Automated Device for Asthma Monitoring: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Automated Device for Asthma Monitoring (NCT01696357), a registered phase not stated study. The primary outcome is Average Number of Coughs Per Hour.

2. Clinical Question

3. Trial Design

The registry identifies this completed study as Automated Device for Asthma Monitoring, sponsored by the University of Rochester. Its start is recorded as 2009-09 and its primary completion as 2013-08. Enrollment is listed as 84. The study description concerns development of a non-invasive prototype using an iPod platform and sound-data analysis to monitor symptoms in adolescents.

The two groups were adolescents with asthma and adolescents without asthma. The asthma group included participants ages 13-17 with an asthma diagnosis, both symptomatic and non-symptomatic, who had currently prescribed asthma medication. The comparison group included participants ages 13-17 without an asthma diagnosis or another respiratory condition presenting with asthma-like symptoms. The record does not state how participants were allocated to these groups. It specifically indicates that this is not a randomized comparison and that the groups are not a like-for-like control comparison.

The named device intervention is ADAM. The registry describes the intended monitoring as including wheezing, coughing, and activity levels, with the device carried by or placed near the adolescent. The primary outcome is the average number of coughs per hour over 7 days. Although the study status is completed, the registry record does not report the analyzed outcome values or a statistical comparison for that endpoint. It lists no analyses. Thus, the design and intended measurement can be described, but the posted record does not support a conclusion about the magnitude or precision of a group difference.

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

Analysis population for the primary outcome, as defined in the registry: One subject in the non-asthma group failed to return the device after the 7-trial. Data from 22 participants (3 from the asthma group, 19 from non-asthma group) were not included in the analysis as no data were recorded (due to a mechanical issue) or data were too extreme (due to cough algorithm failure).

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 Average Number of Coughs Per Hour, measured over 7 days.

The primary endpoint is the average number of coughs per hour, a continuous measure summarized as a mean with standard deviation. The stated timeframe is 7 days. The endpoint description says that the device was supposed to detect and count coughs per 24 hours and register the counts in the device. The registry description of the endpoint is truncated after its account of data exclusions, so further measurement details are not reported in the registry record.

Cough frequency was selected as a measurable symptom that the device was intended to capture. In general, a continuous endpoint can retain information about the level of a measurement rather than reducing it to categories such as present or absent. A mean describes the average recorded level in a group, while a standard deviation describes how spread out individual measurements are around that mean. These summaries could characterize recorded cough frequency, but they do not by themselves establish that the device measures coughs accurately.

The endpoint is a device-recorded measure, not a direct measure of asthma control, quality of life, or clinical outcomes. The registry record does not report observed group means, standard deviations, or a comparison estimate. It therefore does not show how many coughs were recorded or how different the groups were on this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes the primary outcome as a continuous measure: average coughs per hour, summarized by a mean and standard deviation. It does not report group-level summaries or a statistical analysis for this outcome. The explanations below are general statistical context, not descriptions of methods confirmed for this study.

What does a mean describe?

In general, a mean is the arithmetic average of the observations included in an analysis. For a device-monitoring endpoint, its interpretation depends on what counts as an observation. The registry says the device was intended to count coughs per 24 hours, while the named endpoint is average coughs per hour over 7 days. It does not explain how daily recordings were converted into the endpoint, how incomplete recording periods were handled, or whether the mean was calculated across recording periods or participants. Those details are not reported in the registry record.

A mean can be useful for describing a group's overall recorded level, but it does not show the distribution of the measurements. A small number of very high or very low values can influence it. That consideration is particularly relevant because the record says some data were considered too extreme due to cough-algorithm failure. The record does not give the values, criteria, or procedures used to identify such data.

What does a standard deviation describe?

In general, the standard deviation describes the spread of measurements around a mean. A larger standard deviation indicates greater variation among the analyzed measurements; it is not the uncertainty of the mean itself. The registry lists standard deviation as the measure of dispersion for the primary outcome but does not report an observed value. Without the actual mean and standard deviation, neither the average recorded cough frequency nor its variability can be assessed.

How should an estimated group difference be read?

In general, an estimated difference between group means describes the contrast in average measurements for the groups included in the analysis. A confidence interval describes statistical uncertainty around an estimate under the assumptions of the analysis. A narrow interval indicates greater precision than a wide interval, while an interval that includes values compatible with no difference does not establish that the true difference is exactly zero. The registry record does not report an estimate or confidence interval for this endpoint, so these interpretations cannot be applied to an observed result in this study.

A p-value, in general, measures how incompatible the observed data are with a specified null model, given the assumptions of the statistical test. It is not the probability that the null hypothesis is true, and it does not describe the size or practical importance of an effect. No p-value or test for the primary endpoint is reported in the registry record. The record also does not report whether any analysis adjusted for other participant characteristics or repeated measurements.

