The GEM (Goals for Eating and Moving) Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The GEM (Goals for Eating and Moving) Study (NCT03006328), a randomized na study. The primary outcome is Change in Weight From Baseline.

2. Clinical Question

The clinical question concerns a technology-assisted approach embedded in patient-centered medical homes, rather than a medication or procedure. The stated intervention combines the GEM tool with health coaching and coordination. The comparator is also an active form of support: participants were to receive weight-management handouts from health coaches. Thus, the intended contrast is between tailored, coordinated support and non-tailored educational materials within the study’s care setting.

The eligibility description defines the obesity population through body mass index thresholds and the presence or absence of an obesity-associated comorbidity. Further details about how eligibility was assessed, which comorbidities qualified, or whether other eligibility restrictions applied are not reported in the registry record. The record identifies change in weight as the primary outcome and gives baseline and Month 12 as the measurement time points. It does not report the observed mean changes, uncertainty around those changes, or the statistical comparison between groups. Consequently, the clinical question can be described, but the registry information does not establish the direction, magnitude, or precision of the primary outcome difference.

3. Trial Design

The registry describes a randomized, single-masked, parallel-group study with behavioral interventions. It lists NYU Langone Health as the sponsor and the study status as completed. Enrollment was 489. The listed study dates are 2017-11-20 for start and 2020-05-20 for primary completion. The record identifies the study as The GEM (Goals for Eating and Moving) Study and gives the official title, “Technology-Assisted Weight Management Intervention Within Patient-Centered Medical Homes: The GEM (Goals for Eating and Moving) Study.”

Participants were assigned to GEM or enhanced usual care. The GEM arm description includes assignment to the GEM tool and a health coach. The enhanced-usual-care arm description states that participants would receive non-tailored weight-management handouts by health coaches. Both interventions are classified as behavioral. The record states that masking was single, but does not identify which people were masked or describe how masking was maintained. Details of the randomization procedure, allocation concealment, and the unit of randomization are not reported in the registry record.

The study information also flags a cluster-randomized design. In a cluster-randomized trial, groups such as practices or care teams, rather than individual participants, are assigned to study conditions. Outcomes from participants in the same cluster can be more alike than outcomes from unrelated clusters. Therefore, treating every participant as an independent randomized unit can give misleading uncertainty estimates. The particular clusters, their number, and the method used to account for clustering are not reported in the registry record.

The primary outcome was change in weight from baseline to Month 12, expressed as a mean in kilograms with standard deviation as the dispersion measure. The registry lists four outcome measures and zero analyses. Numerical primary-outcome results, the analytical model, and the sponsor’s comparative estimates are not reported in the registry record. The phase is listed as not applicable. Further details about recruitment, follow-up, adherence, and departures from the assigned interventions are also not reported in the registry record.

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

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 Baseline, Month 12.

The primary endpoint is change in weight from baseline to Month 12, measured in kilograms. It is a continuous outcome: each participant’s weight can take a range of values, and the study’s registered summary measure is the mean. The standard deviation is the listed measure of dispersion, describing how spread out individual measurements are around the mean.

Weight change is used to summarize how the measured outcome changes over the specified follow-up period. Comparing mean change between randomized groups can estimate the average difference associated with assignment to the two study approaches, provided the analysis accounts appropriately for the trial design and missing observations. The endpoint alone does not describe every aspect of weight-management care, and a group average does not show how much individual participants varied.

The registry specifies baseline and Month 12 as the time points, but does not report the precise measurement procedures, the definition used for calculating change, or the handling of participants without a Month 12 measurement. Those details are not reported in the registry record. Numerical means, standard deviations, a between-group estimate, and its confidence interval or significance test are also not reported in the registry record. Therefore, the endpoint’s role and measurement period are known, while the observed result and its precision cannot be assessed from the registry information.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Randomization is intended to make groups comparable, on average, with respect to factors that could influence outcomes. In a randomized study, the main comparison is generally based on outcomes between assigned groups. General statistical explanation: random assignment does not guarantee identical groups, especially in a particular trial, and it does not remove the need to account for the actual design. The registry identifies randomization but does not report the sequence-generation procedure, concealment method, or detailed assignment process.

