Intervening in Diabetes With Healthy Eating, Activity, and Linkages To Healthcare - The I-D-HEALTH Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Intervening in Diabetes With Healthy Eating, Activity, and Linkages To Healthcare - The I-D-HEALTH Study (NCT01435603), a randomized na study. The primary outcome is Percent Change in Body Weight.

2. Clinical Question

This is a comparison of two behavioral approaches, not a comparison of lifestyle support with no contact. Both groups received clinical education and brief lifestyle advice. The contrast is the addition of access to the intensive group-based intervention in one arm. The primary outcome is a continuous measure, so the reported group averages describe the mean percentage change rather than the proportion of participants who lost weight or reached a particular target. The record does not report a threshold defining a clinically important difference.

3. Trial Design

The registry describes a randomized, single-masked, parallel-group study. Its stated purpose was to compare ways of giving advice and providing support to improve diet and physical activity in adult primary care patients with elevated body mass index and dysglycemia. The record lists the study phase as NA and the status as completed. The sponsor was Northwestern University.

The two behavioral groups shared core components. Standard lifestyle advice involved primary care-based identification of pre-diabetes and/or type 2 diabetes, standard clinical education offered by participants’ usual primary care providers, and brief lifestyle advice delivered by a study Research Assistant. The advice-plus-intervention group received those components as well as access to an intensive group-based lifestyle intervention. The record does not give further details of the intervention’s content or delivery in the available arm description.

Enrollment is listed as 331. For the primary endpoint, the registry reports 128 participants in the standard lifestyle advice group and 129 in the advice-plus-intervention group. These endpoint counts refer to participants with the required weight measurements, rather than necessarily describing everyone enrolled or assigned. The record does not report the number randomized to each group, the number who began each intervention, or a participant flow account.

The study began in 2011-08, and primary completion is listed as 2016-09. The registry record does not report the masking procedures, the identities of people masked, or whether masking continued through outcome assessment and analysis. It also does not report the original sample-size justification, planned recruitment target by group, or the trial’s original statistical analysis plan. The stated primary endpoint timeframe is baseline to 12 months.

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: Potential participants were prescreened on the phone and attended a baseline visit for further screening and enrollment.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Participants who completed weight measurements at the baseline and 12 month exam

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 Percent Change in Body Weight, measured over Baseline to 12 months.

The primary endpoint is percent change in body weight from baseline to 12 months, summarized as a mean. Its definition is baseline body weight minus body weight at 12 months, divided by baseline body weight. Under this definition, negative values indicate weight loss, while positive values would indicate an increase. The registry identifies the endpoint population as participants who completed weight measurements at both the baseline and 12 month exam.

A percentage-change measure expresses change relative to each participant’s baseline weight, rather than as an absolute change in weight units. Summarizing this continuous measure with a mean allows a comparison of average change between the randomized groups. It does not show how changes were distributed among individuals, whether most participants changed in the same direction, or how many met a particular threshold.

Body weight was selected as the primary endpoint in a study focused on lifestyle advice and support for patients with elevated body mass index and dysglycemia. The registry also lists secondary outcomes concerning costs, health-state utility, cholesterol, A1c, blood pressure, dietary composition, and physical activity. The primary endpoint alone does not establish what happened on those other outcomes. The record does not report a clinically important difference threshold for percent change in body weight.

7. Results Posted in the Registry

• Advice Plus Lifestyle Intervention: mean -1.42 versus -0.15 on control; difference -1.27 (95% CI -2.31 to -0.23), standardized effect d = -0.30; z-test p = 0.017.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant has a percentage change in body weight, and the registry summarizes the groups using means and standard deviations. A mean difference compares the average value in one group with the average value in the other. Here, the posted comparison is advice plus lifestyle intervention minus standard lifestyle advice. A negative contrast means that the mean change was more negative in the intervention group. Because the endpoint definition makes negative changes correspond to weight loss, the contrast’s direction indicates a more negative mean percentage change in that group.

What does the reported regression result represent?

The registry reports a linear regression analysis with a slope of -1.273, a p-value of 0.017, and a superiority label. A regression coefficient summarizes a modeled difference associated with the group comparison. The record does not report the model’s covariates, adjustment strategy, or confidence interval. It is therefore not possible to determine from the registry record whether the slope is adjusted or to reproduce the full sponsor analysis from the posted summary statistics alone.

The recalculated comparison uses the two group means and their standard deviations as an unadjusted comparison. It gives a mean difference of -1.27 and p = 0.017. The recalculation is not a replacement for the sponsor’s analysis. It is a check based on the posted summary values, without any unreported covariate adjustments or other modeling choices. Since the sponsor’s analysis is reported and there is no stated discrepancy, its result remains the primary analysis to interpret.