Why can measurement failures matter?

In general, measurement error can make an observed value differ from the quantity a study intends to measure. For automated cough detection, a missed cough or a false detection can affect the recorded count. A mechanical problem can prevent data capture altogether, while an algorithm failure can make recorded values unreliable. The registry explicitly notes both missing recordings due to a mechanical issue and extreme data attributed to cough-algorithm failure. It does not report validation results, the frequency or pattern of these failures beyond the stated exclusions, or a sensitivity analysis exploring alternative handling.

In general, over-dispersion means that observed variation is greater than a statistical model expects. It can matter when a model's assumed variability is too small, because uncertainty estimates may then be too optimistic. The registry record does not identify a count model, report an over-dispersion assessment, or describe a model-based analysis. Because the listed primary summary is a mean and standard deviation, no particular count-model method should be inferred.

What do planning calculations represent?

In general, sample-size planning illustrations describe the number of participants per arm that a future study might require under stated assumptions about effect size, significance level, allocation, and statistical model. They are not observed results and are not evidence of power achieved in a completed study. The planning assumptions associated with this study are a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects expressed as fractions of the observed effect. These are planning illustrations based on the observed effect, not the trial's own design. The registry does not report its own sample-size justification. Without reported outcome estimates and uncertainty, the planning illustrations cannot substitute for a results analysis or establish the precision of the study's endpoint.

Overall, the named summary statistics are familiar descriptive tools, but the registry does not supply the endpoint values or analysis details necessary to determine how they were applied. The distinction between a recorded count and a valid measure of cough frequency is central to interpreting a device-development study.

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("Adolescents With Asthma", "Adolescents Without Asthma")
n   <- c(39, 22)
m   <- c(57.64, 0.63)
sp  <- c(111.41, 0.76)
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  Adolescents With Asthma  39  57.640  111.410  21.525  93.755
CARM  Adolescents Without Asthma  22  0.630  0.760  0.293  0.967

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and no analyses. It does not report a statistical analysis plan, a prespecified testing hierarchy, adjustment for multiple comparisons, or interim analyses. The record therefore does not establish whether the primary endpoint was the only outcome considered in the investigators' analyses or how any other exploratory measures may have been handled. No multiplicity-adjusted inference can be described from the record.

Missing and excluded data are a substantial interpretive issue. The endpoint description states that data from 22 participants were not included in analysis: 3 from the asthma group and 19 from the non-asthma group. The stated reasons were no recorded data due to a mechanical issue or data considered too extreme due to cough-algorithm failure. The record does not separate the number associated with each reason, explain the criteria for extreme data, or report whether those reasons differed by group. It also does not give results for excluded participants or describe imputation, sensitivity analyses, or other missing-data methods.

The record additionally says that one subject in the non-asthma group failed to return the device after the 7-trial. It does not clarify how that circumstance relates to the stated exclusions or whether it affected the endpoint analysis. The record does not define an analysis population or explain whether all enrolled participants were assessed for the primary endpoint.

In general, complete-case analyses can be biased when the availability of usable measurements is related to the outcome or participant characteristics. Excluding extreme values may also alter a group mean if the observations reflect genuine variation rather than device error. Here, the registry attributes the extreme data to cough-algorithm failure, but does not provide enough detail to independently evaluate that classification. These omissions leave the effect of missingness and exclusions uncertain.

12. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry lists zero serious adverse events among 42 participants at risk in the adolescents-with-asthma group and zero among 42 at risk in the adolescents-without-asthma group. It also lists zero other adverse events among 42 at risk in each group. No serious adverse event terms are listed. The deaths field is not reported in the registry record.

These figures describe what the registry records; they should not be read as proof that no harms occurred. In general, an adverse-event count depends on how events were defined, collected, and reported, and a zero recorded count does not establish that an event is impossible. The record does not describe the adverse-event collection procedures, the observation period for safety reporting, or whether all events were actively solicited.

The groups are not a randomized, like-for-like control comparison, and the registry does not report comparative safety analyses. The listed counts therefore do not establish a causal safety contrast between use of the device and a control condition. The device's non-invasive design is described, but that description does not replace outcome data or safety-reporting detail. The serious and other adverse-event figures are best interpreted narrowly as the counts recorded for the stated groups.

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

The record documents a completed study of an automated symptom-monitoring device for adolescents and identifies average coughs per hour over 7 days as the primary endpoint. It describes a prototype intended to use sound-data analysis to monitor coughs, wheezing, and activity. It also reports that some participant data could not be used because of missing recordings or data attributed to algorithm failure. These registry entries make the study relevant to questions about both measurement feasibility and the limitations of automated symptom capture.

