Computer Automation for Diagnosis and Management of Childhood Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Computer Automation for Diagnosis and Management of Childhood Type 2 Diabetes (NCT01814787), a randomized na study. The primary outcome is Number of Children With Documented Risk Factors for Type 2 Diabetes.

2. Clinical Question

The study’s stated purpose was to put American Diabetes Association screening guidance and management prompts into pediatric primary care through a computer decision support system. Its stated hypothesis concerned improved compliance with recommended screening and better clinical management among children identified with pre-diabetes or type 2 diabetes. The single posted primary outcome, however, is a count of children meeting a documented risk-factor definition. It is not itself a direct measure of screening compliance, management quality, or later health outcomes.

3. Trial Design

The registry describes a randomized, single-masked, parallel trial of the CHICA Type 2 Diabetes Module versus usual care. The study is marked completed, with enrollment of 1369. It lists type 2 diabetes as the condition and Indiana University as sponsor. The phase field is marked NA. The registry record gives a start of 2013-10 and primary completion of 2016-07.

There were two arms. The CHICA Type 2 Diabetes Module arm included 685 participants, and the usual-care arm included 684. The trial description places the module at intervention clinic sites and says the control group had the CHICA system without the diabetes module. The registry record does not identify who was masked, so the meaning of single masking cannot be specified more precisely. It also does not report the allocation sequence method, concealment procedures, or details of any clustering by clinic in the posted outcome analysis.

The primary endpoint was assessed at 12 months and recorded as a count of participants meeting the specified definition of documented risk factors. The posted count was 283 for the module arm and 282 for usual care. No secondary outcomes are posted, and the registry record does not report a statistical analysis for the primary outcome. The registration includes one outcome measure and no posted analyses.

The difference in the level of detail between the study’s broad objective and the posted outcome matters for interpretation. The objective refers to screening and clinical management, while the primary endpoint concerns documented risk-factor status. The posted record does not provide additional outcomes that directly quantify guideline compliance, referrals, treatment changes, or clinical follow-up. It also does not describe the trial’s original design assumptions or the rationale for the enrollment target; those details 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.

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 Number of Children With Documented Risk Factors for Type 2 Diabetes, measured over 12 months.

The primary endpoint is the number of children ages 10 and older with documented risk factors for type 2 diabetes, defined in the registry description as >85%BMI and 2 of 4 risk factors. The assessment timeframe is 12 months. The outcome is recorded as a count of participants, rather than as a continuous measurement or a time-to-event outcome. The registry lists 283 participants in the CHICA Type 2 Diabetes Module arm and 282 in usual care.

This type of endpoint is relevant to a screening-focused intervention because it records whether children meeting a stated risk-factor definition were documented. Counts can be compared between randomized groups, and the registry’s comparison expresses that contrast as a difference between the proportions in the two arms. The distinction between documented status and underlying status is important: the endpoint concerns what was recorded according to the registry definition. It does not directly establish whether every eligible child was screened, whether documentation was complete, or whether clinical management changed.

The registry does not provide more detail about the four risk factors, the exact operational meaning of the BMI criterion, or how missing or incomplete documentation was handled. It also does not report additional primary endpoint components. As a result, the endpoint should be interpreted narrowly as the posted count meeting the stated documentation definition within the 12-month timeframe, not as a general measure of diabetes diagnosis or subsequent health.

7. Results Posted in the Registry

• CHICA Type 2 Diabetes Module: 41.3% (283/685) versus 41.2% (282/684) on control; risk difference 0.1 percentage points (95% CI -5.1 to 5.3), risk ratio 1.00 (0.88–1.14); unadjusted two-proportion z-test p = 0.974.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each child either meets the documented risk-factor definition or does not. The registry reports the number meeting the definition in each arm and compares the proportions. This structure is different from a continuous outcome such as a laboratory value, where the analysis would focus on average measurements, and from a time-to-event outcome, where timing and censoring would be central.

What the risk difference represents

The risk difference is the intervention-group proportion minus the control-group proportion. The posted estimate is 0.001, with a 95% confidence interval from -0.051 to 0.053. The scale is an absolute difference in proportions. Zero represents no difference between groups on this scale; negative values indicate a lower proportion in the module arm, while positive values indicate a higher proportion in that arm. The interval crosses zero, so the data are compatible with values in both directions as well as no difference.

