Apixaban Drug Utilization Study In Stroke Prevention In Atrial Fibrillation (Spaf): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Apixaban Drug Utilization Study In Stroke Prevention In Atrial Fibrillation (Spaf) (NCT03441633), a registered phase not stated study. The primary outcome is Number of Participants by Their Sociodemographic Characteristics: Smoking Habit.

2. Clinical Question

3. Trial Design

The registry describes a completed drug-utilization study of apixaban in stroke prevention in atrial fibrillation, sponsored by Pfizer. It lists atrial fibrillation as the condition and reports enrollment of 51690. The study started on 2016-02-18, and primary completion was 2017-03-08. These dates describe the registry record; they do not specify the observation period for each participant.

The stated population comprises participants diagnosed with nonvalvular atrial fibrillation who had a first-recorded prescription for apixaban, VKA, dabigatran, or rivaroxaban in the SIDIAP database during the study. The registry separates treatment groups by drug and by prior VKA treatment in the 12 months before the index date. The arm descriptions use treatment-naïve and previously VKA-treated categories. Details of how eligibility, index dates, or treatment history were operationalized beyond those descriptions are not reported in the registry record.

Allocation is not stated, and the record explicitly cautions that the groups are not a like-for-like control comparison. The listed drugs should therefore be understood as exposure groups for utilization description, not as randomized arms. The registry does not report a protocol for assigning treatment or a design intended to establish comparative efficacy.

The stated primary outcome is the number of participants by smoking habit, measured at Day 1. The broader purpose described in the registry is to examine whether apixaban use accords with the approved indication and product-characteristic recommendations, including possible misuse or overuse. The registry lists 14 outcome measures and 196 analyses, but details of the analyses and their results are not reported in the registry record. The enrollment figure should not be assumed to be the denominator for every outcome or analysis.

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.

Only the first 6 of 10 groups are shown.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Study population included all participants with a first-recorded prescription of apixaban, VKA, dabigatran or rivaroxaban for SPAF registered in SIDIAP database during the study and diagnosed with NVAF.

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: Study population included all participants with a first-recorded prescription of apixaban, VKA, dabigatran or rivaroxaban for stroke prevention in atrial fibrillation(SPAF) registered in information system for the improvement of research in primary care(SIDIAP) database during the study and diagnosed with non-valvular atrial fibrillation(NVAF).

Only the first 6 of 11 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants by Their Sociodemographic Characteristics: Smoking Habit, measured over Day 1.

The primary endpoint is the number of participants by their sociodemographic characteristic of smoking habit, with a timeframe of Day 1. The outcome description identifies smoking habit as one characteristic used to describe the study population. It is a count-based, descriptive measure rather than a clinical event or a direct measure of stroke prevention.

Describing smoking habit can provide context about the population included in a drug-utilization study. In general, sociodemographic characteristics help readers understand whom the study describes and may be relevant when considering whether findings apply to other populations. The registry does not state a specific reason for selecting smoking habit as the primary measure, define smoking categories, or describe how smoking information was collected. Those details are not reported in the registry record.

No smoking-habit counts, category definitions, or denominators for this endpoint are reported in the registry record. Consequently, the measure’s values and the distribution of smoking habit across the drug groups cannot be assessed from the record. The endpoint also does not, by itself, establish whether apixaban was used according to the approved indication or whether any use constituted misuse or overuse. Those are broader study objectives, not conclusions that can be read from an endpoint title alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry presents a descriptive drug-utilization question. Its primary outcome is a count of participants by smoking habit, and the population is defined by diagnosis of nonvalvular atrial fibrillation and a first-recorded prescription for one of four listed drugs in the SIDIAP database during the study. The record does not report a statistical analysis plan, specific tests, model specifications, or methods for estimating uncertainty. The explanations below are general statistical context, not a description of methods confirmed to have been used in this study.

What does a count outcome tell the reader?

A count records how many participants fall into a defined category. For the stated primary outcome, the categories would depend on how smoking habit was classified, but those categories are not reported in the registry record. A count is easiest to interpret alongside its denominator: the number of people who could have been counted for that category. Without both category definitions and denominators, readers cannot infer the proportion of the population in a category or compare distributions across groups. No such values are reported in the registry record.

The record labels the outcome type as binary and describes its meaning as participants with an event (counts). That label does not supply smoking-category results or explain the statistical model, if any, used to analyze them. A binary outcome generally distinguishes whether a defined characteristic or event is present, while a count reports how many participants meet a definition. The exact coding and analysis for this endpoint are not reported in the registry record.

What does the observational grouping imply?