However, the registry does not report the analyzed endpoint values, an estimate comparing groups, or measures of statistical uncertainty. It therefore does not show the observed average cough frequency in either group, whether a difference was present, or how precisely any difference was estimated. The registry also does not report results on device acceptability or validate its cough-detection accuracy. Completion status and the existence of a named endpoint are not evidence that the stated broader aims were achieved.

The comparison is between adolescents with asthma and adolescents without asthma, and allocation is not stated. The record warns that this is not a randomized, like-for-like control comparison. Any group contrast, if reported elsewhere, would need to be interpreted in light of that design and the possibility that the groups differ in ways beyond the device monitoring itself. It would not by itself establish that the device changed symptoms or improved asthma management.

In general, a measurement tool can be evaluated for whether it captures a target reliably and accurately, while a clinical outcome study asks whether an intervention changes health outcomes. These are distinct questions. The registry describes an intended monitoring strategy and a cough-count endpoint, but does not report enough results to answer either question. The supported conclusion is about the study's stated purpose and documented data limitations, not about treatment benefit, clinical effectiveness, or the accuracy of ADAM.

14. Why This Trial Matters Statistically

This study raises a central statistical issue for digital measurement: the quality of an analysis depends on whether the device produces data that represent the intended phenomenon. A count recorded by a prototype is not automatically equivalent to the number of coughs that occurred. Mechanical failures can leave periods without observations, and algorithm failures can produce implausible recordings. The registry explicitly identifies both types of problem, making data capture and measurement validity important to interpreting its intended endpoint.

The study also illustrates the difference between a device-development question and a randomized treatment question. The groups were defined by asthma status, and the registry says allocation was not stated and the comparison was not like-for-like. This means that a group contrast would not isolate a causal effect of device use. Clear description of group formation, data availability, and measurement procedures is essential when the goal is to assess whether a monitoring technology can characterize symptoms.

Another statistical lesson is that exclusions require transparent reporting. The registry reports 22 participants whose data were not included and gives broad reasons, but does not provide group outcome summaries or detail the analysis of missing and extreme measurements. Without those results, readers cannot judge how exclusions affect the estimated average or its precision. Reporting the distribution of usable recordings and the handling of device failures would help distinguish a measurement problem from variation in the underlying symptom.

The study's broader hypothesis concerned acceptability and more objective symptom assessment, whereas its named primary endpoint was average cough frequency. Keeping these aims distinct matters: a cough-count summary alone cannot establish acceptability, accuracy, or an effect on asthma management. The record's limitations therefore emphasize the value of reporting both technical performance and statistical results when evaluating a new monitoring approach.

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 does not report the primary endpoint results. Group means, standard deviations, an estimated contrast, confidence intervals, and a p-value are not reported in the registry record. As a result, neither the observed cough frequency nor the precision of a group comparison can be assessed from the posted record.

Data availability is a specific concern. The record states that data from 22 participants were excluded because no data were recorded due to a mechanical issue or data were too extreme due to cough-algorithm failure. It gives group counts for these exclusions but not the count associated with each reason, the criteria used to classify extreme data, or results from analyses that include or otherwise address these observations. It also notes a device non-return in the non-asthma group after the 7-trial without clarifying the consequences for analysis.

The registry does not report how cough counts were validated, how periods with incomplete recording were handled, or how the average per hour was derived from the intended daily counts. The endpoint description is truncated, and the record does not provide enough procedural detail to assess the measurement process. The intended monitoring of wheezing and activity is described, but those measures are not identified as the primary endpoint in the record.

The group comparison has design limitations: allocation is not stated, and the registry says the groups are not a like-for-like control comparison. The record also does not report a sample-size justification, detailed statistical analysis plan, analysis population, multiplicity approach, or missing-data sensitivity analysis. These omissions limit assessment of bias, uncertainty, and reproducibility. They do not establish that any particular analysis was or was not performed outside what the registry record reports.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Automated Device for Asthma Monitoring — NCT01696357.
• Registry record last updated 2014-12-10; status last verified 2014-12.
• Results first posted 2014-12-10.
• 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 registry record describes the completed University of Rochester study Automated Device for Asthma Monitoring, with enrollment listed as 84. The recorded start is 2009-09 and primary completion is 2013-08. The study concerned a non-invasive prototype using an iPod platform to monitor symptoms in adolescents.

Groups comprised adolescents ages 13-17 with asthma and adolescents ages 13-17 without asthma. The primary endpoint was average coughs per hour over 7 days, summarized as a mean and standard deviation. The registry reports that data from 22 participants were not included because of missing recordings or data considered too extreme due to cough-algorithm failure. It lists no analyses and does not report the analyzed endpoint results, so the magnitude and precision of any group difference are not available in the registry record.