Effectiveness of Dispatch-Assisted Cardiopulmonary Resuscitation (CPR) Instructions: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effectiveness of Dispatch-Assisted Cardiopulmonary Resuscitation (CPR) Instructions (NCT00664443), a registered phase not stated study. The primary outcome is Sensitivity.

2. Clinical Question

The broader study objectives include assessing dispatchers’ ability to diagnose cardiac arrest over the phone, the frequency and impact of perceived agonal breathing, completion of dispatch-assisted CPR instructions, and their impact on bystander CPR and survival rates. These objectives place the primary endpoint within a wider evaluation of call recognition and subsequent events. However, the registry information available for this analysis does not report numerical findings for those objectives. The primary endpoint therefore addresses recognition at the call, not whether CPR instructions were completed or whether a patient survived.

3. Trial Design

This is a completed registry study sponsored by the Ottawa Hospital Research Institute. The record gives a planned enrollment of 2260 and describes the study as an evaluation of 9-1-1 calls involving unconscious patients in participating communities. Its listed study arm is “Potential cardiac arrest calls,” with the stated purpose of determining diagnostic accuracy by 9-1-1 call takers. The registry does not describe a randomized allocation or a like-for-like control group; allocation is not stated.

The study dates in the registry are 2007-07 for the start and 2010-06 for primary completion. The arm description specifies that the relevant calls were to be included over the timeframe January 2008 and October 2009 in participating communities. These dates describe the registry record and the stated call-inclusion window; they do not establish additional details about how calls were selected, classified, or reviewed.

The study’s stated overall aim was to understand factors associated with successful dispatch-assisted CPR instructions and ultimately support improved outcomes for cardiac arrest patients. Its specific objectives extend beyond the primary endpoint to phone-based diagnosis, perceived agonal breathing, successful completion of CPR instructions, bystander CPR, and survival. The registry lists two outcome measures and two analyses, but details for the additional measure and analyses are not reported in the registry record.

Because the record describes a call-based diagnostic assessment rather than a randomized comparison, the primary endpoint should be read as a measure of recognition within the defined call population. The record does not provide enough information to describe further design features, such as the method used to confirm cardiac arrest, whether assessments were blinded, or how calls with incomplete information were handled. Those design 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

Analysis population for the primary outcome, as defined in the registry: Confirmed cardiac arrest calls (true positive)

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.

Population: Potential cardiac arrest audio recordings.

6. Primary Endpoint

The registry lists the primary outcome as Sensitivity, measured over At time of 9-1-1 call.

The primary endpoint is titled “Sensitivity.” Its operational description is the number of true cardiac arrest cases correctly identified, measured at the time of the 9-1-1 call. The stated population is confirmed cardiac arrest calls, described as true positive. In practical terms, the endpoint concerns whether a dispatcher identifies cardiac arrest when the call is ultimately confirmed to involve cardiac arrest.

Sensitivity is a diagnostic concept: it describes recognition of cases that truly have the condition of interest. It is relevant to a dispatch setting because recognition can determine whether CPR instructions are initiated. As a general statistical point, a sensitivity estimate is usually expressed as a proportion of confirmed cases correctly identified. The registry, however, describes this endpoint using a participant count and does not report a calculated proportion, the observed count, or the denominator used for any sensitivity estimate. The label and operational description therefore should not be taken as an available numerical estimate.

The endpoint is specific to recognition at the call. It does not by itself measure the successful completion of CPR instructions, bystander CPR, or survival. Those outcomes appear among the broader study objectives, but their numerical results are not reported in the registry record. The primary endpoint’s timeframe is “At time of 9-1-1 call,” so it addresses an immediate diagnostic decision rather than a later clinical outcome.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry labels the primary outcome “Sensitivity,” while describing its measure as the number of true cardiac arrest cases correctly identified. The distinction matters. A count reports how many cases were identified; sensitivity, as a diagnostic statistic, relates correctly identified cases to all cases that truly have the condition. The record gives the intended population as confirmed cardiac arrest calls, but does not report the observed count, the total number of confirmed cases used for calculation, or a sensitivity estimate. Consequently, the outcome definition can be explained, but the statistic cannot be reconstructed from the registry information.