Why does the cluster design matter?

In cluster randomization, assignment occurs at the level of a group rather than an individual. People within a cluster may share clinicians, care routines, location, or other influences. Their outcomes can therefore be correlated. General statistical explanation: when within-cluster correlation is present, an analysis that treats all individuals as fully independent can make standard errors too small and confidence intervals too narrow. Analyses commonly account for clustering through a suitable model or variance adjustment, but the specific method used in this study is not reported in the registry record.

The study information cautions that simple individual-level re-analysis would be misleading and says only the sponsor’s own analysis should be shown. This is an important constraint on interpretation. An analyst cannot reconstruct a valid cluster-adjusted comparison from aggregate outcome statistics alone unless the information required by the method is available. No cluster-level structure or sponsor estimate is reported in the registry record.

What does the mean and standard deviation describe?

The registered outcome is a continuous measure summarized by its mean and standard deviation. The mean is the arithmetic average of the measured changes in a group. The standard deviation describes the spread of participant-level values around that average. It is not a measure of uncertainty in the group mean; that uncertainty is usually conveyed by a standard error or confidence interval. General statistical explanation: two groups can have similar means but different variability, or different means with substantial overlap in their individual observations. The group means and standard deviations for the primary outcome are not reported in the registry record.

A between-group estimate would express the contrast in average change between the randomized groups. A confidence interval would show the statistical uncertainty around that contrast under the chosen analysis. A p-value, if reported, would summarize how compatible the data are with a specified null hypothesis under the model. None of these primary-outcome quantities is reported in the registry record. Statistical significance and clinical importance are also different concepts: a p-value alone does not communicate the size of an effect or whether it matters in practice.

How should safety counts be interpreted?

Safety summaries count participants or events according to defined categories and denominators. The registry presents serious-adverse-event and other-adverse-event counts by arm, along with deaths. These summaries are descriptive unless a comparative analysis is reported. General statistical explanation: raw counts depend on the number at risk and the period during which participants are observed. Interpreting between-arm differences also requires definitions of events, ascertainment procedures, and an appropriate analysis. Further details on these aspects are not reported in the registry record.

What can planning calculations show?

A planning sample-size calculation asks how many participants would be needed under specified assumptions to detect a chosen effect with a chosen statistical error rate and power. It is a prospective illustration, not evidence about what the trial observed. General statistical explanation: such calculations depend on the assumed effect, variability, allocation, analysis method, and—in a cluster-randomized study—the clustering structure. The registry’s planning assumptions specify a two-sided significance level of 0.05, equal allocation, a normal approximation, and a target effect expressed as a fraction of the observed effect. The record does not report numerical planning outputs or the trial’s own design justification. These assumptions should not be mistaken for the original protocol’s rationale or for observed power.

Power is a design quantity: under stated assumptions, it is the probability of detecting an effect of a specified size using a planned test. General statistical explanation: calculating “observed power” from an observed effect does not add useful information beyond the effect estimate and its uncertainty. Any sample-size illustrations based on an observed effect are planning illustrations, not observed power. In this case, the primary effect estimate is not reported in the registry record, so its magnitude and the precision of any effect-based planning illustration cannot be independently interpreted from the registered outcome information.

Finally, numerical analysis requires more than an endpoint name and an enrollment count. The analysis population, missing-data approach, model specification, and any adjustments for clustering or baseline values can affect estimates and uncertainty. These details are not reported in the registry record. Without them and without the sponsor’s numerical results, the statistical methods can be explained in general terms, but the primary analysis cannot be independently reconstructed.

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("Participants With Obesity + GEM", "Participants With Obesity + Enhanced Usual Care")
n   <- c(220, 269)
m   <- c(-1.4, -0.8)
sp  <- c(0.8, 0.9)
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  Participants With Obesity + GEM  220  -1.400  0.800  -1.506  -1.294
CARM  Participants With Obesity + Enhanced Usual Care  269  -0.800  0.900  -0.908  -0.692

11. Multiplicity, Missing Data and Analysis Populations

The registry lists four outcome measures and zero analyses. The available record does not identify which outcomes were treated as secondary, how outcome testing was ordered, or whether any adjustment for multiple comparisons was planned or used. Multiplicity matters when several outcomes or repeated tests are considered: the more opportunities there are to identify a statistically unusual result, the greater the chance of at least one such finding under a collection of null hypotheses. General statistical explanation: prespecified outcome hierarchies or suitable adjustment procedures can help control this issue. The registry record does not report whether either approach was used.