Relative effect measures

The risk ratio compares the intervention proportion with the control proportion, with 1 representing equal proportions. The posted risk ratio is 1.002, with a confidence interval from 0.883 to 1.137. The odds ratio compares the odds of meeting the endpoint in each arm; 1 likewise represents equal odds. The posted odds ratio is 1.004, with a confidence interval from 0.809 to 1.244. Since both intervals include 1, neither relative estimate rules out no difference. These measures describe the same binary outcome on different scales, so their numerical values should not be compared as if they were interchangeable.

How to read the p-value

The unadjusted p-value is 0.974. In general, a p-value quantifies how inconsistent the observed data are with a specified null model, assuming that model and the analysis conditions hold. It is not a measure of effect size, and it does not indicate the probability that a treatment is effective or ineffective. It also does not establish equivalence. The registry does not state a full statistical analysis plan, including whether a particular test was prespecified, so details beyond the posted unadjusted result are not reported in the registry record.

Randomization and analysis

Random assignment is intended to make treatment groups comparable, on average, with respect to factors that could affect outcomes. This supports causal interpretation when allocation, follow-up, outcome ascertainment, and analysis are appropriately conducted. The registry identifies the design as randomized, but does not describe the method of sequence generation, allocation concealment, adherence, or analysis population. The counts and estimates should therefore be understood as the posted comparison, without assumptions about unreported procedures.

The registry reports a parallel design and two arms. It does not report whether the statistical analysis accounted for the clinic sites mentioned in the intervention description. When participants are treated within clinics, outcomes can be correlated within the same clinic; general statistical methods account for such clustering when the design and data warrant it. Ignoring relevant clustering can affect uncertainty estimates. Whether this issue applies to the posted estimates cannot be determined from the registry record.

Why documentation matters

The endpoint is based on documented risk factors. In general, an outcome that depends on clinical documentation can reflect both the underlying state and the process of recording it. A decision-support tool may be intended to influence identification or documentation, so the distinction is especially relevant to interpretation. The registry does not report validation of the endpoint, missing-data rules, or completeness of risk-factor documentation. It is therefore not possible to determine from the posted result how those features affected the comparison.

General statistical interpretation also distinguishes a nonsignificant test from evidence of no meaningful difference. A narrow confidence interval concentrated near zero might support a more precise conclusion than a wide interval spanning both directions. Here, the interval and estimates should be considered together: the point estimate is close to zero, but the interval covers a range of differences. No equivalence margin or target precision is reported, so the result cannot be recast as an equivalence finding.

The registry lists no secondary outcomes and no analyses. It does not report adjusted estimates, subgroup analyses, interim analyses, missing-data procedures, or an analysis population such as intention-to-treat. Those elements cannot be inferred from the primary comparison. The displayed estimates are useful for describing the posted endpoint, while their interpretation remains limited by the analysis details absent from the registry record.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.001, and its 95% confidence interval runs from -0.051 to 0.053. The point estimate is close to zero, but the interval includes negative and positive differences. This means the data do not pinpoint a single effect value; they are compatible with a range of values under the analysis method. The interval’s inclusion of zero is consistent with the unadjusted p-value of 0.974.

A confidence interval is a statistical measure of uncertainty, not a range containing a fixed proportion of individual participants or a guarantee about the true value. In general, a narrower interval indicates greater precision for the estimated contrast, while a wider interval indicates greater uncertainty. Precision is distinct from whether an estimate is clinically important. The registry does not report a threshold defining an important difference, so the bounds cannot be categorized as clinically acceptable or unacceptable from this record.

The relative estimates also have uncertainty: the risk ratio is 1.002 with a confidence interval from 0.883 to 1.137, and the odds ratio is 1.004 with a confidence interval from 0.809 to 1.244. Both intervals include the no-difference value of 1. The numerical endpoints differ because these measures use different scales; they should be interpreted according to their own null values.

Precision is limited to the endpoint and analysis reported. The registry does not describe adjustment, clustering, missing data, or the exact statistical method used to construct the intervals. These omissions limit assessment of whether the reported uncertainty reflects all relevant features of the design. The estimates are informative as posted, but they do not resolve uncertainty about outcomes that were not reported.

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 participants per arm, not total events. They are illustrations based on the observed effect, rather than the trial’s original design calculation. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The trial’s own sample-size assumptions are not reported in the registry record.

