Validity of the PortionSize App to Measure Children's Food Intake: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Validity of the PortionSize App to Measure Children's Food Intake (NCT05587816), a na na study. The primary outcome is Estimation of Energy Intake (kcal) by Children Using the PortionSize App.

2. Clinical Question

3. Trial Design

The registry record describes a single-group device evaluation titled “Validity of the PortionSize App to Measure Children's Food Intake.” The intervention is training and use of the PortionSize app by children and parents or caregivers. Use is described in the clinic of Pennington Biomedical Research Center and at home under free-living conditions; participants also rate satisfaction with the app. The record states that the study is not a randomized comparison, and allocation is listed as NA. Masking is listed as none.

The registered enrollment is 31. The study status is completed, with a start date of 2022-10-28 and a primary completion date of 2023-03-06. The sponsor is Pennington Biomedical Research Center. The record lists nutritional assessment as the condition. Further information about recruitment, eligibility, participant characteristics, adherence, or the conduct of the clinic and home activities is not reported in the registry record.

The primary measure concerns children's energy intake estimated using the app. Its stated timeframe is one study visit of approximately 2.5 hours. The description says app estimates are compared with estimates from weighed meals, the criterion measure. That comparison concerns measurement validity and agreement, rather than a comparison of health outcomes between randomized treatment groups. Secondary outcomes concern user satisfaction with the app for recording food intake by children and by parents or caregivers. Numerical findings for these measures are not reported in the registry record.

The registry does not describe a control arm or a between-group treatment contrast. Its single-group structure is important when interpreting any planning illustration that assumes equal allocation between two arms: that assumption does not describe the registered design. The reason for the study's design, and any protocol-level sample-size justification, are 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.

Before assignment: Screening and enrollment were conducted at Pennington Biomedical Research Center. * Dyads enrolled: 31 * Children included: 31 * Caregivers/parents included: 31 - Among 31 parents, 21 were unique, as nine had a second child who participated in the study. * No randomization; single group assignment. * Intervention: training on and use of the PortionSize app by children and by caregivers/parents to log children's meals; validation in clinic and free-living use at home.

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 Estimation of Energy Intake (kcal) by Children Using the PortionSize App, measured over One study visit of ~ 2.5 hours.

The primary endpoint is estimation of energy intake, measured in kcal, by children using the PortionSize app. The registry record describes the endpoint as a mean and identifies the standard deviation as its dispersion measure. The stated timeframe is one study visit of approximately 2.5 hours. The app-derived estimates are compared with estimates from weighed meals, identified in the record as the criterion measure.

This type of endpoint is used to evaluate a measurement tool: it asks how the app's estimates relate to a reference approach, rather than whether an intervention changes a clinical outcome. The registry names percent mean error and Bland-Altman analysis for assessing errors in the estimates. In general, an average error can describe systematic over- or underestimation, while methods focused on agreement can help describe how closely measurements correspond across observations. These are general explanations of the methods, not findings from this study.

The exact operational definition of percent mean error, the way measurements were paired, and the detailed analysis specification are not reported in the registry record. Nor are the observed mean, standard deviation, error estimates, or other numerical results. Accordingly, the endpoint description establishes what was intended to be evaluated, but does not establish how accurate the app was in the enrolled participants.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a continuous primary endpoint: children's estimated energy intake in kcal using the PortionSize app. It identifies the mean as the parameter and standard deviation as the dispersion measure. The comparison reference is the estimate from weighed meals. The registry names percent mean error and Bland-Altman analysis as methods for determining errors in estimates. Detailed specifications and numerical results are not reported in the registry record.

What does percent mean error describe?

In general, an error measure summarizes the difference between a measurement tool and a reference. A percent-based summary expresses error relative to a reference quantity, which can help describe the scale of error in context. Its exact meaning depends on the formula, direction convention, and aggregation rule. The registry record does not define the formula for percent mean error, state whether positive values denote overestimation or underestimation, or specify how participant-level errors are combined. No study-specific value is reported.

An average error can be informative about systematic bias, but it does not by itself show how closely individual app estimates match the criterion measure. Opposing errors can offset one another in an average. A small average error therefore need not mean that each paired estimate is close. This is a general statistical point, not a finding about this app or study.

Why use Bland-Altman analysis?

In general, Bland-Altman analysis assesses agreement between two measurement methods by examining paired differences in relation to the measurements' average. It can help reveal systematic bias and whether disagreement changes across the measurement range. A usual presentation includes an estimate of average difference and limits describing the spread of paired differences. Those summaries are useful for judging the range of disagreement, but they require actual paired observations and assumptions that must be considered for the particular data.

The registry states that Bland-Altman analysis will be performed, but does not report the estimated bias, limits of agreement, confidence intervals, or any assessment of assumptions. The record also does not state whether the analysis was performed on raw values, transformed values, or another scale. Thus the method's general purpose can be explained, but its study-specific result cannot be interpreted from the registry record.