The primary outcome is described as change in weight from baseline to Month 12. The record does not report the number of participants with an observed Month 12 value, the number included in the primary analysis, or how missing measurements were handled. Missing outcomes can affect both the precision and the validity of a comparison, depending on why observations are missing and on the analysis assumptions. General statistical explanation: complete-case analysis, imputation, and likelihood-based approaches rely on different assumptions; none should be attributed to this trial without a report of the method.

The analysis population is also not reported in the registry record. In general, an intention-to-treat approach analyzes participants according to randomized assignment, while other analysis sets may impose additional conditions. The record does not state which population the sponsor used, how participants who changed or did not follow their assigned intervention were treated, or whether any sensitivity analyses were performed.

Because the trial is cluster-randomized, missingness and analysis populations may also need to be considered in relation to clusters. The record does not report cluster counts, cluster participation, or the handling of clusters with incomplete outcome data. These omissions prevent assessment of the analysis decisions and the robustness of the primary comparison.

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: foot fracture (1.9%), wrist/hand injury (1.9%), abdominal surgery (1.8%), chest pain (1.8%), gallbladder disease (1.5%).

The registry lists safety figures separately for Participants With Obesity + GEM and Participants With Obesity + Enhanced Usual Care. For the GEM arm, 220 were at risk for serious adverse events and the serious-adverse-event count was 53; 220 were at risk for other adverse events and the other-adverse-event count was 36. The listed number of deaths was 1. For enhanced usual care, 269 were at risk for serious adverse events and the serious-adverse-event count was 81; 269 were at risk for other adverse events and the other-adverse-event count was 43. The listed number of deaths was 3.

These are descriptive counts. They do not, by themselves, establish a causal difference between groups or account for differences in exposure time, event definitions, or follow-up. The registry also lists serious-event terms with percentages: Foot fracture, 1.9%; Wrist/hand injury, 1.9%; Abdominal Surgery, 1.8%; Chest Pain, 1.8%; Gallbladder Disease, 1.5%; and Back pain, 1.5%. The record does not give further context for interpreting these term percentages, including their denominators or group attribution.

General statistical explanation: safety counts are most informative when read with clear definitions, time-at-risk information, and details on how events were collected and classified. Such details and formal between-group safety analyses are not reported in the registry record. The figures should therefore be read as registered descriptive information, not as a standalone comparative safety conclusion.

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

The registered information establishes that the GEM Study was a randomized comparison of a technology-assisted, health-coach-supported intervention with enhanced usual care, and that the primary endpoint was mean weight change from baseline to Month 12. It also establishes that the study was completed, had an enrollment of 489, and was identified as cluster-randomized. These registry entries describe the question and design, but they do not provide the primary outcome result.

No numerical estimate of the difference in weight change, no confidence interval, and no p-value are reported in the registry record. Thus, the available information does not show the direction or magnitude of the primary outcome contrast, nor how precisely it was estimated. The design cannot substitute for results: randomization supports a comparative interpretation when the analysis is appropriate, but it does not indicate what the measured outcome was.

The result also cannot be inferred from the intervention descriptions. A tool intended to tailor care and coordinate services is a description of the intervention’s goals, not evidence of its effect on the endpoint. Likewise, completion status does not imply a favorable, unfavorable, or null finding. The serious adverse-event figures are descriptive safety counts and do not provide the missing primary efficacy estimate.

General statistical explanation: an estimate and its uncertainty answer different questions. The estimate describes the observed contrast; the interval conveys how much uncertainty surrounds it under the analysis assumptions. Neither should be replaced by a binary statement based only on a significance threshold. In this record, neither the estimate nor its uncertainty is available, so even that more nuanced interpretation is not possible.

