|---|
| Start | 2012-04 |
| Primary completion | 2015-03 |
| Total enrollment | 3,169 |
| Number of arms | 2 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
The registry's stated limitation is particularly important for interpreting exposure to the intervention: because passive enrollment was permitted, all eligible patients identified by automated record review could be enrolled, while very few veterans randomized to the intervention actually used the program.
5. Primary Endpoint
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Change in Weight (From Baseline) | Weight as recorded in the medical record during a 6-month period around the follow-up point. | Measured at 12 and 24 months | Continuous; pounds |
The endpoint is a change score rather than a time-to-event outcome. That distinction determines the appropriate statistical framework: the registry reports a two-sided t-test and a mean difference rather than a hazard ratio or survival-analysis estimate.
6. Statistical Methodology
Reported comparison: two-sided t-test
The registry identifies the reported method as a t-test, 2 sided. The statistical method is a t-test, with the method category classified as linear models. The reported effect measure is a mean difference.
The registry's analysis notes explicitly state that the difference is expressed as the MOVE! group minus the MOVE OUT group. Therefore, the sign of the estimate must be interpreted using that direction.
Two-sided inference
A two-sided test evaluates departures in either direction from the null comparison. For a mean difference, the corresponding null value is zero: a mean difference of zero represents no average difference between the two randomized groups under the specified comparison.
Confidence intervals
Each posted primary analysis includes a 95% two-sided confidence interval. For a mean difference, the interval describes statistical uncertainty around the estimated difference under the analysis framework. Because both reported intervals include zero, the data are compatible with no average difference as well as with differences in either direction within the respective intervals.
Repeated measures and random effects in the analysis notes
The registry's analysis notes state that the analysis accounted for repeated measures and random effects. At the same time, the registry's formal method field identifies the method as a two-sided t-test. These statements should be kept distinct: the formal registry method field is the reported t-test, while the accompanying analysis text describes accounting for repeated measures and random effects. The available trial data do not provide enough additional model specification to reconstruct the exact repeated-measures or random-effects structure.
Missing follow-up observations
For the analysis, participants lost to follow-up were included using last observation carried forward to impute missing values. This is a missing-data assumption rather than a direct observation. Consequently, the reported estimates depend partly on how the imputation rule treats participants without an observed follow-up measurement.
7. Primary Results at 12 Months
The first posted primary analysis evaluates change in weight at 12 months relative to baseline. The comparison is MOVE! minus MOVE OUT.
Mean difference in change in weight
95% CI: −.7347 to 1.5229 · P = 0.4937
Two-sided t-test; mean difference expressed as MOVE! (Arm 2) minus MOVE OUT (Arm 1).
| 12-month analysis | Reported result |
|---|---|
| Outcome | Change in Weight (From Baseline) |
| Outcome unit | Pounds |
| Comparison | Arm 1: MOVE OUT vs Arm 2: MOVE! |
| Method | t-test, 2 sided |
| Effect measure | Mean Difference (Final Values) |
| Estimate | 0.3941 |
| 95% CI | −.7347 to 1.5229 |
| P-value | 0.4937 |
| Missing-data approach | Last observation carried forward for those lost to follow-up |
The estimated mean difference was 0.3941 pounds, with the difference defined as MOVE! minus MOVE OUT. Thus, the positive sign indicates that the estimated average change in weight was 0.3941 pounds higher in MOVE! than in MOVE OUT under the reported analysis.
This estimate does not mean that every participant differed by 0.3941 pounds, nor does it describe the weight change of an individual veteran. It is an average between-group difference under the specified analysis.
The 95% confidence interval extends from −.7347 to 1.5229 pounds. Because zero lies inside this interval, the interval includes no average difference as a statistically compatible value under the stated framework. The interval also illustrates the uncertainty around the point estimate: the data are compatible with a negative difference as well as a positive difference within the reported interval.
The P-value of 0.4937 does not measure the size of the treatment effect. A p-value describes the compatibility of the observed data with the null hypothesis under the specified testing framework. Effect magnitude and precision are better conveyed by the mean difference and its confidence interval.
Finally, interpretation is conditioned on the analysis population and missing-data rule. The registry states that last observation carried forward was used for participants lost to follow-up, and it also reports that very few veterans randomized to MOVE OUT used the program. These features are important when moving from the randomized comparison to a practical interpretation of intervention exposure.
8. Primary Results at 24 Months
The second posted primary analysis evaluates the same registered endpoint at 24 months relative to baseline.
