This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
TACTiC was a completed phase 3, single-group study in consumers with high cholesterol. The study evaluated 5 mg rosuvastatin calcium with a Web App and reported three primary endpoints: two binary TASS outcomes and percent change from baseline in verified LDL-C at the final assessment.
| Feature | TACTiC |
|---|---|
| Trial name | Technology-Assisted Cholesterol Trial in Consumers (TACTiC) |
| Phase | Phase 3 |
| Status | Completed |
| Condition | High Cholesterol |
| Design | Single-group |
| Allocation | NA |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 1196 |
| Arms | 1 |
| Intervention | 5 mg rosuvastatin calcium with a Web App (combination product) |
| Lead sponsor | AstraZeneca |
| Sponsor type | Industry |
| Start | 2021-07-08 |
| Primary completion | 2023-03-01 |
| ClinicalTrials.gov | NCT04964544 |
2. Clinical Question
The statistical question in TACTiC is different from the usual two-arm randomized comparison. The study enrolled one treatment group and evaluated whether participants achieved correct TASS outcomes at specified assessments and whether verified LDL-C changed from baseline by the final assessment.
Population
Consumers with high cholesterol enrolled in a phase 3, single-group treatment study.
Intervention
5 mg rosuvastatin calcium with a Web App, described in the registry as a combination product.
Comparator
There was no separate comparator arm. The registry describes the allocation as NA and the design model as single-group.
Primary question
What proportion of participants had a correct TASS outcome at the initial and final assessments, and what was the percent change from baseline in verified LDL-C at 180 days?
3. Trial Design
Nonprescription Rosuvastatin
- 5 mg rosuvastatin calcium
- Web App
- Combination product
- Home use period through 180 days
No randomized comparator
- Allocation: NA
- Design model: single-group
- Masking: none
- Primary purpose: treatment
The single-group design is central to interpreting the reported estimates. The percentages for the TASS endpoints describe the observed proportion within the study population; they are not treatment-versus-control effect estimates. Likewise, the reported LDL-C change describes change within the study population rather than a randomized between-group difference.
4. Study Timeline and Assessment Framework
Study start
The registry records the start of TACTiC on July 8, 2021.
Initial TASS assessment
The first co-primary endpoint evaluated the proportion of participants with a correct TASS outcome at their initial TASS assessment.
End of home use period
The second co-primary endpoint and the verified LDL-C endpoint were assessed at the end of the home use period at 180 days.
Primary completion
The registry records March 1, 2023 as the primary completion date.
5. Primary Endpoints
TACTiC registered three primary endpoints. The first two are proportions of participants with a correct TASS outcome. The third is percent change from baseline in verified LDL-C, regardless of final use outcome, in the AUS ITT population.
| Endpoint | Time frame | Definition / role |
|---|---|---|
| Overall Correct Initial TASS Outcome, Worst Case Imputation, First Co-Primary Endpoint (Self-Selection Population) | Study day -30 to -1, at initial TASS assessment | Proportion of participants that had a correct tass outcome at their initial TASS assessment. |
| Overall Correct Final Use Outcome With Mitigation, Second Co-primary Endpoint (Per Protocol Population) | From enrollment to end of the home use period at 180 days, at final TASS Assessment | Proportion of participants that had a correct tass outcome at their final TASS assessment. |
| Percent Change From Baseline in Verified LDL-C Regardless of Final Use Outcome (AUS ITT Population) | From enrollment to end of the home use period at 180 days, at final assessment | Percent change from baseline in verified LDL-C regardless of final use outcome. |
6. Statistical Methodology
What the registry reports
The ClinicalTrials.gov record posts three formal statistical analyses corresponding to the three primary endpoints. For all three analyses, the statistical method is listed as not reported. The reported effect measure is a percentage for the two TASS endpoints and mean percent change from baseline for the verified LDL-C endpoint.
| Primary analysis | Analysis population | Reported measure | Method |
|---|---|---|---|
| Initial TASS outcome | Self-Selection Population | Percentage | Not reported |
| Final TASS outcome | Per Protocol Population | Percentage | Not reported |
| Verified LDL-C change | AUS ITT Population | Mean percent change from baseline | Not reported |
Binary endpoint analysis
The two TASS endpoints are binary outcomes: each participant either had a correct TASS outcome or did not. The registry reports the resulting proportion and a two-sided 95% confidence interval. A confidence interval for a binary proportion quantifies uncertainty around the estimated proportion, but the registry does not identify the particular interval construction used.
