This page separates reported trial results from statistical interpretation. Numerical results are restricted to the ClinicalTrials.gov record. The registry notes that the March data cutoff of 17 May 2018 was not the final analysis for overall survival, time to first subsequent therapy or death, time to second subsequent therapy or death, and time to treatment discontinuation or death.
1. Trial at a Glance
SOLO1 was a randomized, parallel-group, quadruple-masked phase 3 trial evaluating olaparib 300 mg tablets as maintenance monotherapy versus placebo in patients with BRCA-mutated advanced ovarian cancer following first-line platinum-based chemotherapy and a clinical complete or partial response.
| Feature | SOLO1 |
|---|---|
| Trial name | SOLO1 |
| ClinicalTrials.gov identifier | NCT01844986 |
| Phase | Phase 3 |
| Status | Active, not recruiting |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 450 |
| Lead sponsor | AstraZeneca |
| Sponsor type | Industry |
| Primary endpoint type | Time-to-event |
| Results posted | Yes |
| Outcome measures posted | 9 |
| Statistical analyses posted | 16 |
2. Clinical Question
The primary question was whether olaparib maintenance monotherapy improved progression-free survival compared with placebo in patients with BRCA-mutated high-risk advanced ovarian cancer who were in clinical complete response or partial response following first-line platinum-based chemotherapy.
Population
Patients with newly diagnosed, advanced ovarian cancer, including FIGO stage III-IV disease, BRCA mutation, and clinical complete or partial response following first-line platinum chemotherapy.
Intervention
Olaparib 300 mg tablets as maintenance monotherapy.
Comparator
Placebo tablets.
Primary question
Does olaparib maintenance monotherapy improve investigator-assessed progression-free survival compared with placebo?
3. Trial Design
Olaparib 300 mg tablets
- Olaparib 300 mg tablets
- Maintenance monotherapy
- Global cohort and China cohort analyses were posted
Placebo tablets
- Placebo tablets
- Maintenance comparator
- Global cohort and China cohort analyses were posted
The trial's statistical profile is dominated by time-to-event analysis. The registry reports Cox proportional-hazards models for progression-free survival and several subsequent efficacy endpoints, together with a mixed-effects model for longitudinal health-related quality of life.
4. Primary Endpoint
| Endpoint | Registry definition and time frame | Statistical approach |
|---|---|---|
| Progression Free Survival (PFS) Using Investigator Assessment According to Modified Response Evaluation Criteria in Solid Tumours (RECIST 1.1) | To determine the efficacy by progression free survival (PFS) using investigator assessment according to modified Response Evaluation Criteria in Solid Tumours (RECIST 1.1) of olaparib maintenance monotherapy compared to placebo in BRCA mutated high risk advanced ovarian cancer patients who are in clinical complete response or partial response following first line platinum based chemotherapy. Radiologic scans performed at baseline then every 12 weeks up to 156 weeks, then every 24 weeks thereafter until objective radiological disease progression. DCO: 17 May 2018 | Cox proportional-hazards model; superiority hypothesis |
The registry identifies PFS as the sole registered primary endpoint. The posted primary analyses include both the global cohort and the China cohort. Both use hazard ratios, 95% two-sided confidence intervals, and Cox regression.
5. Statistical Analysis Populations
The primary global analysis used a Global Full Analysis Set (FAS) consisting of all patients randomized as part of global recruitment, including 5 patients randomized in China. The registry also identifies a China Full Analysis Set for the China-specific analyses.
| Population | Role in the posted analyses |
|---|---|
| Global Full Analysis Set | All patients randomized as part of global recruitment, including 5 patients randomized in China; used for the global efficacy analyses. |
| China Full Analysis Set | Used for the China cohort analyses. |
| Global HRQoL analysis population | Global Full Analysis Set with a baseline and post-baseline TOI score available. |
| Myriad gBRCAm population | Global Full Analysis Set consisting of patients confirmed as Myriad gBRCAm for the BRCA-specific PFS analysis. |
The repeated appearance of the Full Analysis Set in the registry is statistically important. Treatment comparisons remain anchored to randomized assignment rather than being restricted to patients who completed a particular amount of treatment. For the HRQoL analysis, however, the registry specifically requires baseline and post-baseline TOI scores.