Mean difference in change in weight
95% CI: −0.5395 to 1.7181 · P = 0.3062
Two-sided t-test; mean difference expressed as MOVE! (Arm 2) minus MOVE OUT (Arm 1).
| 24-month analysis | Reported result |
|---|---|
| Outcome | Change in Weight (From Baseline) |
| Outcome unit | Pounds |
| Comparison | Arm 1: MOVE OUT vs Arm 2: MOVE! |
| Method | t-test, 2 sided |
| Effect measure | Mean Difference (Final Values) |
| Estimate | 0.5893 |
| 95% CI | −0.5395 to 1.7181 |
| P-value | 0.3062 |
| Missing-data approach | Last observation carried forward for those lost to follow-up |
The estimated mean difference at 24 months was 0.5893 pounds, again defined as MOVE! minus MOVE OUT. The positive estimate therefore corresponds to a higher estimated average change in weight in MOVE! than MOVE OUT under the reported comparison.
The estimate should not be interpreted as a statement about individual participants. It is a between-group average difference and does not show how the distribution of individual weight changes behaved within either arm.
The 95% confidence interval runs from −0.5395 to 1.7181 pounds. Because the interval includes zero, the reported estimate is compatible with no average difference under the stated inferential framework. Its width also shows that the point estimate should not be treated as an exact quantity.
The P-value of 0.3062 is evidence about the statistical test, not a measure of the magnitude or clinical importance of the estimated difference. A larger or smaller p-value does not by itself tell us how large the underlying mean difference is.
The same missing-data consideration applies here: the analysis included participants lost to follow-up through last observation carried forward. The registry also notes low actual use of the intervention among veterans randomized to MOVE OUT, which is relevant when interpreting the randomized assignment effect.
9. Comparing the Two Posted Primary Analyses
| Feature | 12 months | 24 months |
|---|---|---|
| Endpoint | Change in Weight (From Baseline) | Change in Weight (From Baseline) |
| Estimate | 0.3941 pounds | 0.5893 pounds |
| 95% CI | −.7347 to 1.5229 | −0.5395 to 1.7181 |
| P-value | 0.4937 | 0.3062 |
| Test | t-test, 2 sided | t-test, 2 sided |
| Difference direction | MOVE! − MOVE OUT | MOVE! − MOVE OUT |
| Imputation | Last observation carried forward | Last observation carried forward |
The point estimates are both positive, but the two confidence intervals both include zero. The registry therefore provides two time-specific estimates of the same primary outcome rather than a single pooled effect. The available data do not provide a separate longitudinal effect estimate that should replace these reported analyses.
10. Safety Results
The registry reports serious adverse events by randomized arm using affected participants over those at risk.
| Safety measure | MOVE OUT | MOVE! |
|---|---|---|
| Serious adverse events | 32/1571 affected/at risk | 29/1598 affected/at risk |
These counts describe serious adverse events separately from the primary weight-change endpoint. They should not be combined with the efficacy estimate into a single numerical benefit-risk measure.
MOVE OUT
32 participants with serious adverse events among 1571 at risk.
MOVE!
29 participants with serious adverse events among 1598 at risk.
The ClinicalTrials.gov record does not provide a formal between-arm statistical analysis of serious adverse events. Accordingly, this page reports the registry counts without adding an unreported hypothesis test or derived safety conclusion.
11. Missing Data and Last Observation Carried Forward
Missing-data handling is central to this trial because the registry explicitly states how participants lost to follow-up were incorporated into the primary analysis.
The registry states that those lost to follow-up were included using the last observation carried forward to impute missing values.
What last observation carried forward does
Last observation carried forward, or LOCF, substitutes the most recent available observation for a later missing measurement. This allows a participant with missing follow-up information to remain in the analysis rather than being removed entirely.
Why the assumption matters
The method implicitly makes an assumption about what the missing outcome would have looked like. If the true unobserved weight change differs systematically from the carried-forward value, the resulting treatment comparison can differ from what would have been obtained using another missing-data strategy.
LOCF is not neutral merely because it preserves sample size
Retaining participants in an analysis does not remove the statistical consequences of the assumptions used to fill in missing values. A missing-data method should therefore be considered part of the estimand and analysis strategy rather than a purely technical preprocessing step.
12. Statistical Methods Explained
Why was a t-test used?
The registered primary outcome is continuous change in weight, measured in pounds. A t-test is a standard method for comparing the mean of a continuous outcome between two groups. Here, the registry specifically reports a two-sided t-test and a mean difference as the effect measure.
What does the mean difference of 0.3941 mean at 12 months?
The registry defines the direction as MOVE! minus MOVE OUT. Therefore, 0.3941 means that the estimated average change in weight was 0.3941 pounds higher in MOVE! than MOVE OUT under the reported 12-month analysis. It is not a statement that each participant differed by that amount.
Why is zero important for a mean difference?
For a difference between two group means, zero represents equality of the group means. A confidence interval containing zero therefore includes the possibility of no average difference under the statistical model and sampling framework used to construct the interval.
What does the 95% confidence interval tell us?