What does the mean and standard deviation add?

In general, a mean describes the average level of a continuous measurement, and a standard deviation describes variability among measurements. They provide a basic description of a distribution, but they are not substitutes for a direct agreement assessment. The registry identifies these summary types for the primary endpoint, yet the mean and standard deviation values are not reported in the registry record. It is also not reported whether these summaries refer to app estimates, criterion estimates, paired differences, or more than one of these quantities.

Uncertainty is another part of interpretation. A confidence interval is a general statistical range intended to communicate the precision of an estimate under the analysis assumptions. Its width depends on the data and method. No confidence interval for the primary endpoint or agreement measures is reported in the registry record, so the precision of the study's estimates cannot be assessed from this entry.

What analysis choices remain unclear?

For a method comparison, analysts generally need to define which observations form a pair, how repeated observations are handled, and whether differences are evaluated on an absolute or relative scale. The registry identifies the two measurement approaches but does not report these detailed choices. It does not describe model terms, transformations, outlier handling, or whether measurements from clinic and home settings are analyzed separately. Each of these decisions could affect the resulting summaries, so they should not be assumed.

The registry lists six outcome measures and four analyses. It does not report the full analysis plan, how the analyses map to the outcomes, or whether any multiplicity adjustment was used. The term “analyses” alone does not establish the statistical model or the inferential procedure. The analysis populations, missing-data methods, and handling of incomplete app records are also not reported in the registry record.

How should planning illustrations be read?

The stated sample-size assumptions describe a planning illustration computed from an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. The record says that non-time-to-event illustrations are expressed as participants per arm. These are assumptions for planning illustrations, not evidence about the trial's achieved power or a description of its registered single-group design. The observed effect used in the illustration and the resulting participant counts are not reported in the registry record.

General statistical practice distinguishes a design-stage sample-size justification from a retrospective illustration based on an observed estimate. The latter can vary with the observed effect and should not be presented as observed power. The registry record does not report its own sample-size justification, so no conclusion about the original design rationale can be made.

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("Experimental")
n   <- c(30)
m   <- c(717.9)
sp  <- c(284.6)
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  Experimental  30  717.900  284.600  611.629  824.171

11. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures and four analyses. It does not report how the analyses correspond to the outcomes, which analyses are confirmatory, or whether there was a prespecified hierarchy. A multiplicity adjustment or other strategy for interpreting multiple statistical tests is not reported in the registry record. Without those details and the numerical results, the overall false-positive risk and the intended status of any individual finding cannot be assessed.

The record identifies the primary endpoint as children's energy-intake estimation with the app, compared with estimates from weighed meals. It also identifies satisfaction outcomes for children and for parents or caregivers. The registry does not report which outcomes have numerical results, whether all registered outcomes were analyzed, or whether outcome definitions changed. Any such information is not reported in the registry record.

The enrolled number is listed as 31, but the numbers included in each analysis and the analysis-population definitions are not reported in the registry record. The adverse-event figures have a separate at-risk denominator; that denominator should not be treated as the primary endpoint analysis population. Reasons for missing observations, loss to follow-up, unusable app records, and incomplete paired measurements are not reported. Nor does the registry explain whether missing data were excluded, imputed, or handled through a model.

For this measurement question, missing paired observations could matter because a comparison requires app and criterion estimates that can be meaningfully matched. This is a general methodological consideration, not evidence that paired data were missing in this study. The registry does not describe the completeness of pairing or any sensitivity analysis. Consequently, the extent to which missing data might affect the validity estimates is not reported in the registry record.

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.

The registry lists one experimental group for adverse-event reporting. The serious-event figures show 0 affected among 30 at risk; the other-event figures show 0 affected among 30 at risk. The deaths figure is 0. These are the figures recorded for the listed group, and they should be read with their stated at-risk denominators.

The registry's enrollment is 31, while the adverse-event entries use 30 as the at-risk count. The reason for the difference is not reported in the registry record, and it should not be inferred. The entry does not list serious-event terms, and it does not provide a narrative explaining the event ascertainment process or observation period beyond the study's separately stated primary-endpoint timeframe. Further detail about how events were collected is not reported in the registry record.

A zero count in the registry means that no affected participants are recorded in that group for the listed category; it does not establish that an event was impossible or that all potential harms were exhaustively captured. These figures are descriptive safety entries, not a comparison against another group. The study is registered as single-group, and no comparative safety estimate is reported.

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

The registry record establishes the study's intended measurement question and the methods it names for addressing it. The app is evaluated for estimating children's energy intake, with estimates from weighed meals serving as the criterion measure. Percent mean error and Bland-Altman analysis are named in the primary outcome description. The record also describes app use in clinic and at home and includes satisfaction outcomes for children and parents or caregivers.

These details do not establish the app's observed accuracy. Numerical estimates for the primary endpoint, error summaries, agreement measures, and their uncertainty are not reported in the registry record. It is therefore not possible to determine the direction or magnitude of error, the degree of agreement for individual observations, or the precision of the estimates. The absence of numerical results in this registry entry is not evidence that the app did or did not perform well.