6. Results: Primary Progression-Free Survival
The registry reports a formal primary analysis of investigator-assessed PFS for both the global cohort and the China cohort. The analysis uses Cox regression, with a hazard ratio as the effect measure and a superiority hypothesis.
Global Cohort
Hazard ratio for progression-free survival
95% CI: 0.23–0.41 · P < 0.0001
Olaparib 300 mg tablets versus placebo; two-sided 95% confidence interval.
A hazard ratio of 0.30 means that, under the fitted Cox model, the estimated instantaneous rate of progression or the relevant PFS event in the olaparib group was approximately 30% of that in the placebo group. Equivalently, the estimated hazard was approximately 70% lower with olaparib under this model.
The hazard ratio does not mean that 70% of patients avoided progression, that every patient experienced the same reduction, or that the probability of remaining progression-free at every particular time point was exactly 70% higher.
The 95% confidence interval of 0.23–0.41 describes statistical uncertainty around the estimated hazard ratio under the analysis model. It is not a range containing the individual treatment effects experienced by patients.
The reported P < 0.0001 addresses the statistical evidence against the null hypothesis specified for the superiority analysis. It does not measure the size or clinical importance of the effect. The effect size is conveyed by the hazard ratio and its confidence interval.
Because the analysis is a Cox proportional-hazards analysis, interpretation of a single hazard ratio also depends on the proportional-hazards framework. The registry does not provide additional information here that would allow a separate assessment of that assumption.
China Cohort
China cohort PFS hazard ratio
95% CI: 0.23–0.97 · Two-sided
Olaparib 300 mg tablets versus placebo.
The China-cohort estimate of 0.46 corresponds to an estimated hazard approximately 46% of that in the placebo group under the fitted model. In relative terms, this is consistent with an estimated hazard approximately 54% lower under the model.
The 95% CI of 0.23–0.97 is substantially wider than the global-cohort interval of 0.23–0.41. That difference in precision is important when interpreting the China analysis: a subgroup or cohort-specific estimate is not interchangeable with the global estimate.
The registry states that the model includes the stratification variable of response to first-line platinum chemotherapy as a covariate. It also identifies intention-to-treat analysis, covariate adjustment, and stratified analysis as concepts associated with this analysis.
No P-value is reported in the registry analysis for the China cohort. The confidence interval therefore provides the principal registry-reported measure of statistical uncertainty for this cohort-specific result.
7. Primary PFS Results Side by Side
| Analysis | Population | Effect measure | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Primary global analysis | Global Full Analysis Set | Hazard ratio | 0.30 | 0.23–0.41 | <0.0001 |
| Primary China analysis | China Full Analysis Set | Hazard ratio | 0.46 | 0.23–0.97 | Not reported in the ClinicalTrials.gov record |
8. Secondary Time-to-Event Results
The registry contains formal Cox-model analyses for several secondary time-to-event endpoints. These analyses provide a broader statistical picture of disease progression, subsequent therapy, treatment discontinuation, and survival.
Overall Survival
Global overall survival
95% CI: 0.60–1.53 · P = 0.8903
Olaparib versus placebo in the Global Full Analysis Set.
The global OS hazard ratio of 0.95 is close to 1. Under the Cox model, the estimated instantaneous rate of death was therefore similar between the randomized groups in this posted analysis.
The 95% CI of 0.60–1.53 is wide enough to encompass values below and above 1. This means the ClinicalTrials.gov record does not establish a precise direction or magnitude of an OS effect from this analysis alone.
The P = 0.8903 result is evidence that this particular superiority comparison did not produce a statistically small P-value. It should not be interpreted as proof that the two treatments are identical, and the P-value itself is not a measure of treatment effect size.
Overall Survival — China Cohort
| Endpoint | Estimate | 95% CI | P-value |
|---|---|---|---|
| Overall survival, China cohort | HR 0.77 | 0.29–2.28 | Not reported |
The China-cohort OS estimate has a substantially wider confidence interval than the global estimate. The interval includes values below and above 1, so the ClinicalTrials.gov record does not establish a precise treatment effect for this cohort-specific analysis.