The 12-month interval, −.7347 to 1.5229, and the 24-month interval, −0.5395 to 1.7181, quantify uncertainty around the corresponding mean-difference estimates. They are not ranges containing 95% of individual participants' weight changes, nor are they predictions for individual patients.
Why doesn't the p-value measure effect size?
A p-value evaluates the compatibility of the observed data with a null hypothesis under a specified testing procedure. It does not tell us how large the treatment difference is. The mean difference provides the effect estimate, while the confidence interval provides information about its precision.
Why does the missing-data method matter?
The primary analysis includes participants lost to follow-up using last observation carried forward. The resulting estimate therefore reflects both observed data and the assumptions embedded in that imputation rule. A different missing-data strategy could potentially produce a different estimate.
Why should the reported t-test and repeated-measures language be read together?
The registry's formal method field says "t-test, 2 sided," while the analysis notes say that the analysis accounted for repeated measures and random effects. This indicates that the registry record contains more than one level of methodological description.
13. Randomization and What It Contributes
Randomization is one of the central statistical features of MOVe-OUT. Participants were assigned to two parallel behavioral-intervention arms rather than choosing the intervention themselves.
Randomization does not guarantee that every participant uses the assigned intervention. The registry specifically notes that very few veterans randomized to the intervention used the program. That limitation is particularly relevant when distinguishing the effect of being assigned to MOVE OUT from the effect of actually engaging with MOVE OUT.
14. No Masking: Statistical Implications
The registry identifies masking as none. In an unmasked behavioral trial, participants and study personnel can know which intervention has been assigned.
For an objectively recorded outcome such as weight recorded in the medical record, lack of masking does not automatically invalidate the outcome comparison. However, knowledge of assignment can influence behavior, participation, follow-up, and other aspects of trial conduct. The ClinicalTrials.gov record does not quantify those mechanisms, so they should be recognized as design considerations rather than assigned numerical effects.
15. Repeated Measurements at 12 and 24 Months
The primary endpoint was measured at both 12 and 24 months. The registry analysis notes also state that the analysis accounted for repeated measures and random effects.
| Assessment | Primary outcome | Reported analysis |
|---|---|---|
| 12 months | Change in Weight (From Baseline) | Two-sided t-test; mean difference 0.3941 pounds |
| 24 months | Change in Weight (From Baseline) | Two-sided t-test; mean difference 0.5893 pounds |
Repeated measurements create correlation because observations from the same participant are not independent. The registry's analysis notes explicitly recognize this issue by stating that repeated measures and random effects were accounted for. The exact model specification, however, is not provided in the ClinicalTrials.gov record.
A statistical model that accounts for within-participant correlation can distinguish repeated observations from independent observations across different participants. The ClinicalTrials.gov record states that repeated measures and random effects were accounted for, but does not provide enough information to reproduce the exact model.
16. Multiplicity and the Two Time Frames
The registry identifies one registered primary endpoint measured at two time points: 12 and 24 months. Two formal statistical analyses are posted for that endpoint.
| Feature | Registry information |
|---|---|
| Registered primary endpoints | 1 |
| Primary endpoint | Change in Weight (From Baseline) |
| Time frames | 12 and 24 months |
| Statistical analyses posted | 2 |
| Primary-endpoint analyses | 2 |
| Analyses with estimate + CI | 2 |
| Hypothesis type | Superiority |
The ClinicalTrials.gov record does not state an alpha-allocation procedure, hierarchical testing strategy, or other formal multiplicity adjustment for the two time points. Therefore, this page does not assign a familywise error interpretation that is not documented in the registry data.
17. Non-Inferiority, Bayesian Methods, and Interim Analysis
The ClinicalTrials.gov record classifies the hypothesis type as superiority. No non-inferiority margin is reported, so there is no non-inferiority framework to apply to the primary analysis.
The trial data also provide no Bayesian analysis, Bayesian posterior probability, prior distribution, or Bayesian decision rule. Accordingly, the primary results are interpreted using the reported frequentist t-test, mean difference, confidence interval, and p-value.
No interim analysis procedure is reported in the ClinicalTrials.gov record. This page therefore does not attribute an interim-monitoring boundary, alpha-spending method, or stopping rule to MOVe-OUT.
18. What the Mean Difference Does — and Does Not — Mean
The 12-month mean difference of 0.3941 pounds is the estimated average difference in change in weight between the two randomized groups, with the direction defined as MOVE! minus MOVE OUT.
It does not mean that the intervention caused every participant to gain or lose exactly 0.3941 pounds, nor does it describe the distribution of individual responses.
The 95% confidence interval of −.7347 to 1.5229 pounds shows that the point estimate is uncertain. The interval includes zero, so no average difference is among the values excluded by the reported 95% confidence interval.