The treatment groups are based on the drug participants were receiving, and the registry states that allocation is not stated. In general, treatment groups in a nonrandomized utilization study may differ in ways that affect both prescribing and measured outcomes. A direct comparison can therefore mix the association of interest with differences in participant characteristics or treatment history. The registry’s warning that the groups are not a like-for-like control comparison is important: group labels alone do not justify a causal interpretation.

The arm descriptions distinguish participants who had no prior VKA prescription in the 12 months before the index date from participants who had been treated with VKA during that period. This is a stated classification, not evidence that all other relevant clinical or demographic differences were measured or controlled. The record does not report adjustment variables, matching, weighting, stratification, or another method for addressing differences between groups.

How should uncertainty be considered?

In general, a point estimate is a single summary of the observed data, while a confidence interval gives a range of values compatible with the data under the assumptions of the method. The interval’s width conveys precision: a wider interval indicates more uncertainty than a narrower one. The registry does not report estimates or confidence intervals for the primary outcome, so the precision of the smoking-habit counts cannot be assessed from the record.

For count or binary data, interpretation depends on how the outcome is defined, how many participants contribute information, and whether the analysis accounts for the study design and relevant differences between groups. A large overall enrollment does not guarantee precise estimates for every category or subgroup. The registry does not report outcome-specific denominators, missingness, or the distribution of observations needed to assess this.

What do the planning assumptions mean?

The registry’s sample-size assumptions describe a planning illustration computed from the observed effect, rather than the study’s own design. The stated 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 described as a stated fraction of the observed effect. These assumptions concern an illustration, not evidence that the study was designed around that calculation.

In general, a sample-size calculation estimates the amount of information needed to detect a specified effect under stated assumptions. Such calculations are sensitive to the chosen effect, significance level, allocation, and statistical method. An illustration based on an observed effect is not an observed-power analysis and does not show how likely the completed study was to detect an effect. The registry does not report a sample-size justification for the study itself. The numerical planning outputs and the observed effect used for the illustration are not reported in the registry record, so their implications cannot be evaluated from this record.

The registry also distinguishes the reporting unit for time-to-event sample-size illustrations from other outcomes: time-to-event entries give a total number of events, while other entries give participants per arm. That distinction is about the units used in planning displays; it does not establish that the primary smoking-habit endpoint was analyzed as a time-to-event outcome. No time-to-event primary endpoint is identified in the registry record.

What can be inferred from the analysis count?

The record lists 14 outcome measures and 196 analyses. In general, evaluating many outcomes or analyses raises a multiplicity issue: the more statistical tests considered, the greater the chance that some findings may appear notable by chance if each is interpreted without accounting for the full set. Approaches to multiplicity can include prespecifying a primary analysis or adjusting inference, but the registry does not report whether any approach was used. The analysis count alone does not indicate that the results are invalid; it does mean the reader needs analysis details and a stated interpretation plan, which are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 14 outcome measures and 196 analyses. It does not report whether these analyses were prespecified, how the primary outcome was distinguished from other outcomes for inferential purposes, or whether any correction or other strategy for multiplicity was used. In general, when many outcomes or analyses are examined, isolated findings can be difficult to interpret without a clear analysis plan. The counts alone do not establish that a multiplicity problem occurred, but the record does not provide the information needed to assess how it was handled.

Missing-data procedures are not reported in the registry record. The record does not specify how missing smoking-habit information or missing covariates were identified, whether participants with incomplete data were excluded, or whether any imputation or sensitivity analysis was undertaken. Without these details, the completeness of the primary endpoint and the potential effect of missingness on its interpretation cannot be assessed.

The registry defines the study population as participants with a first-recorded prescription for apixaban, VKA, dabigatran, or rivaroxaban in SIDIAP during the study and a diagnosis of nonvalvular atrial fibrillation. It does not report outcome-specific analysis populations, exclusions after eligibility, or denominators for the registered analyses. Enrollment of 51690 should therefore not be treated as the denominator for every analysis. Details about participant flow, the handling of repeated prescriptions, and the statistical population for each outcome are not reported in the registry record.

Allocation is not stated, and the registry cautions that the groups are not a like-for-like control comparison. Methods addressing differences between treatment groups, if any, are not reported in the registry record. This limits assessment of both comparative interpretation and the robustness of any estimates.

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

The registry record establishes that a completed drug-utilization study was registered with enrollment of 51690 and that its stated primary measure was the number of participants by smoking habit at Day 1. It also describes a broader objective: assessing whether apixaban was used in line with its approved indication and product-characteristic recommendations, including estimating possible misuse or overuse. These are study aims; the numerical results needed to judge them are not reported in the registry record.

The record does not report smoking-habit counts, outcome-specific denominators, or estimates of use in accordance with recommendations. It therefore does not show the distribution of smoking habit, how use varied between drug groups, or the extent of possible misuse or overuse. Nor does the endpoint title establish whether the broader utilization objective was achieved.