Time to Earliest Progression by RECIST, CA-125, or Death
Global cohort
95% CI: 0.23–0.40 · P < 0.0001
This endpoint broadens the event definition to the earliest progression by RECIST or Cancer Antigen (CA-125), or death. The global hazard ratio of 0.30 indicates an estimated event hazard approximately 70% lower with olaparib under the Cox model.
The 95% CI of 0.23–0.40 indicates relatively tight statistical precision around the estimated relative hazard. The P-value of <0.0001 indicates strong statistical evidence under the specified superiority comparison, but it does not itself quantify the size of the treatment effect.
| Analysis | Estimate | 95% CI | P-value |
|---|---|---|---|
| Global cohort | HR 0.30 | 0.23–0.40 | <0.0001 |
| China cohort | HR 0.47 | 0.23–0.99 | Not reported |
Time From Randomization to Second Progression
Global cohort
95% CI: 0.35–0.72 · P = 0.0002
The global hazard ratio of 0.50 corresponds to an estimated instantaneous event rate approximately half that of the placebo group under the Cox model. The registry defines this endpoint as time from randomization to second progression, following first progression disease and then assessment per local practice every 12 weeks until second progression.
The 95% CI of 0.35–0.72 provides the uncertainty interval around the estimate. The P-value of 0.0002 indicates evidence against the null hypothesis specified for this superiority analysis, but it does not mean there is a 0.02% probability that the null hypothesis is true.
| Analysis | Estimate | 95% CI | P-value |
|---|---|---|---|
| Global cohort | HR 0.50 | 0.35–0.72 | 0.0002 |
| China cohort | HR 0.55 | 0.23–1.35 | Not reported |
Time to First Subsequent Therapy or Death (TFST)
Global cohort
95% CI: 0.22–0.40 · P < 0.0001
The global TFST hazard ratio of 0.30 indicates an estimated event hazard approximately 70% lower with olaparib under the fitted Cox model. Here the event is time to first subsequent therapy or death rather than progression itself.
The 95% CI of 0.22–0.40 is relatively narrow, indicating more statistical precision than the China-cohort estimate. The P-value of <0.0001 indicates strong evidence against the specified null under the posted superiority analysis.
| Analysis | Estimate | 95% CI | P-value |
|---|---|---|---|
| Global cohort | HR 0.30 | 0.22–0.40 | <0.0001 |
| China cohort | HR 0.58 | 0.29–1.23 | Not reported |
Time to Second Subsequent Therapy or Death (TSST)
Global cohort
95% CI: 0.32–0.63 · P < 0.0001
The global TSST hazard ratio of 0.45 corresponds to an estimated event hazard approximately 55% lower with olaparib under the fitted Cox model. This endpoint extends the treatment-pathway perspective beyond the first subsequent therapy.
The 95% CI of 0.32–0.63 quantifies uncertainty around the relative hazard estimate. The P-value of <0.0001 indicates statistical evidence under the specified superiority analysis; it should not be confused with the magnitude or clinical importance of the effect.
| Analysis | Estimate | 95% CI | P-value |
|---|---|---|---|
| Global cohort | HR 0.45 | 0.32–0.63 | <0.0001 |
| China cohort | HR 0.55 | 0.25–1.26 | Not reported |
Time to Study Treatment Discontinuation or Death (TDT)
Global cohort
95% CI: 0.51–0.79 · P < 0.0001
The global TDT hazard ratio of 0.63 indicates an estimated event hazard approximately 37% lower with olaparib under the Cox model. This endpoint is different from PFS: it measures time from randomization to study treatment discontinuation or death.
The 95% CI of 0.51–0.79 provides the registry-reported uncertainty interval. The P-value of <0.0001 addresses statistical evidence for the superiority comparison and does not measure the practical importance of the effect.
| Analysis | Estimate | 95% CI | P-value |
|---|---|---|---|
| Global cohort | HR 0.63 | 0.51–0.79 | <0.0001 |
| China cohort | HR 0.81 | 0.47–1.42 | Not reported |
9. PFS in Patients With a Deleterious or Suspected Deleterious BRCA Variant
The registry separately reports PFS in the Global Full Analysis Set among patients confirmed as Myriad gBRCAm. This is a defined analysis population rather than a new randomized comparison.
BRCA-specific PFS
95% CI: 0.22–0.40 · P < 0.0001
The estimated hazard ratio of 0.30 indicates an estimated instantaneous PFS event rate approximately 70% lower with olaparib under the fitted Cox model in the registry-reported Myriad gBRCAm analysis population.