What sensitivity describes

As a general statistical concept, sensitivity evaluates case recognition among people or events that truly have the condition being assessed. It answers a different question from specificity, which concerns recognition of people or events without the condition. A sensitivity measure is useful when missed cases matter to the diagnostic process. In a dispatch context, the relevant classification is whether cardiac arrest is identified during the call. The record does not report specificity, false-positive counts, or other diagnostic classifications, so those aspects of performance cannot be assessed.

A point estimate alone would not describe its precision. As a general concept, a confidence interval gives a range of values compatible with the data and the method used, reflecting sampling uncertainty. A narrower interval indicates greater precision under that method; a wider interval indicates less precision. No confidence interval for the primary endpoint is reported in the registry record. Without an estimate and its uncertainty, the degree of precision cannot be evaluated.

Counts and denominators

Counts depend on the number of observations eligible to contribute. The planned enrollment is 2260, but that number does not establish how many calls were confirmed cardiac arrests or included in the primary analysis. A diagnostic rate requires a clearly defined denominator as well as a numerator. The registry describes the numerator conceptually, as true cases correctly identified, but does not report its observed value or an analysis denominator. It is therefore not appropriate to convert the enrollment figure into a diagnostic rate.

As a general statistical issue, changes in eligibility rules or case confirmation can change which observations enter a denominator. For a call-based evaluation, the method used to establish whether cardiac arrest truly occurred is important to interpreting diagnostic classifications. The registry record does not report the confirmation method, review procedures, or how uncertain cases were categorized. No assumptions about those procedures can be made from the endpoint wording alone.

Design and comparison

The record describes a group of potential cardiac arrest calls, not a randomized treatment comparison. It states that allocation is not stated and that the arms are not a like-for-like control comparison. This limits the kinds of statistical conclusions available: the listed design does not establish a randomized contrast from which to estimate a treatment effect. The primary endpoint instead concerns diagnostic recognition within the stated call population. Details of any subgroup comparisons or adjusted analyses are not reported in the registry record.

The registry lists two analyses, but does not report their specifications or results. It does not state whether the primary analysis was adjusted, how uncertainty was calculated, or what statistical model was used. These details should not be inferred from the endpoint title or from general practice in diagnostic studies.

Planning illustrations

The registry’s planning assumptions describe an illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The assumptions also state that the effect to detect is a stated fraction of the observed effect. For non-time-to-event outcomes, the illustration is expressed as participants per arm; time-to-event planning uses a total number of events. These are planning conventions, not reported trial results.

The registry explicitly characterizes these calculations as planning illustrations rather than the study’s own design. A sample-size illustration does not show what was observed and should not be interpreted as observed power. Because the study is not described as a randomized comparison and the registry does not report a numerical observed effect for the primary endpoint, the planning assumptions do not establish that the actual study had a particular power or allocation. The illustration’s numerical sample-size results are not reported in the registry record.

More generally, sample-size planning depends on assumptions about the effect to be detected, variability or event frequency, the analysis method, and the desired error properties. If the assumptions do not match the data-generating process or the final analysis, a planned sample size may not deliver the anticipated precision. These general considerations do not identify whether any particular assumption was appropriate in this study; the registry does not report a full protocol or sample-size justification.

Finally, statistical significance and clinical importance are distinct concepts. A p value, when reported, addresses compatibility of observed data with a specified statistical model and null hypothesis; it does not by itself measure the size, precision, or practical importance of an effect. The registry does not report a p value or a numerical primary result. Thus, neither statistical significance nor the magnitude of diagnostic performance can be judged from the listed endpoint definition.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Potential Cardiac Arrest Calls")
n      <- c(1536)
events <- c(1012)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Potential Cardiac Arrest Calls  1536  1012  65.9  63.5  68.3  63.5  68.2  63.5  68.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and two analyses, but details beyond the primary endpoint are not reported in the registry record. It does not state which outcomes were designated primary or secondary beyond the listed primary endpoint, whether multiple hypotheses were tested, or whether any adjustment for multiple testing was planned. When several outcomes or analyses are considered, the chance of at least one apparently unusual result can increase; as a general statistical approach, prespecified outcome hierarchies or multiplicity adjustments can help address that issue. No such approach is reported, so none should be assumed.