Continuous change endpoint
The verified LDL-C endpoint is reported as mean percent change from baseline. For an endpoint of this type, an analysis of within-participant change would ordinarily summarize the distribution of change and quantify uncertainty around the mean. The registry reports the mean percent change and its two-sided 95% confidence interval, but does not report the underlying statistical method.
Per-protocol analysis
The second co-primary endpoint was analyzed in the Per Protocol Population. A per-protocol analysis focuses on participants meeting the prespecified protocol criteria rather than automatically retaining every enrolled participant. This can provide information about outcomes among participants who followed the protocol-defined treatment and assessment framework, but it can also introduce selection concerns because protocol adherence may be related to participant characteristics or outcomes.
Worst case imputation
Both TASS primary analyses explicitly include worst case imputation in their registered endpoint names and analysis descriptions. This means the reported TASS proportions are intended to incorporate the registry-specified treatment of missing outcome information rather than simply excluding all participants without an observed assessment. The record does not provide the detailed imputation algorithm beyond the designation of worst case imputation.
Confidence intervals
All three posted primary analyses report two-sided 95% confidence intervals. The intervals quantify statistical uncertainty around the corresponding point estimates. They do not describe the range of individual participant responses and should not be interpreted as a prediction interval for a future participant.
7. Primary Results
The registry posts formal analyses for all three primary endpoints. Because TACTiC has a single treatment group, the estimates are descriptive within the relevant analysis populations rather than randomized treatment effects against a comparator.
First Co-Primary Endpoint: Initial TASS Outcome
Overall correct initial TASS outcome
95% CI: 88.9–92.3 · Two-sided 95% CI
Self-Selection Population; worst case imputation; study day -30 to -1, at initial TASS assessment.
The estimate of 90.7% means that the reported proportion of participants in the Self-Selection Population with a correct initial TASS outcome was 90.7%, using the registry-specified worst case imputation.
It does not represent a treatment effect relative to another therapy because the study has one group and no comparator arm. It also does not mean that 90.7% of all consumers with high cholesterol would necessarily obtain a correct TASS outcome outside the study population.
The 95% CI of 88.9–92.3 describes uncertainty around the estimated proportion under the statistical framework used for the reported analysis. The relatively narrow interval indicates that the estimated proportion was measured with greater precision than would be conveyed by the point estimate alone.
No p-value is reported for this analysis in the ClinicalTrials.gov record. A p-value would answer a hypothesis-testing question defined relative to a null hypothesis; it would not measure the size or practical meaning of the 90.7% estimate.
Second Co-Primary Endpoint: Final TASS Outcome
Overall correct final use outcome with mitigation
95% CI: 97.1–98.8 · Two-sided 95% CI
Per Protocol Population; worst case imputation; final TASS assessment at the end of the home use period at 180 days.
The estimate of 98.1% means that 98.1% of participants in the Per Protocol Population were reported to have a correct final TASS outcome with mitigation at the final TASS assessment.
The estimate is not a randomized comparison and therefore does not quantify the incremental effect of rosuvastatin or the Web App versus another treatment. It describes the observed outcome proportion within the specified per-protocol population.
The 95% CI of 97.1–98.8 expresses uncertainty around the proportion. Because this endpoint was evaluated in a Per Protocol Population, the precision of the estimate should not be confused with the representativeness of that population for all enrolled participants or all consumers with high cholesterol.
No p-value is reported for this analysis in the ClinicalTrials.gov record. The p-value, if one were specified against a particular null proportion, would not quantify the magnitude of the observed 98.1% estimate.
Third Co-Primary Endpoint: Verified LDL-C
Mean percent change from baseline in verified LDL-C
95% CI: −36.63 to −34.33 · Two-sided 95% CI
AUS ITT Population; regardless of final use outcome; final assessment after the 180-day home use period.
The estimate of −35.48% means that the reported mean percent change from baseline in verified LDL-C at the final assessment was negative 35.48% in the AUS ITT Population.
In descriptive terms, the negative sign indicates a decrease from baseline. It does not represent a treatment-versus-control difference because TACTiC did not include a separate comparator group. It also should not be interpreted as saying that every participant experienced exactly a 35.48% reduction.
The 95% CI of −36.63 to −34.33 describes uncertainty around the mean percent-change estimate. It provides a statistical interval for the population mean change under the analysis framework represented by the reported estimate.
No p-value is reported for this analysis in the ClinicalTrials.gov record. A p-value would depend on the null hypothesis being tested and would not itself quantify the magnitude of the LDL-C change.