The confidence interval of 0.22–0.40 indicates relatively tight statistical uncertainty around the estimate. The analysis remains a time-to-event comparison and therefore inherits the interpretive considerations of censoring and the Cox model.
Because this analysis is defined by confirmation of a BRCA-related population within the randomized trial, its interpretation should remain tied to the population specified in the registry rather than generalized to patients without the stated BRCA confirmation.
10. Health-Related Quality of Life
The registry includes a secondary continuous endpoint measuring change from baseline in health-related quality of life using the Trial Outcome Index (TOI) of the Functional Assessment of Cancer Therapy - Ovarian (FACT-O).
| Feature | Registry information |
|---|---|
| Endpoint | Change From Baseline in Health-Related Quality of Life (HRQoL) as Assessed by the Trial Outcome Index (TOI) of the Functional Assessment of Cancer Therapy - Ovarian (FACT-O) |
| Time frame | Questionnaires at baseline, Day 29 and then every 12 weeks for 156 weeks, then every 24 weeks or until the data cut off for the PFS analysis, change in TOI over 24 months reported |
| Analysis | Mixed Models Analysis |
| Effect measure | Mean Difference (Final Values) |
| Estimate | -3.00 |
| 95% CI | -4.779 to -1.216 |
| P-value | 0.0010 |
Final-value mean difference
95% CI: -4.779 to -1.216 · P = 0.0010
The reported mean difference of -3.00 represents the model-based difference in final TOI values between the randomized treatment groups according to the direction used in the registry analysis. Because the ClinicalTrials.gov record does not provide the underlying scale orientation or a minimally important difference, the numerical difference should not be translated into a statement of clinical importance solely from the P-value.
The 95% CI of -4.779 to -1.216 describes uncertainty around the model-based mean difference. The P-value of 0.0010 provides evidence against the specified null for this comparison, but it does not indicate whether the magnitude is clinically meaningful.
The model included fixed effects for treatment, visit and baseline TOI, treatment-by-visit and baseline-TOI-by-visit interactions, together with a random patient effect. This is appropriate conceptually for repeated measurements because observations from the same patient are correlated rather than independent.
11. Statistical Methodology
Cox proportional-hazards model
The dominant statistical method in SOLO1 is the Cox proportional-hazards model. The registry reports Cox regression for the primary PFS analysis and for the posted secondary time-to-event analyses.
The hazard ratio for a treatment indicator is represented by exp(β). A hazard ratio below 1 indicates a lower modeled event hazard in the treatment group relative to the comparator.
For the global PFS analysis, the registry notes that the model includes a factor for response to previous platinum chemotherapy. For the China analyses, the model includes the stratification variable of response to first-line platinum chemotherapy as a covariate.
Hazard ratio
The hazard ratio is a relative time-to-event measure. An HR of 0.30 does not mean that 30% of patients had an event, nor does it mean that the median survival is 70% longer. It describes the relative event hazard under the fitted model.
The corresponding relative reduction in estimated hazard is approximately 70%. This is a statement about the model-based relative hazard, not an absolute probability.
Intention-to-treat analysis
The registry identifies intention-to-treat analysis as an associated concept for the primary global PFS analysis and multiple secondary efficacy analyses. The essential principle is that randomized patients remain associated with their randomized treatment group for the efficacy comparison.
This preserves the treatment contrast created by randomization. It also means that the treatment-effect estimate can represent the effect of assignment to the strategy being compared rather than merely the effect among patients who remain on treatment.
Covariate adjustment and stratified analysis
The registry explicitly identifies covariate adjustment and stratified analysis for the China cohort analyses. The China PFS model includes the stratification variable of response to first-line platinum chemotherapy as a covariate.
Adjustment does not turn a randomized comparison into an observational study. Instead, it incorporates prespecified information into the statistical model used to estimate the treatment effect. The key interpretive requirement is to distinguish the covariates used in the model from variables that define a separate subgroup analysis.
Mixed-effects model for HRQoL
The HRQoL endpoint was analyzed using a mixed-effects model. The registry specifies fixed effects for treatment, visit and baseline TOI, treatment-by-visit and baseline-TOI-by-visit interactions, together with a random patient effect.