What does a p-value tell the reader?

In general, a p-value measures how surprising data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the model assumptions were true. It is not the chance that the null hypothesis is correct, and it does not measure the size or importance of an effect. The posted p-value is 0.017. Its interpretation depends on the analysis and assumptions, including how the comparison was specified. The registry record does not report whether the primary p-value was adjusted for testing multiple outcomes or analyses.

Why is the standard deviation relevant?

In general, a standard deviation describes how spread out individual observations are around their group mean. The reported standard deviations are 4.05 for standard lifestyle advice and 4.45 for advice plus intervention. These summarize variation in participants’ percentage changes; they are not standard errors or confidence limits for the means. Standard errors quantify uncertainty in an estimated average or contrast and depend on both variability and the amount of information available. The registry also reports standard errors for the two group means, but they do not replace an interval for the between-group effect.

In general, over-dispersion means that observed variability is greater than a model expects. When that occurs, model-based uncertainty can be understated if the model does not accommodate the extra variation. The posted endpoint is continuous and the registry gives group standard deviations, but it does not report an over-dispersion assessment or identify this as a concern. No claim about over-dispersion in this study can be made from the registry record.

How should the confidence interval be read?

In general, a confidence interval describes a range of effect values reasonably compatible with the data and the method used to calculate the interval. It conveys uncertainty about the effect estimate, rather than the range of individual participants’ outcomes. The recalculated interval extends from -2.31 to a value close to zero. Its endpoints indicate that the unadjusted estimate is compatible with a more negative difference as well as a difference near zero. It does not include zero. The registry’s own regression result has no confidence limits reported, so the recalculated interval should not be mistaken for an interval from that model.

The registry’s endpoint population is defined by completion of weight measurements at both visits. This is an analysis restriction with implications for interpretation: it describes the observed endpoint results among those with the required measurements. The record does not report missing-data methods, imputation, sensitivity analyses, or the population used in the sponsor’s regression. Those omissions prevent a full assessment of how missing outcomes may affect the estimate.

10. Confidence Intervals and Precision

The recalculated unadjusted mean difference is -1.27, with a confidence interval from -2.31 to a value close to zero. The interval gives a sense of the range of group differences compatible with the summary data under the calculation used. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty about the effect’s size. The interval is not a range containing a specified proportion of individual participants’ changes.

The interval excludes zero, but its upper endpoint is close to zero. This means the recalculated result supports a negative mean difference while leaving uncertainty about how large that difference is. It is not enough to focus only on whether zero falls inside or outside the interval: the range and the context of the endpoint also matter. The registry does not report a clinical importance threshold, so the interval cannot be judged against a registry-stated threshold for meaningful change.

This interval is an unadjusted recalculation from the posted group summaries. It is distinct from the sponsor’s linear regression result, for which the registry reports no confidence limits. The reported regression slope is -1.273 and p = 0.017. The similarity of the estimates and p-values is reassuring as a numerical comparison, but does not make their methods identical. The registry record does not report further information about the regression model or the uncertainty calculation used by the sponsor.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrations based on the observed mean difference, not the trial’s original design assumptions. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row uses the stated fraction of the observed effect as the effect to detect. These calculations are participants per arm, not total participants or total events.

For planning for 100% of the observed effect, the figures are 177 participants per arm for 80% power and 236 participants per arm for 90% power. For planning for 75% of the observed effect, they are 314 participants per arm for 80% power and 420 participants per arm for 90% power. For planning for 50% of the observed effect, they are 705 participants per arm for 80% power and 944 participants per arm for 90% power.

In general, power is the chance that a specified statistical test will detect an effect of the chosen size if that effect is present and the assumptions hold. The 80% power column gives the per-arm planning figure associated with an 80% chance of detection under those assumptions; the 90% column uses a higher target chance of detection. Smaller effects need more participants to detect with the same power target, which is why the planning figures increase as the assumed fraction of the observed effect gets smaller. These values are not observed power and do not describe the probability that the trial’s result is true.

The sample-size assumptions for the trial’s original design, including any original target and rationale, are not reported in the registry record. The planning figures should not be read as evidence that the study was designed around its observed result. They are forward-looking illustrations under stated assumptions, not a retrospective measure of the trial’s success.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.270; pooled standard deviation σ = 4.255.
2. n = 2 × (1.96 + 0.842)² × 4.255² / -1.270² = 176.2, rounded up to 177 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.27                      # observed difference in means
sigma <- 4.25548665450459                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1435603)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 177 (80% power), 236 (90% power)
 75% of observed effect: n per arm = 314 (80% power), 420 (90% power)
 50% of observed effect: n per arm = 705 (80% power), 944 (90% power)
power.t.test: 177.2 
Simulated power at n = 177 per arm: 0.799 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor’s own analysis is reported as a linear regression with a slope of -1.273 and p = 0.017, and is labelled as a superiority analysis. The registry does not report confidence limits for this analysis or the regression model’s covariates and other details. That result is distinct from the unadjusted recalculation, which uses the posted group means and standard deviations and gives a mean difference of -1.27 with p = 0.017.