8. Comparing the Three Primary Endpoints
| Endpoint | Population | Estimate | 95% CI | Interpretive role |
|---|---|---|---|---|
| Initial TASS outcome | Self-Selection Population | 90.7% | 88.9–92.3 | Proportion with a correct initial TASS outcome |
| Final TASS outcome | Per Protocol Population | 98.1% | 97.1–98.8 | Proportion with a correct final TASS outcome |
| Verified LDL-C change | AUS ITT Population | −35.48% | −36.63 to −34.33 | Mean percent change from baseline |
The three estimates should not be placed on a common effect scale. The first two are proportions of participants achieving a binary outcome, while the third is a mean percent change from baseline. In addition, the first two use different analysis populations. Consequently, the numerical estimates describe different aspects of the study rather than three interchangeable measures of treatment efficacy.
9. Safety Results
The registry reports serious adverse events by arm as 27/1177 for the Nonprescription Rosuvastatin group. Because TACTiC is a single-group study, there is no randomized safety comparator in the registry record.
| Safety measure | Nonprescription Rosuvastatin |
|---|---|
| Serious adverse events | 27/1177 affected / at risk |
The denominator reported for the serious-adverse-event summary is 1177, compared with 1196 participants enrolled in the study. The registry therefore presents the safety measure using its own affected/at-risk population rather than the overall enrollment count.
10. Statistical Methods Explained
Why are the TASS endpoints analyzed as proportions?
Each TASS endpoint is represented as a binary outcome: participants either had a correct TASS outcome or did not. The natural summary is therefore the proportion with the specified outcome. The registry reports 90.7% for the initial assessment and 98.1% for the final assessment.
What does a 95% confidence interval around a proportion mean?
A confidence interval describes uncertainty around the estimated population proportion. For the initial TASS endpoint, the estimate is 90.7% with a 95% CI of 88.9–92.3. For the final TASS endpoint, the estimate is 98.1% with a 95% CI of 97.1–98.8. The interval is about uncertainty in the estimated proportion, not about the range of individual participant outcomes.
Why does the Per Protocol Population matter for the final TASS endpoint?
The final TASS endpoint is specifically defined for the Per Protocol Population. A per-protocol analysis focuses on participants meeting protocol-defined criteria. That can describe outcomes under protocol adherence, but it can also produce a population that differs systematically from the full enrolled population. Therefore, a highly precise per-protocol estimate is not automatically a broadly generalizable estimate.
What does the −35.48% LDL-C estimate mean?
The reported mean percent change from baseline was −35.48% in the AUS ITT Population at the final assessment. The negative value represents a decrease from baseline on the percent-change scale. The estimate is a within-study change, not a randomized treatment effect against a control group.
Why is the LDL-C confidence interval also important?
The 95% CI of −36.63 to −34.33 shows the uncertainty around the mean percent-change estimate. It gives more information than the point estimate alone because two estimates with the same mean change can have very different levels of statistical precision.
Why can't the TASS estimates be interpreted as treatment effects?
A treatment effect requires a comparison that defines what would have happened under an alternative condition. TACTiC is a single-group study with no separate comparator arm. Its reported TASS percentages therefore describe observed outcomes in the study population rather than a randomized contrast between two treatment strategies.
What is the role of worst case imputation?
The first and second primary endpoint names explicitly specify worst case imputation. Missing TASS outcomes are therefore not simply ignored in the reported endpoint definition. The registry does not provide further detail on the imputation algorithm, so the interpretation should remain tied to the registry's stated worst-case framework rather than assuming a particular mathematical implementation.
11. Confidence Intervals Versus P-Values
All three primary analyses report two-sided 95% confidence intervals. None of the three posted analyses includes a p-value in the reported analysis information.
Confidence interval
The interval communicates statistical uncertainty around the corresponding estimate. It also provides a range of effect values that can be considered in interpreting the estimate under the specified statistical framework.
P-value
A p-value is defined relative to a null hypothesis. It does not measure the size of an effect, the clinical importance of an effect, or the probability that the null hypothesis is true.
For TACTiC, the absence of a reported p-value is particularly important because the study is single-group. A hypothesis test for a proportion or mean change would require an explicitly specified reference value. The registry does not state such a null value for the three posted primary analyses.