The 24-month estimate of 0.5893 pounds is interpreted in the same direction: MOVE! minus MOVE OUT. Its 95% confidence interval, −0.5395 to 1.7181 pounds, also includes zero.
The reported p-values, 0.4937 at 12 months and 0.3062 at 24 months, quantify evidence against the corresponding null hypothesis under the reported two-sided testing procedure. They do not quantify clinical importance, probability that one intervention is better, or the probability that the null hypothesis is true.
19. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The posted primary analyses estimate mean differences of 0.3941 pounds at 12 months and 0.5893 pounds at 24 months, with both 95% confidence intervals containing zero.
Trial-context interpretation
The registry reports that very few veterans randomized to MOVE OUT used the program. This limits how directly an assignment-level result can be interpreted as the effect of actual program use.
These two perspectives should remain separate. The statistical result describes the randomized comparison under the reported analysis. The intervention-use limitation describes an important aspect of how that comparison maps onto actual exposure to MOVE OUT.
20. Important Limitations and Interpretation Issues
- Low intervention use: the registry states that very few veterans randomized to the intervention used the program. This is an important limitation for interpreting the effect of assignment to MOVE OUT as the effect of actual program participation.
- Passive enrollment: the registry states that passive enrollment was allowed by the Institutional Review Board, permitting enrollment of all eligible patients identified by automated record review.
- Unmasked design: masking was none. Knowledge of assignment can affect behavior, engagement, follow-up, or other trial processes, although the ClinicalTrials.gov record does not quantify such effects.
- Missing-data assumptions: participants lost to follow-up were incorporated using last observation carried forward. This makes the primary estimate dependent on the assumptions of the imputation approach.
- Repeated measurements: the endpoint was assessed at 12 and 24 months, and the analysis notes state that repeated measures and random effects were accounted for. The ClinicalTrials.gov record does not specify the full model structure.
- Multiplicity: the trial reports two analyses of the primary endpoint at different time points, but the ClinicalTrials.gov record does not document a multiplicity-control procedure.
- Safety information: serious adverse events are reported as affected/at-risk counts, but the ClinicalTrials.gov record does not provide a formal between-arm statistical analysis of those events.
- Registry detail: the formal method field identifies a two-sided t-test, while the analysis notes provide additional language about repeated measures and random effects. The exact underlying model cannot be reconstructed from the ClinicalTrials.gov record alone.
21. Why This Trial Matters Statistically
MOVe-OUT is a useful teaching case because it illustrates a different statistical structure from trials dominated by time-to-event endpoints. The primary outcome is a continuous change score, analyzed as a mean difference, with two follow-up time points and an explicit missing-data strategy.
| Concept | How it appears in MOVe-OUT |
|---|---|
| Randomization | Participants were randomized to two parallel behavioral-intervention arms. |
| Continuous endpoint | The primary endpoint was change in weight from baseline, measured in pounds. |
| Mean difference | The effect measure was the difference between the two groups' mean changes. |
| Two-sided testing | The registry reports a two-sided t-test for both primary analyses. |
| Confidence interval | Each primary analysis provides a 95% two-sided confidence interval. |
| P-value | Each primary analysis provides a p-value associated with the reported t-test. |
| Repeated measures | The analysis notes state that repeated measures and random effects were accounted for. |
| Missing data | Participants lost to follow-up were included using last observation carried forward. |
| Multiple time points | The primary endpoint was measured at both 12 and 24 months. |
| Unmasked design | The registry identifies masking as none. |
| Implementation limitation | The registry reports that very few veterans randomized to the intervention used the program. |
22. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
23. Related Statistical Calculators
24. Sources
- ClinicalTrials.gov: MOVe-OUT, NCT01339390.
Continue with Clinical Biostats statistical methods
Explore tutorials and statistical calculators covering the methods used to interpret randomized clinical-trial data.
25. Record Summary
MOVe-OUT provides a useful example of statistical analysis for a randomized behavioral intervention with a continuous primary endpoint. The registry reports change in weight from baseline at 12 and 24 months, analyzed using a two-sided t-test with mean difference as the effect measure. The reported difference was defined as MOVE! minus MOVE OUT, with an estimate of 0.3941 pounds at 12 months and 0.5893 pounds at 24 months. The corresponding 95% confidence intervals were −.7347 to 1.5229 and −0.5395 to 1.7181, with p-values of 0.4937 and 0.3062, respectively.
The statistical interpretation must also incorporate the trial's analysis choices. Participants lost to follow-up were included using last observation carried forward, while the analysis notes state that repeated measures and random effects were accounted for. The trial was unmasked, and the registry specifically reports that very few veterans randomized to MOVE OUT used the program. These features do not change the numerical estimates reported by the registry, but they matter when determining what those estimates can and cannot establish.