The recalculation is close to the sponsor’s reported estimate and p-value, but it does not reconstruct the regression model. The registry record does not identify any discrepancy between the sponsor analysis and the recalculation, so the sponsor’s reported analysis takes precedence. The recalculation is best understood as a basic consistency check from summary statistics, rather than a substitute or an alternative primary analysis.

For interpretation, the sponsor’s slope indicates a negative group contrast on the percentage-change scale, and its p-value is the registry’s reported test result. The confidence interval calculated from the summaries describes uncertainty for the unadjusted contrast only. It should not be attributed to the sponsor’s regression analysis, whose interval is not reported in the registry record.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists eight secondary outcomes: incremental costs, changes in health-state utility, percent change in body weight, percent change in blood total cholesterol, percent change in A1c, percent change in blood pressure, percent change in dietary composition, and percent change in physical activity. It also reports one analysis for the primary result. The record does not report a multiplicity adjustment, a hierarchy among outcomes, or a strategy for controlling the chance of false-positive findings across multiple tests.

In general, testing multiple outcomes or multiple contrasts can increase the chance that at least one result appears statistically notable by chance. A pre-specified testing hierarchy or adjustment can address that issue, but the registry record does not say whether one was planned or applied. The primary result should therefore be interpreted as the reported endpoint comparison, without assuming that it was protected from multiplicity by an unreported procedure. The record does not provide enough information to assess multiplicity across all analyses that may have been performed.

The primary endpoint population includes participants who completed weight measurements at baseline and the 12 month exam. The registry record does not report the amount or pattern of missing endpoint data, reasons for missingness, or the methods used to handle missing values. It does not report whether imputation, a missing-data model, or sensitivity analyses were used. Consequently, the potential influence of missing measurements on the estimate cannot be assessed from the record.

The registry does not report whether the primary analysis included all randomized participants, followed an intention-to-treat principle, or used another analysis population. The endpoint counts are reported for participants with required measurements, but the number randomized to each group and the relationship between randomization and endpoint analysis are not stated. These details limit assessment of whether the analysis preserves the original randomized comparison.

The registry also does not report a full statistical analysis plan, the covariates in the sponsor’s linear regression, or the rationale for any exclusions. These are limitations of the available record, not evidence that particular procedures were or were not used. A more complete report would be needed to evaluate those methods.

15. 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 5% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: hypoglycemia (severe: admitted to hospital, emergency room, or needed assistance by other person) (2.4%), balloon angioplasty or stenting of coronary arteries (percutaneous transluminal coronary angioplasty (1.8%), fracture (1.8%), heart attack (myocardial infarction) (0.6%), congestive heart failure (0.6%).

The registry reports serious adverse events for both groups. In standard lifestyle advice, 10 serious adverse events are reported among 167 at risk; in advice plus lifestyle intervention, 10 are reported among 164 at risk. Deaths are reported as 1 in the standard advice group and 0 in the advice-plus-intervention group. The registry reports 0 other adverse events in each group, with 167 and 164 at risk, respectively.

The listed serious-event terms include severe hypoglycemia requiring hospital admission, an emergency-room visit, or assistance from another person at 2.4%; balloon angioplasty or stenting of coronary arteries at 1.8%; fracture at 1.8%; heart attack at 0.6%; congestive heart failure at 0.6%; and stroke at 0.6%. The registry presents these as serious-event terms and percentages, without giving a group assignment for each term in this list. They should not be assigned to a particular arm from the information reported.

These figures are descriptive safety counts in the registry record. The record does not provide enough detail to establish whether the figures represent events or participants consistently, how events were adjudicated, or the duration of safety follow-up. The numbers alone do not establish that a treatment caused or prevented an event. They also do not provide a complete comparison of harms beyond the listed registry data.

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

The primary result describes an average difference in percent change in body weight from baseline to 12 months among participants with measurements at both visits. The advice-plus-intervention group had a more negative mean change than the standard-advice group, and the sponsor’s reported regression result is a negative slope with p = 0.017. The unadjusted recalculation is similar in direction and magnitude. Together these statistics describe the posted endpoint comparison, not every participant’s outcome.