12. Analysis Populations and What They Answer
| Population | Primary endpoint | What the analysis describes |
|---|---|---|
| Self-Selection Population | Initial TASS outcome | Correct initial TASS outcomes using worst case imputation |
| Per Protocol Population | Final TASS outcome | Correct final TASS outcomes with mitigation using worst case imputation |
| AUS ITT Population | Verified LDL-C | Mean percent change from baseline regardless of final use outcome |
The distinction between these populations is statistically important. An ITT-style population generally preserves the broad enrollment framework, while a per-protocol population conditions the analysis on protocol-defined participation. The TACTiC registry explicitly assigns different primary endpoints to different analysis populations, so the three estimates should be interpreted within their respective populations.
13. Single-Group Design and Causal Interpretation
The design model is single-group, with allocation listed as NA and masking listed as none. This design can efficiently characterize outcomes after exposure to an intervention, but it does not provide the same internal randomized comparison as a parallel-group controlled trial.
A within-group outcome can be estimated precisely without establishing how the outcome would have differed under another intervention or under no treatment.
This distinction applies to both major result types in TACTiC. A 98.1% final TASS outcome is a proportion within the specified Per Protocol Population. A −35.48% mean percent change in verified LDL-C is a within-study change. Neither estimate is a between-arm causal contrast.
14. Interpreting the LDL-C Endpoint
The third primary endpoint is explicitly defined as percent change from baseline in verified LDL-C regardless of final use outcome, measured from enrollment through the end of the 180-day home use period at final assessment.
Reported mean change
95% CI: −36.63 to −34.33
AUS ITT Population; final assessment at 180 days.
Percent change is a relative within-participant measure. Conceptually, it compares a participant's final measurement with that participant's baseline measurement before summarizing those changes across the analysis population. Because the registry reports the result as a mean percent change, the reported −35.48% should be understood as the average on that percent-change scale, not necessarily the percent change of an average LDL-C value.
15. Interpreting the Two TASS Endpoints
The two TASS endpoints examine different stages of the study experience. The first evaluates the initial TASS assessment during study days -30 to -1. The second evaluates the final use outcome with mitigation at the final TASS assessment after the 180-day home use period.
| Feature | Initial TASS | Final TASS |
|---|---|---|
| Estimate | 90.7% | 98.1% |
| 95% CI | 88.9–92.3 | 97.1–98.8 |
| Population | Self-Selection Population | Per Protocol Population |
| Imputation | Worst case | Worst case |
| Assessment | Initial TASS assessment | Final TASS assessment |
| Time frame | Study day -30 to -1 | Enrollment through 180 days |
The numerical increase from 90.7% to 98.1% should not be interpreted as a formal longitudinal treatment effect. The two estimates come from different analysis populations and correspond to different registered endpoint definitions. A direct comparison would therefore mix both endpoint definition and population changes.
16. Precision of the Reported Estimates
The confidence intervals are relatively narrow around both TASS point estimates. The initial TASS interval spans 88.9–92.3, while the final TASS interval spans 97.1–98.8. For LDL-C, the mean percent-change estimate has a 95% CI spanning −36.63 to −34.33.
Precision and validity are different concepts. A narrow confidence interval indicates statistical precision for the estimated quantity; it does not by itself establish that the study population is representative of a broader population or that a single-group association is causal.
17. What the Registry Does Not Establish From These Estimates
The reported results support several descriptive statements, but the design places important limits on comparative interpretation.
- No randomized comparator: the registry describes a single-group study with no separate control arm.
- No between-group treatment effect: the reported percentages and mean change are not treatment-versus-control contrasts.
- No p-values for the posted primary analyses: the registry reports point estimates and two-sided 95% confidence intervals but does not report p-values for these analyses.
- Different analysis populations: the initial and final TASS endpoints use different populations, while the LDL-C endpoint uses the AUS ITT Population.
- Method not reported: the statistical method underlying each posted analysis is listed as not reported.
- Limited interpretation of precision: confidence intervals quantify uncertainty around estimates but do not resolve generalizability or causal questions created by the single-group design.
18. Study Limitations
- Single-group design: without a randomized comparator, the observed TASS and LDL-C outcomes cannot be interpreted as randomized treatment effects.
- Different analysis populations: the three primary endpoints are tied to different analysis populations, which complicates direct numerical comparison across endpoints.
- Per-protocol analysis: the final TASS endpoint is evaluated in the Per Protocol Population, which can differ systematically from the enrolled population.
- Worst case imputation: the first two primary endpoints incorporate worst case imputation, but the registry does not provide the detailed imputation algorithm.
- Statistical method not reported: the registry posts estimates and confidence intervals but identifies the statistical method for all three analyses as not reported.