The drug groups are not described as randomized, allocation is not stated, and the record cautions that they are not a like-for-like control comparison. Any group contrasts, even if reported elsewhere, would require careful interpretation because the record does not describe how differences between groups were addressed. The available information cannot support a causal claim that one drug caused a difference in a characteristic or outcome.

The adverse-event summary lists zero counts in its reported fields across the listed groups and no top serious terms. That summary should not be read as proof that no adverse events occurred. The registry does not describe how events were collected, the completeness of reporting, or the interpretation of those zero entries. In general, an absence of recorded events is not equivalent to evidence that an event could not occur.

Overall, the registry record describes the intended population, drug groups, and primary descriptive measure, but does not provide the outcome values or analysis details needed to quantify findings or their precision. Completion status and enrollment do not fill that evidential gap.

12. Why This Trial Matters Statistically

Drug-utilization studies address how medicines are used in practice, a question distinct from whether a medicine is efficacious under randomized trial conditions. This record frames the study around apixaban use among people with nonvalvular atrial fibrillation and explicitly identifies the assessment of use according to approved recommendations, including possible misuse or overuse. That focus can be important for understanding prescribing patterns and the population receiving treatment.

The registry also illustrates why a clearly named endpoint matters. Its primary measure is a participant count by smoking habit, a descriptive sociodemographic characteristic, whereas the broader stated objective concerns use in accordance with recommendations. These questions are related only insofar as a population description can provide context; the smoking-habit endpoint alone does not answer whether prescribing was appropriate. Outcome definitions and reported results are needed to connect a descriptive measure to the broader objective.

The record’s explicit caution about non-comparable groups is statistically important. When treatment exposure is not assigned at random, differences across drug groups should not automatically be interpreted as effects of the drugs. Clear descriptions of the population, treatment history, analysis approach, and uncertainty are essential for readers to judge what a utilization study can support.

Finally, a registry entry can convey design and intended outcomes without conveying results. Here, the record lists a large enrollment and numerous outcome and analysis entries, but does not report the primary counts, comparative estimates, or methods needed to assess precision and multiplicity. Distinguishing study aims, observed findings, and limitations of interpretation is central to responsible statistical reading.

13. 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 smoking-habit counts, their category definitions, or the corresponding denominators. This prevents assessment of the endpoint’s distribution and precision. Although enrollment is reported as 51690, the record does not show how many participants contributed to each outcome or analysis, so that figure cannot be assumed to represent the analysis population for every result.

The record describes treatment groups based on apixaban, VKA, dabigatran, and rivaroxaban, with further distinctions by prior VKA treatment. Allocation is not stated, and the registry explicitly says the arms are not a like-for-like control comparison. The record does not report methods for controlling confounding or other differences between groups. Thus, the registry information does not support causal interpretation of group contrasts.

Important operational details are not reported in the registry record, including the definitions and collection of smoking habit, outcome-specific eligibility and exclusions, missing-data handling, and participant flow. The record lists 14 outcomes and 196 analyses but does not describe the analysis plan, statistical methods, or multiplicity procedures. These omissions limit evaluation of how the results were generated and how selective interpretation was avoided.

The adverse-event summary contains zero counts in its listed fields and no top serious terms, but the record does not explain ascertainment or reporting completeness. Those entries cannot establish the absence of adverse events. Finally, the registry’s brief summary states the utilization aims, but the findings needed to judge whether apixaban use followed recommendations or involved misuse or overuse are not reported in the registry record.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Apixaban Drug Utilization Study In Stroke Prevention In Atrial Fibrillation (Spaf) — NCT03441633.
• Registry record last updated 2023-06-09; status last verified 2023-06.
• Results first posted 2019-04-19.
• 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.

17. Record Summary

This completed Pfizer-sponsored drug-utilization study concerns apixaban use in stroke prevention among people with atrial fibrillation. The registry reports enrollment of 51690, a study start date of 2016-02-18, and primary completion on 2017-03-08. The stated population comprises participants diagnosed with nonvalvular atrial fibrillation and a first-recorded prescription for apixaban, VKA, dabigatran, or rivaroxaban in the SIDIAP database during the study.

The primary outcome is the number of participants by smoking habit at Day 1. The broader objective is to assess whether apixaban use followed its approved indication and product-characteristic recommendations and to estimate possible misuse or overuse. Allocation is not stated, and the registry cautions that the drug groups are not like-for-like controls. Although 14 outcomes and 196 analyses are listed, outcome values, analysis methods, and missing-data and multiplicity procedures are not reported in the registry record. The adverse-event summary lists zero counts in its reported fields, with no top serious terms; reporting details are not given.