The random patient effect accounts for the fact that repeated TOI measurements from one patient are correlated. Treating every observation as independent would generally fail to represent that within-patient correlation.
12. Statistical Methods Explained
Why was a Cox model used for PFS?
PFS is a time-to-event endpoint. Some patients may experience progression or death during follow-up, while others may remain event-free when observation ends. Cox regression is designed for this structure and provides a hazard ratio comparing the event rates between treatment groups while accommodating right censoring.
What does an HR of 0.30 mean?
An HR of 0.30 means that the estimated instantaneous event hazard under the fitted model is 30% of the comparator hazard. It can therefore be described as an approximately 70% lower estimated hazard. It does not mean that 70% of patients are protected from progression, and it does not directly give an absolute risk difference.
Why does the confidence interval matter?
A point estimate is only one estimate of the underlying treatment effect. The 95% confidence interval provides a range reflecting statistical uncertainty under the specified model and sampling framework. The global PFS interval of 0.23–0.41 is considerably narrower than the China PFS interval of 0.23–0.97, illustrating how cohort-specific estimates can be less precise.
Why does the P-value not measure effect size?
A P-value measures the statistical evidence against a specified null hypothesis under the assumptions of the analysis. It does not tell us how large the treatment effect is. For example, the global PFS estimate of 0.30 and its 0.23–0.41 confidence interval describe effect size and precision; the P-value of <0.0001 describes evidence against the null.
Why is the HRQoL analysis different?
HRQoL is measured repeatedly on a continuous scale rather than as a time-to-event outcome. The registry therefore uses a mixed-effects model rather than a Cox model. The random patient effect accounts for correlation among repeated observations from the same participant.
Why should the China cohort estimates be interpreted cautiously?
The registry-reported China analyses have wider confidence intervals than the global analyses. For example, the China PFS estimate is 0.46 with a 95% CI of 0.23–0.97, compared with 0.30 and 0.23–0.41 globally. Wider intervals indicate greater uncertainty; they should not be treated as evidence that the treatment effect is necessarily different between cohorts.
What does a time-to-second-progression endpoint add?
Time from randomization to second progression extends the outcome beyond the first progression event. It can therefore provide information about disease control through a subsequent progression event. It remains a time-to-event endpoint and is analyzed with the Cox model in the ClinicalTrials.gov record.
13. Interpreting the Pattern of Results
The posted results show a consistent pattern across several global time-to-event endpoints. The estimated hazard ratios are below 1 for PFS, earliest progression/CA-125/death, second progression, first subsequent therapy or death, second subsequent therapy or death, and treatment discontinuation or death.
| Endpoint | Global estimate | 95% CI | P-value |
|---|---|---|---|
| Primary PFS | HR 0.30 | 0.23–0.41 | <0.0001 |
| Overall survival | HR 0.95 | 0.60–1.53 | 0.8903 |
| Earliest progression by RECIST, CA-125, or death | HR 0.30 | 0.23–0.40 | <0.0001 |
| Second progression | HR 0.50 | 0.35–0.72 | 0.0002 |
| First subsequent therapy or death | HR 0.30 | 0.22–0.40 | <0.0001 |
| Second subsequent therapy or death | HR 0.45 | 0.32–0.63 | <0.0001 |
| Treatment discontinuation or death | HR 0.63 | 0.51–0.79 | <0.0001 |
The distinction between these endpoints is important. PFS measures a disease-related time-to-event outcome, whereas overall survival measures death from any cause. TFST, TSST and TDT capture different points in the treatment pathway. Similar-looking hazard ratios therefore do not imply that the endpoints are interchangeable.