The registry does not describe missing-data rates, reasons for missingness, imputation, or sensitivity analyses. As a general matter, missing observations can affect both the population represented by an analysis and the precision of its estimate. The direction and size of any resulting bias depend on why data are missing and how the analysis handles them. Since those details are not reported, the possible influence of missing data on the primary endpoint cannot be assessed.

Analysis populations are also not described in sufficient detail. The endpoint refers to confirmed cardiac arrest calls, but the registry does not report how many such calls were identified, how many entered the analysis, or whether any alternative analysis populations were used. The planned enrollment of 2260 should not be treated as the primary-analysis denominator.

The registry does not report a statistical analysis plan, interim analyses, or procedures for selecting among analyses. These items are not reported in the registry record. As a result, readers cannot determine from this information whether the reported outcome definitions and analyses were prespecified or whether changes were made during the study.

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’s serious adverse event fields list “Potential Cardiac Arrest Calls” with 2260 at risk, zero affected, and zero deaths. The corresponding other-category fields also list 2260 at risk and zero affected. These are the posted counts in the registry; they should be read as entries in those fields, not as an independent comparison between groups. The record lists no serious adverse event terms.

Zero recorded events means that the registry fields contain no affected participants or deaths for the listed categories. It does not establish that every event was actively sought, that every possible event type was covered, or that event reporting was complete. The registry does not describe safety surveillance methods, definitions used for serious events, or how follow-up was conducted. Those details are not reported in the registry record.

The safety figures also do not answer the primary diagnostic question. They are separate from the endpoint concerning correct identification of cardiac arrest at the time of the 9-1-1 call. The registry does not report numerical results for that endpoint, so the safety entries cannot be used to infer diagnostic performance or the effect of dispatch-assisted instructions.

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

The record establishes that the study was designed to examine recognition of cardiac arrest during 9-1-1 calls and that its primary endpoint was the number of true cardiac arrest cases correctly identified at the time of the call. It also records a planned enrollment of 2260, the study’s completed status, and its stated broader objectives. These details explain the research question and its context.

They do not establish how often dispatchers identified true cardiac arrest cases. The observed primary count and a sensitivity estimate are not reported in the registry record. There is also no reported confidence interval or other measure of precision for the primary endpoint. Consequently, the record does not support a quantitative conclusion about diagnostic performance, nor can it show how stable an estimate would be across comparable calls.

The broader aims include perceived agonal breathing, successful completion of CPR instructions, bystander CPR, and survival. These are objectives, not findings. Numerical results for them are not reported in the registry record. The primary endpoint is recognition at the call and does not, by itself, demonstrate that instructions were completed or that later outcomes changed.

The study is not described as a randomized comparison, and allocation is not stated. Its listed group is potential cardiac arrest calls, rather than a stated control condition. The record therefore does not support interpreting the primary endpoint as a randomized causal effect of dispatch-assisted CPR instructions. It also does not provide enough detail to assess possible confounding or other sources of bias.

The serious adverse event fields record zero affected participants and zero deaths for the listed group, but the registry does not describe the methods used to collect or verify safety events. These entries should be interpreted narrowly and separately from efficacy or diagnostic performance. Overall, the registry provides a useful outline of the study’s purpose and endpoint, but the absence of reported outcome estimates means the measured result and its uncertainty cannot be evaluated.

14. Why This Trial Matters Statistically

This study is statistically interesting because it frames emergency call handling as a diagnostic classification problem. The relevant question is not simply whether CPR instructions were mentioned, but whether cardiac arrest was correctly recognized among calls later confirmed to involve cardiac arrest. That distinction clarifies what the primary endpoint is intended to capture and why the endpoint is tied to the time of the call.