The result does not establish that every participant lost weight, or that the intervention caused a particular individual’s change. The group standard deviations show variation around the means, while the registry does not report individual outcome distributions or the proportion of participants who gained or lost weight. It also does not state a threshold for a clinically important difference. Statistical evidence against a null comparison is not, by itself, evidence that the magnitude is clinically important.

The primary result does not answer whether the groups differed in the listed secondary outcomes, including A1c, blood pressure, cholesterol, diet, physical activity, costs, or health-state utility. Those outcomes require their own results and uncertainty estimates. Nor does the result establish effects beyond the baseline-to-12-month timeframe reported for the primary endpoint.

Interpretation is also limited by missing information about the analysis population, missing measurements, regression adjustments, and multiplicity. The endpoint population consists of participants with measurements at both visits, and the record does not explain how participants without these measurements were handled. Such omissions do not negate the posted estimate, but they limit assessment of its robustness and its applicability to everyone enrolled or randomized.

The safety record reports serious adverse events and deaths in both groups, as well as the listed serious-event terms. Those entries are descriptive and do not establish comparative causation. Overall, the registry supports a precise description of the reported mean comparison and its statistical uncertainty, while leaving clinical importance, robustness to missing data, and broader effects unresolved.

17. Why This Trial Matters Statistically

This study illustrates the distinction between a mean difference and an individual response. The endpoint is a continuous percentage-change measure, and the reported contrast concerns group averages. A mean can communicate the direction and central magnitude of a comparison, but it cannot by itself show how consistently participants changed or how much their outcomes varied. The reported standard deviations provide a reminder that participant-level changes vary around the means.

It also shows why both an estimate and its uncertainty matter. The sponsor reports a regression slope and p-value, while the summary-statistic recalculation gives a similar mean difference and p-value with an interval that reaches close to zero at its upper end. A p-value alone does not describe the plausible range of effect sizes. The interval contributes information about precision, while still not answering whether the difference meets a clinically important threshold.

The comparison is between two active behavioral approaches: both groups received standard clinical education and brief lifestyle advice, while one group had access to an intensive group-based intervention. This makes it important to interpret the contrast as the difference between these approaches, rather than as an intervention-versus-no-support comparison. The trial’s description and endpoint population also matter: the primary result is for those with weight measurements at both specified visits.

Finally, the record highlights the value of transparent methods reporting. The original sample-size assumptions, missing-data procedures, analysis population, multiplicity strategy, and regression details are not reported. These omissions make it harder to judge design intent and robustness, even though the primary estimate and sponsor p-value are available. Statistical interpretation depends not only on the numerical result but also on understanding how the analysis was planned and conducted.

18. 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 are specific to the information reported for this trial. The primary endpoint is defined only for participants completing weight measurements at baseline and the 12 month exam. The registry does not state how many randomized participants lacked one or both measurements, why data were missing, or how missingness was addressed. It is therefore not possible to assess whether the complete-measurement endpoint population differed from those without endpoint data.

The registry gives enrollment and endpoint group counts, but does not report the number randomized to each group or a participant flow account. It does not describe adherence, participation in the group-based intervention, or exposure to other lifestyle support. The available intervention description also ends without further detail about the intensive group-based program. These omissions limit understanding of intervention delivery and engagement.

The sponsor’s linear regression is reported with a slope and p-value but no confidence limits, model covariates, or fuller specification. The registry does not report the original sample-size justification, interim analyses, a full statistical analysis plan, multiplicity procedures, or sensitivity analyses. The registry record does not report these details, so their use should not be assumed either way.

The primary result concerns body-weight change over the stated baseline-to-12-month timeframe. The record lists secondary outcomes, but the available results do not describe them. It also provides no registry-stated threshold for clinical importance. Safety entries are reported, but the record does not explain event adjudication, event-level versus participant-level counting, or the duration of safety follow-up. Each of these gaps narrows what can be concluded from the posted result.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Intervening in Diabetes With Healthy Eating, Activity, and Linkages To Healthcare - The I-D-HEALTH Study — NCT01435603.
• Registry record last updated 2018-04-27; status last verified 2016-10.
• Results first posted 2018-04-27.
• 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.

22. Record Summary

NCT01435603 is a completed randomized, single-masked, parallel-group study sponsored by Northwestern University. It compared standard lifestyle advice with the same core advice and clinical education plus access to an intensive group-based lifestyle intervention. Enrollment is listed as 331, and the study began in 2011-08, with primary completion in 2016-09.

The primary endpoint was mean percent change in body weight from baseline to 12 months among participants with measurements at both visits. The posted means were -0.15 for standard lifestyle advice and -1.42 for advice plus intervention; the posted mean difference was -1.27. The sponsor reports a regression slope of -1.273 and p = 0.017. Further information about missing data, analysis populations, multiplicity, original design assumptions, and regression details is not reported in the registry record.