The design also limits what the study question represents. This is a single-group validation evaluation, not a randomized comparison of treatments. Its primary reference is weighed-meal estimation, not a clinical outcome. The record does not establish whether results differ between clinic and home use, whether performance varies across participant characteristics, or whether the app performs similarly for children's own records and parent or caregiver records. Those questions would require relevant analyses and results, which are not reported in the registry record.

The registered enrollment and safety entries should not be mistaken for the primary endpoint analysis size. The analysis denominator, completeness of paired observations, and missing-data procedures are not reported. The registry also does not report enough detail to assess the practical meaning of any potential error threshold, or whether an accuracy criterion was specified in advance.

The most defensible conclusion from the registry record is descriptive: it records a completed single-group study intended to evaluate a dietary-intake measurement app against weighed meals and to assess user satisfaction. It does not report the numerical findings needed to conclude how close the estimates were to the criterion measure or how certain those estimates were.

14. Why This Trial Matters Statistically

Measurement validity is a statistical question with practical consequences for how dietary intake data are interpreted. A tool can produce an average that appears close to a reference while still disagreeing substantially for individual observations. Conversely, an average difference alone may conceal patterns in which error changes with the amount being measured. The registry's choice to name both percent mean error and Bland-Altman analysis reflects distinct aspects of measurement error: relative error summaries and assessment of agreement. This is a general explanation of why both kinds of information can matter.

The study also describes use in more than one setting: at the clinic and at home under free-living conditions. In general, measurement performance can be context-dependent, so it is important to distinguish the settings and the populations contributing to each analysis. The registry does not report setting-specific validity results or whether those settings were analyzed separately. This makes the study's intended context clear while leaving its observed performance unresolved in the registry record.

Including children as app users and parents or caregivers as reporters raises a further measurement question: the quality of recorded intake may depend on who uses the tool. The registry lists separate satisfaction outcomes for children and for parents or caregivers, but does not provide numerical results or details about any comparison between reporting groups. Satisfaction and measurement agreement are different constructs; a satisfaction rating alone would not establish accuracy. This is a general distinction, not a study finding.

Finally, the record illustrates why the analysis denominator and paired-data handling matter in validation work. A method comparison depends on suitably matched measurements, and the number contributing to each analysis can differ from enrollment. Those analysis details are not reported in the registry record. The study's statistical relevance lies in the validity question it asks, but its empirical contribution cannot be characterized without the actual estimates and their precision.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record provides a concise description of the study but omits details needed to assess the validity analysis fully. Participant eligibility criteria, age range, demographic characteristics, recruitment methods, and the distribution of participants across settings are not reported in the registry record. The enrollment is recorded as 31, but the primary analysis denominator and the number with complete app and weighed-meal measurements are not reported.

The weighed meals are identified as the criterion measure, but the procedures for weighing, food selection, meal recording, and matching app entries to reference values are not described in the registry record. The exact definition and direction of percent mean error are also not reported. The record does not specify whether measurements were repeated, whether app estimates were independently checked, or how observations from clinic and home use were handled.

Important statistical details are absent, including the confidence intervals or other uncertainty measures, assessment of Bland-Altman assumptions, treatment of extreme observations, and any model adjustment. The record does not report missing-data methods, analysis-population definitions, or a multiplicity strategy. It also does not identify a prespecified threshold for acceptable measurement error. These omissions prevent a detailed assessment of precision, robustness, and the practical interpretation of any estimate.

The single-group design is appropriate to describe as a method evaluation, but it does not provide a randomized treatment comparison. The planning assumptions in the record refer to equal allocation between two arms, so they do not describe the registered single-group design. The protocol's own sample-size rationale is not reported in the registry record. These limitations concern the available registry information and should not be taken as proof that the omitted procedures were not conducted.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Validity of the PortionSize App to Measure Children's Food Intake — NCT05587816.
• Registry record last updated 2026-04-15; status last verified 2026-03.
• Results first posted 2026-04-15.
• 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

NCT05587816 is a completed, single-group evaluation of the PortionSize app, sponsored by Pennington Biomedical Research Center. The registry lists enrollment of 31, a start date of 2022-10-28, and a primary completion date of 2023-03-06. Masking is listed as none, and allocation is listed as NA.

The primary endpoint concerns children's energy-intake estimates in kcal from the app, compared with estimates from weighed meals. The registered description names percent mean error and Bland-Altman analysis. The study also describes clinic and home use and lists satisfaction outcomes for children and parents or caregivers. Numerical primary and secondary outcome findings, detailed analysis populations, and missing-data methods are not reported in the registry record.

The safety entry records 0 serious events among 30 at risk, 0 other events among 30 at risk, and 0 deaths. The reason the at-risk count differs from the enrollment is not reported in the registry record.