For planning for 100% of the observed effect, the figures are 5,168,641 participants per arm for 80% power and 6,919,343 participants per arm for 90% power. For planning for 75% of the observed effect, they are 9,187,985 participants per arm for 80% power and 12,300,105 participants per arm for 90% power. For planning for 50% of the observed effect, they are 20,671,367 participants per arm for 80% power and 27,673,095 participants per arm for 90% power.

In general, power is the probability that a planned study will detect an effect of the specified size under the assumptions used, when that effect is present. The 80% and 90% columns represent different target probabilities of detection; the higher target requires more participants for each listed effect. Smaller effects are harder to distinguish from random variation and therefore require more participants under the same assumptions. These figures are planning illustrations, not observed power, and do not describe what the completed trial was originally designed to detect.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.412; treatment proportion p₁ = 0.413; average p̄ = 0.413.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6962 = 1.3646.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6962 = 0.5860.
4. n = (1.3646 + 0.5860)² / (0.0009)² = 5,168,640.4, rounded up to 5,168,641 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
p0 <- 0.412280701754386                          # control proportion
p1 <- 0.413138686131387                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1814787)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 5,168,641 (80% power), 6,919,343 (90% power)
 75% of observed effect: n per arm = 9,187,985 (80% power), 12,300,105 (90% power)
 50% of observed effect: n per arm = 20,671,367 (80% power), 27,673,095 (90% power)
power.prop.test: 5168640 
Simulated power at n = 5,168,641 per arm: 0.798 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when several outcomes, comparisons, subgroups, or repeated analyses are tested. The more statistical tests considered, the greater the chance that at least one result appears notable by chance if no adjustment or prespecified hierarchy is used. The registry lists one primary outcome measure, no secondary outcomes, and no posted analyses. It does not report whether additional outcomes or comparisons were collected or analyzed outside the posted record, nor whether a multiplicity strategy was planned.

Missing-data handling is also not described. The registry reports arm sizes of 685 and 684 and endpoint counts of 283 and 282, but it does not explain whether every randomized participant had an outcome available, how incomplete records were treated, or whether any imputation or sensitivity analysis was used. The posted arm sizes should not automatically be interpreted as the number with complete endpoint data unless the registry explicitly establishes that.

The analysis population is not reported in the registry record. In general, an intention-to-treat analysis keeps participants in their randomized groups regardless of adherence, while per-protocol analyses apply additional conditions related to treatment received or protocol adherence. The registry does not state which population generated the reported comparison. It also does not describe exclusions after randomization, protocol deviations, or how any participants were handled in the analysis.

These omissions do not invalidate the posted estimates, but they restrict the ability to assess how robustly they represent the randomized comparison. The registry does not report interim analyses, stopping rules, or a complete statistical analysis plan. No conclusions about those procedures should be inferred from the absence of detail in the record.

14. 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’s adverse-event record lists zero serious adverse events and zero other adverse events in the CHICA Type 2 Diabetes Module arm, among 685 participants at risk. It also lists zero serious adverse events and zero other adverse events in usual care, among 684 participants at risk. The deaths field is not populated for either arm, so deaths are not reported in the registry record.

These entries describe what was recorded in the registry; they do not by themselves establish that no adverse events occurred during the trial. The record does not give definitions, collection procedures, observation duration for adverse events, or details about completeness of reporting. The primary outcome timeframe is 12 months, but the adverse-event listing does not separately state its assessment timeframe. Zero posted counts should therefore be read as zero events recorded in these fields, not as proof of absence of harm.

The safety entries also do not support a comparative safety conclusion beyond the posted counts. Without further information about event ascertainment, severity criteria, exposure, and missing reporting, the figures are limited descriptions of the registry record.

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

The result describes a narrow comparison: the number of children meeting the posted documented-risk-factor definition at 12 months in a randomized comparison of a computer decision-support module with usual care. The counts were 283 and 282. The risk-difference estimate was 0.001, and its 95% confidence interval ranged from -0.051 to 0.053. The unadjusted p-value was 0.974. The estimate is close to zero, while the interval leaves uncertainty about the size and direction of the underlying difference.

This result does not establish that the module and usual care are equivalent. Equivalence requires a prespecified margin and an analysis directed at showing that differences fall within that margin. Neither is reported in the registry record. A nonsignificant result is also not proof that an intervention has no effect. It indicates that the posted comparison did not demonstrate a statistically clear difference under the reported analysis.