14. Multiplicity, Interim Analysis, and Other Design Features
The ClinicalTrials.gov record identifies a superiority hypothesis for the posted analyses, but they do not provide details of a multiplicity adjustment strategy, interim-analysis boundary, alpha-spending plan, non-inferiority margin, or Bayesian analysis.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Superiority | Yes. The posted primary and secondary analyses are identified as superiority hypotheses. |
| Multiplicity | No specific multiplicity procedure is provided in the ClinicalTrials.gov record. |
| Interim analysis | No specific interim-analysis method or boundary is provided in the ClinicalTrials.gov record. |
| Non-inferiority margin | Not applicable to the registry-reported superiority analyses; no non-inferiority margin is reported. |
| Bayesian methods | No Bayesian method is reported in the ClinicalTrials.gov record. |
| Crossover | No crossover procedure is reported in the ClinicalTrials.gov record. |
| Factorial design | No factorial structure is reported; the design model is parallel. |
| Missing-data/imputation method | No specific imputation procedure is reported in the ClinicalTrials.gov record. |
These omissions matter because statistical interpretation should not assume methods that are not documented in the available record. The posted analyses can be explained in terms of their reported Cox and mixed-effects methods without inventing an unreported multiplicity or interim-monitoring strategy.
15. Serious Adverse Events
The ClinicalTrials.gov record reports serious adverse events by arm for both the global and China cohorts.
| Cohort | Treatment | Affected / at risk | Proportion |
|---|---|---|---|
| Global | Olaparib 300 mg tablets | 54 / 260 | 54/260 |
| Global | Placebo tablets | 16 / 130 | 16/130 |
| China | Olaparib 300 mg tablets | 13 / 44 | 13/44 |
| China | Placebo tablets | 3 / 20 | 3/20 |
The registry data provide affected and at-risk counts rather than a formal between-group statistical analysis for these serious adverse events. Accordingly, the counts should be presented descriptively rather than converted into an inferred hypothesis test or relative-risk estimate.
The denominators also differ between the global and China cohorts. These values therefore should not be pooled or treated as though they represented a single common analysis population.
16. Limitations
- Data cutoff: The registry states that the data presented based on the March data cutoff of 17 May 2018 are not the final analyses for OS, TFST, TSST, and TDT. Further analyses were planned when the prespecified number of OS events were achieved.
- Different endpoint definitions: PFS, earliest progression by RECIST/CA-125/death, second progression, TFST, TSST and TDT are distinct endpoints. Their hazard ratios cannot be treated as interchangeable measures of one outcome.
- Hazard-ratio interpretation: Cox-model estimates rely on the model framework, including the proportional-hazards interpretation. The ClinicalTrials.gov record does not provide a separate diagnostic assessment of that assumption.
- Censoring: Time-to-event analyses depend on the handling of censored observations. The ClinicalTrials.gov record provides the analysis method but not patient-level censoring information.
- China-cohort precision: China-specific confidence intervals are wider for the posted efficacy analyses than the corresponding global intervals, reflecting greater uncertainty around cohort-specific estimates.
- HRQoL population: The mixed-effects HRQoL analysis requires baseline and post-baseline TOI scores, so its analysis population is defined differently from the general global FAS.
- Serious adverse events: The ClinicalTrials.gov record is descriptive affected/at-risk counts. No formal statistical comparison is posted on ClinicalTrials.gov for these events.
- Unreported design details: The ClinicalTrials.gov record does not specify multiplicity control, interim-analysis boundaries, imputation procedures, Bayesian methods, crossover, or a non-inferiority framework.
- Results not equivalent to final analyses: The registry specifically cautions that OS, TFST, TSST and TDT analyses from the registry-reported cutoff were not final analyses.
17. Why This Trial Matters Statistically
SOLO1 is a useful statistical teaching case because it combines randomized treatment assignment with several related time-to-event endpoints and a longitudinal patient-reported outcome.
| Concept | How it appears in SOLO1 |
|---|---|
| Randomization | Randomized, parallel-group phase 3 design. |
| Blinding | Quadruple masking. |
| Intention-to-treat analysis | Identified as an analysis concept for the primary and multiple secondary efficacy analyses. |
| Time-to-event endpoints | PFS, overall survival, earliest progression/CA-125/death, second progression, TFST, TSST and TDT. |
| Cox model | Primary PFS and multiple secondary efficacy analyses. |
| Hazard ratio | Primary and secondary time-to-event effect measure. |
| Confidence intervals | 95% two-sided intervals accompany the posted estimates. |
| Covariate adjustment | Response to previous or first-line platinum chemotherapy is included in specified models. |
| Stratified analysis | Explicitly identified for the China cohort analyses. |
| Mixed-effects model | Used for repeated HRQoL TOI measurements. |
| Continuous outcome analysis | HRQoL reported as a mean difference in final values. |
| Safety denominators | Serious adverse events reported as affected patients over patients at risk by cohort and arm. |
The statistical lesson is that a trial should be read endpoint by endpoint. The same randomized comparison can generate very different estimates depending on whether the event is progression, death, second progression, subsequent therapy, or treatment discontinuation.