The record also illustrates why outcome labels and operational definitions both matter. “Sensitivity” names a diagnostic performance measure, while the recorded operational description is a count of true cases correctly identified. A count and a sensitivity estimate are not interchangeable: interpreting a rate requires an appropriate denominator. The registry record does not report the observed count, denominator, or calculated sensitivity, so the distinction is important when assessing what can be learned from the record.

The wider study objectives span multiple stages: recognition, perceived breathing patterns, completion of instructions, bystander CPR, and survival. These are related questions, but each represents a different outcome and may require different definitions and analyses. The registry lists two outcome measures and two analyses, yet does not report details for the additional measure or the analysis methods. It is therefore not possible to determine how the objectives were linked analytically.

The design description also matters. The record does not describe a randomized comparison or a like-for-like control group. That structure can support descriptive assessment of calls, but it does not by itself establish a causal effect of an intervention. Clear reporting of denominators, case confirmation, missing observations, and uncertainty would be needed to interpret the diagnostic estimate fully. Those details and the numerical primary result are not reported in the registry record.

Finally, the registry’s completed status and enrollment figure should not be mistaken for evidence of a particular finding. Enrollment describes planned study size, while outcome results describe what was observed. Keeping design, endpoint definition, and measured results distinct is essential to sound interpretation.

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 gives an endpoint definition but does not report its observed result. It therefore does not permit assessment of the magnitude or precision of the primary diagnostic measure. Although the endpoint is titled “Sensitivity,” the record describes it as a count and does not state the numerical count, the denominator, or a calculated sensitivity estimate.

The record does not describe how cardiac arrest was confirmed, who reviewed the calls, or how ambiguous calls were classified. These details matter to the interpretation of diagnostic accuracy because they define what counts as a true case and how dispatcher recognition is compared with that classification. These procedures are not reported in the registry record.

The arm description refers to unconscious patients and potential cardiac arrest calls, while the primary endpoint population is confirmed cardiac arrest calls. The registry does not report the flow from initial calls to confirmed cases or to the analyzed population. The planned enrollment of 2260 cannot substitute for that information. Missing-data handling and reasons for exclusions are also not reported.

The registry describes no randomized comparison, says allocation is not stated, and does not identify a like-for-like control group. Thus, the record does not support a comparative causal interpretation. It also does not provide the methods used to address confounding, if any. Details about adjustment, subgroup analysis, or prespecification are not reported in the registry record.

Although the broader objectives mention instructions, bystander CPR, and survival, numerical findings for those outcomes are not reported. The serious adverse event fields show zero affected and zero deaths, but safety ascertainment methods are also not described. These limitations leave the record informative about purpose and intended measurement, but insufficient for evaluating the study’s results or the robustness of its conclusions.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Effectiveness of Dispatch-Assisted Cardiopulmonary Resuscitation (CPR) Instructions — NCT00664443.
• Registry record last updated 2024-09-19; status last verified 2024-04.
• Results first posted 2024-09-19.
• 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 record describes a completed study sponsored by the Ottawa Hospital Research Institute, with planned enrollment of 2260. The study evaluates 9-1-1 calls involving unconscious patients in participating communities, with an arm described as potential cardiac arrest calls. The study dates are 2007-07 to 2010-06, and the stated call-inclusion timeframe is January 2008 and October 2009.

The primary endpoint is titled “Sensitivity” and defined as the number of true cardiac arrest cases correctly identified at the time of the 9-1-1 call. The record does not report the observed endpoint result or a sensitivity estimate. It describes no randomized comparison, and allocation is not stated.

The broader objectives include phone-based diagnosis, perceived agonal breathing, completion of CPR instructions, bystander CPR, and survival. Numerical findings for these objectives, analysis details, missing-data methods, and the method of confirming cardiac arrest are not reported in the registry record. Serious adverse event fields list zero affected and zero deaths for the stated group, with no serious event terms listed.