The endpoint is documentation of risk factors, not a direct measure of whether screening guidance was followed for all eligible children, whether children received a diagnosis, whether clinicians changed management, or whether later health outcomes changed. The trial’s stated purpose included screening and management, but those broader aims cannot be judged fully from this single posted outcome. The record does not report secondary outcomes that would address those questions.

Nor do the results identify why the counts were similar. The endpoint can reflect underlying risk status and the completeness or consistency of documentation. Details of measurement validation, missing-data handling, and analysis population are not reported. The comparison therefore supports description of the posted outcome, but not a definitive explanation of the processes behind it.

The figures should not be used as medical guidance. They are statistical results from one trial record and concern a specified outcome and timeframe. Broader claims about clinical effectiveness, long-term outcomes, or performance in other settings would require evidence not reported in this registry record.

16. Why This Trial Matters Statistically

This trial raises an important statistical distinction between a program’s objective and the outcome used to evaluate it. The stated aim concerned implementing screening guidance and management prompts, while the posted primary outcome counts children documented as meeting a risk-factor definition. An endpoint should be interpreted according to what it actually measures, not as a substitute for every process or clinical outcome the intervention may be intended to influence.

The result also illustrates why an estimate, confidence interval, and p-value should be considered together. The risk-difference estimate of 0.001 is near zero, the interval extends from -0.051 to 0.053, and the p-value is 0.974. Together, these describe little observed separation on the endpoint while retaining uncertainty about the underlying difference. A test result alone cannot establish equivalence or rule out meaningful effects without a prespecified criterion for what differences matter.

Documentation-based endpoints are especially relevant in studies of decision support. In general, a system that changes the information available to clinicians might affect the recording of risk factors as well as clinical actions. Knowing whether documentation completeness was measured and how missing records were handled would help distinguish those possibilities. Those details are not reported in the registry record.

Finally, the posted sample-size illustrations emphasize how difficult it can be to plan for a small observed contrast under stated assumptions. They are not the trial’s original design justification, and they do not indicate the study’s observed power. Separating planning calculations from the completed trial’s results is essential: the randomized comparison describes the data, while planning calculations describe hypothetical future studies under explicit assumptions.

17. 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 limited detail about how the trial was conducted and analyzed. It identifies randomization, single masking, and a parallel design, but does not report the sequence-generation method, allocation concealment, who was masked, or how adherence and protocol deviations were handled. It names intervention clinic sites but does not describe the number of sites or whether the primary analysis accounted for possible correlation among participants treated at the same site.

The endpoint is described as children ages 10 and older with documented risk factors, defined as >85%BMI and 2 of 4 risk factors. Further details of the risk factors, measurement procedures, and completeness of documentation are not reported. Because the endpoint depends on recorded information, the absence of documentation methods limits interpretation of what the counts represent.

The record does not report missing-data procedures, the analysis population, adjusted analyses, subgroup analyses, or a complete statistical analysis plan. It posts no secondary outcomes and no analyses beyond the primary comparison. The trial’s original sample-size assumptions and justification are also not reported. The planning illustrations based on the observed effect do not fill that gap.

Safety reporting is limited to zero recorded serious and other adverse events in each arm; deaths are not reported in the registry record. Event definitions, ascertainment, and reporting completeness are not described. These limitations mean that the posted findings are best treated as a focused description of one endpoint rather than a complete account of all trial procedures, outcomes, and potential effects.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Computer Automation for Diagnosis and Management of Childhood Type 2 Diabetes — NCT01814787.
• Registry record last updated 2017-05-15; status last verified 2017-04.
• Results first posted 2017-03-23.
• 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.

21. Record Summary

The registry record describes a completed randomized, single-masked, parallel trial comparing the CHICA Type 2 Diabetes Module with usual care. Enrollment was 1369, with 685 participants in the module arm and 684 in usual care. The listed condition is type 2 diabetes, the sponsor is Indiana University, and the phase field is marked NA.

The primary outcome was the count of children ages 10 and older with documented risk factors at 12 months. Counts were 283 and 282. The risk difference was 0.001, with a 95% confidence interval from -0.051 to 0.053 and p = 0.974. The registry lists no secondary outcomes or posted analyses. Adverse-event fields show zero serious and other events recorded in each arm; deaths are not reported in the registry record.

The result describes the posted endpoint and does not resolve questions about screening processes, management, or longer-term outcomes. Important details of the statistical plan and outcome handling are not reported in the registry record.