18. Clinical Biostats Interpretation of the Primary Result
The global primary PFS analysis reported an HR of 0.30 with a two-sided 95% CI of 0.23–0.41 and P < 0.0001. Under the reported Cox model, this corresponds to an estimated PFS event hazard approximately 70% lower with olaparib than placebo.
The HR does not provide an absolute probability of remaining progression-free, a median PFS value, the proportion of patients cured, or the treatment effect for every individual patient. None of those quantities should be inferred from the hazard ratio alone.
The 95% CI of 0.23–0.41 gives the statistical uncertainty around the global PFS estimate. The relatively narrow interval indicates greater precision than the China-specific interval of 0.23–0.97.
The P-value of <0.0001 indicates strong statistical evidence against the null hypothesis specified for the superiority analysis. It is not a probability that the null hypothesis is true, and it does not measure the clinical magnitude of the treatment effect.
The global analysis uses the Full Analysis Set, including all patients randomized as part of global recruitment and including 5 patients randomized in China. The analysis therefore should not be described as if it were limited to patients who completed treatment or who remained progression-free through a particular visit.
19. Global Versus China Cohort: A Statistical Comparison
| Endpoint | Global cohort | China cohort |
|---|---|---|
| PFS | HR 0.30 (95% CI 0.23–0.41), P < 0.0001 | HR 0.46 (95% CI 0.23–0.97) |
| Overall survival | HR 0.95 (95% CI 0.60–1.53), P = 0.8903 | HR 0.77 (95% CI 0.29–2.28) |
| Earliest progression by RECIST/CA-125/death | HR 0.30 (95% CI 0.23–0.40), P < 0.0001 | HR 0.47 (95% CI 0.23–0.99) |
| Second progression | HR 0.50 (95% CI 0.35–0.72), P = 0.0002 | HR 0.55 (95% CI 0.23–1.35) |
| TFST | HR 0.30 (95% CI 0.22–0.40), P < 0.0001 | HR 0.58 (95% CI 0.29–1.23) |
| TSST | HR 0.45 (95% CI 0.32–0.63), P < 0.0001 | HR 0.55 (95% CI 0.25–1.26) |
| TDT | HR 0.63 (95% CI 0.51–0.79), P < 0.0001 | HR 0.81 (95% CI 0.47–1.42) |
The cohort-specific estimates should be compared descriptively rather than by looking only at whether an individual confidence interval crosses 1. A formal claim that treatment effects differ between cohorts would require an appropriate interaction or heterogeneity analysis; the ClinicalTrials.gov record does not report such a test.
20. Related Tutorials
Learn more about the methods used in this trial:
21. Related Calculators
22. Sources
- ClinicalTrials.gov: SOLO1, NCT01844986.
- Linked publication: PubMed record for PMID 39695768.
- Linked publication: PubMed record for PMID 36082969.
- Linked publication: PubMed record for PMID 35170751.
- Linked publication: PubMed record for PMID 34715071.
- Linked publication: PubMed record for PMID 33862001.
Continue learning from the statistical methods
Explore the underlying survival-analysis, clinical-trial, and longitudinal-data methods used to interpret randomized clinical trials.
23. Record Summary
SOLO1 provides a useful example of how a randomized phase 3 trial can combine a primary time-to-event endpoint with multiple secondary time-to-event outcomes and a longitudinal quality-of-life endpoint. The primary global PFS analysis reported a hazard ratio of 0.30 with a 95% CI of 0.23–0.41 and P < 0.0001. The registry also reports hazard ratios below 1 for several secondary global endpoints, while the registry-reported global OS analysis reports an HR of 0.95 with a 95% CI of 0.60–1.53 and P = 0.8903.
The most important statistical distinction is between the different endpoint definitions. PFS, overall survival, second progression, subsequent therapy, and treatment discontinuation describe different events and different points in the patient's treatment pathway. Their hazard ratios should therefore be interpreted individually rather than combined into one overall treatment-effect statistic.