- No reported p-values: the posted analyses provide confidence intervals without corresponding p-values or stated null hypotheses.
- COVID-19 pandemic: the registry notes that TACTiC was conducted during the COVID-19 pandemic, while also stating that the pandemic was not judged to adversely impact the quality of the study.
- Safety comparison: serious adverse events are reported for the single study group, so there is no randomized comparator for contextualizing that safety count.
19. COVID-19 Pandemic Consideration
This is a study-level contextual consideration rather than a numerical efficacy result. It is relevant when considering the operational environment in which participant assessments and home use occurred, but the registry does not provide a quantitative analysis attributing any portion of the reported endpoint estimates to the pandemic.
20. Why This Trial Matters Statistically
TACTiC is useful as a teaching case because it demonstrates a different statistical structure from the randomized two-arm trials that dominate clinical-trial examples. Its central challenge is not estimation of a hazard ratio between treatment groups; it is understanding how proportions, mean change, analysis populations, missing-data handling, and confidence intervals function in a single-group study.
| Concept | How it appears in TACTiC |
|---|---|
| Single-group design | One intervention group with allocation listed as NA |
| Binary endpoint | Initial and final TASS outcomes are registered as binary endpoints |
| Proportion | Initial TASS estimate of 90.7% and final TASS estimate of 98.1% |
| Confidence interval | Two-sided 95% CIs accompany all three primary estimates |
| Per-protocol analysis | Final TASS endpoint analyzed in the Per Protocol Population |
| Worst case imputation | Specified for both TASS primary endpoints |
| ITT population | Verified LDL-C endpoint analyzed in the AUS ITT Population |
| Mean percent change | Verified LDL-C reported as mean percent change from baseline |
| Analysis-population effects | Different primary endpoints use different populations |
| Statistical-method reporting | Posted analyses provide estimates and CIs while the statistical method is listed as not reported |
21. Statistical Interpretation: Precision, Effect Size, and Causal Effect
The confidence intervals communicate how precisely the study estimates the reported quantities. For example, the final TASS estimate is 98.1% with a 95% CI of 97.1–98.8. Precision is a property of the estimate and its uncertainty; it is not evidence of comparative treatment benefit.
The magnitude of the reported quantities is directly visible in the point estimates: 90.7%, 98.1%, and −35.48%. But the meaning of each number depends on its endpoint definition and analysis population. They should not be converted into a common effect measure merely for ease of comparison.
Because TACTiC is single-group, the observed outcomes cannot by themselves establish how outcomes would have differed under another intervention. A causal treatment effect requires a suitable counterfactual comparison, which is not reported by a one-arm design.
22. Results in Statistical Context
The three primary estimates provide complementary information:
- 90.7% initial TASS outcome: describes the proportion with a correct initial TASS assessment in the Self-Selection Population, with worst case imputation.
- 98.1% final TASS outcome: describes the proportion with a correct final TASS assessment in the Per Protocol Population, with mitigation and worst case imputation.
- −35.48% verified LDL-C change: describes the mean percent change from baseline in verified LDL-C in the AUS ITT Population at final assessment.
The first two outcomes concern correct TASS assessments, while the third concerns a biomarker change. Their statistical scales, populations, and interpretations differ. The most defensible reading therefore keeps each estimate attached to its original endpoint definition rather than treating the three values as components of a single composite efficacy statistic.
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: TACTiC (NCT04964544).
- Linked publication: PubMed record for PMID 38599257.
Continue through Clinical Biostats
Explore statistical tutorials and calculators that develop the estimation and analysis concepts illustrated by this trial.
26. Record Summary
TACTiC provides a useful example of statistical interpretation in a phase 3 single-group clinical study. The registry reports three primary endpoints with formal posted analyses: a 90.7% overall correct initial TASS outcome with a 95% CI of 88.9–92.3, a 98.1% overall correct final use outcome with a 95% CI of 97.1–98.8, and a −35.48% mean percent change from baseline in verified LDL-C with a 95% CI of −36.63 to −34.33. The first two endpoints use worst case imputation, while the final TASS endpoint is evaluated in the Per Protocol Population and the LDL-C endpoint in the AUS ITT Population.
The central statistical lesson is that precise estimation and causal comparison are different objectives. TACTiC provides confidence intervals around descriptive single-group estimates, but its single-group design does not supply a randomized comparator against which a treatment effect can be estimated. The analysis-population definitions, endpoint-specific time frames, worst case imputation, and distinction between binary proportions and mean percent change are therefore essential to interpreting the reported results.