The cluster-randomized design further limits what can be recovered from summary information. The study information cautions against a simple individual-level re-analysis and indicates that the sponsor’s own analysis takes precedence. The sponsor’s numerical primary analysis is not reported in the registry record. No conclusion about whether GEM changed average weight relative to enhanced usual care can therefore be drawn from the registered information alone.

14. Why This Trial Matters Statistically

Statistically, the trial is notable for asking an effectiveness question within patient-centered medical homes and for evaluating a behavioral intervention that combines technology and health coaching. Its comparator was not an absence of support: enhanced usual care included non-tailored weight-management handouts by health coaches. This makes the intended comparison one between two different forms of weight-management support.

The cluster-randomized design is central to interpreting any result. When groups such as care teams or practices are randomized, people sharing a group may have correlated outcomes. General statistical explanation: valid uncertainty estimates should reflect the assignment structure, rather than treating every participant as independently randomized. The study information explicitly warns that an individual-level re-analysis would be misleading and points to the sponsor’s analysis as the relevant one. This illustrates why study design details must guide analysis and why an aggregate participant count is not enough to reproduce a valid comparison.

The trial also illustrates the difference between a registered endpoint and a reportable result. The primary endpoint and time points are specified, but numerical group outcomes and a comparative estimate are not reported in the registry record. Without those results, readers cannot assess effect size or precision, regardless of how clearly the intervention and outcome are described.

Finally, the safety record presents counts by arm and lists selected serious-event terms, but does not provide the context needed for a full comparative assessment. The statistical value of a registry entry depends not only on its design information but also on complete reporting of analysis methods, denominators, missingness, and outcome estimates. Here, the design raises important analytical considerations, while the absence of primary numerical results limits what can be concluded statistically.

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.

A central limitation for interpreting this trial from the registry record is that the primary outcome results are not reported. Group means, standard deviations, the between-group estimate, and measures of uncertainty are unavailable. As a result, the magnitude, direction, and precision of the primary comparison cannot be assessed.

The cluster-randomized design is identified, but the record does not describe the clusters or report the cluster-adjusted analytical method. The study information cautions against a simple individual-level re-analysis and says only the sponsor’s analysis should be shown. The sponsor’s numerical analysis is not reported in the registry record, limiting independent assessment of the primary finding.

The record does not report the primary analysis population, the number contributing endpoint measurements, missing-data methods, or sensitivity analyses. These omissions leave uncertainty about how incomplete Month 12 measurements and departures from assigned interventions were addressed. The masking designation is single, but the masked parties and procedures are not described. Randomization procedures, allocation concealment, and measurement procedures are also not reported in the registry record.

The safety counts and serious-event terms are presented without detailed definitions, follow-up exposure, or a reported comparative safety analysis. Their denominators and the context for listed percentages are not fully described. The registry lists four outcome measures but does not report a multiplicity strategy or the analysis plan for outcomes beyond the primary endpoint. These reporting gaps constrain interpretation; they do not establish that the corresponding procedures were absent from the study.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: The GEM (Goals for Eating and Moving) Study — NCT03006328.
• Registry record last updated 2023-04-27; status last verified 2023-04.
• Results first posted 2023-04-27.
• 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 GEM Study (NCT03006328) is listed as a completed, randomized, single-masked, parallel-group behavioral trial sponsored by NYU Langone Health. It enrolled 489 participants and is identified as cluster-randomized. The listed start date is 2017-11-20 and primary completion date is 2020-05-20.

The study compared the GEM tool with health coaching against enhanced usual care consisting of non-tailored weight-management handouts by health coaches. The population description specifies body mass index of ≥30kg/m2 or body mass index of ≥25 kg/m2 with an obesity-associated comorbidity. The primary endpoint was mean change in weight, in kilograms, from baseline to Month 12, with standard deviation as the dispersion measure.

The record lists four outcome measures and zero analyses. Numerical primary-outcome results and details of the analysis population, missing-data handling, and cluster-adjusted methods are not reported in the registry record. Safety information includes serious and other adverse-event counts and deaths by arm, plus selected serious-event terms. These figures are descriptive and do not supply the missing primary comparison.