← Clinical Trials
Late Infantile Metachromatic Leukodystrophy Phase 1, Phase 2 Continuous Endpoint NCT00633139

Long-term Metazym Treatment of Patients With Late Infantile Metachromatic Leukodystrophy (MLD): Complete Statistical Analysis

This analysis examines the registry record for a single-center, open-label, non-randomized study of Metazym in late infantile metachromatic leukodystrophy. It focuses on the planned motor-function endpoint, the study design, and how to interpret the reported safety counts in the absence of numerical primary-endpoint results.

Trial start: 2007-01-22  ·  Enrollment: 13  ·  Sponsor: Shire
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.

1. Trial at a Glance

This page covers the registry results for Long-term Metazym Treatment of Patients With Late Infantile Metachromatic Leukodystrophy (MLD) (NCT00633139), a non-randomized phase 1, phase 2 study. The primary outcome is Relative Changes (%) in Gross Motor Function Measurement (GMFM).

13
Enrollment
Phase 1, Phase 2
n/a
Arms
In the primary outcome
n/a
Effect
No re-analysis shown
n/a
95% CI
See limitations
FeatureRegistry value
ClinicalTrials.gov identifierNCT00633139
PhasePhase 1, Phase 2
ConditionLate Infantile Metachromatic Leukodystrophy
AllocationNon-Randomized
Design modelParallel
MaskingNone (open label)
Enrollment13
Arms in the primary outcomen/a
Lead sponsorShire
Results postedYes

2. Clinical Question

The study asked how motor function changed during continuing Metazym treatment in patients with late infantile metachromatic leukodystrophy (MLD). Its primary endpoint was relative change in Gross Motor Function Measurement (GMFM) from baseline to Week 52. The registry describes three dose cohorts, but no untreated or active-control group, so the study does not directly compare treatment with a control condition.

Population

Patients with late infantile MLD. The study enrolled 13 participants. The stated intent-to-treat population included all participants in the study.

Intervention

Recombinant human arylsulfatase A (rhASA), also called Metazym. Cohorts received 50 U/kg, 100 U/kg, or 200 U/kg.

Comparator

There was no randomized or untreated comparator. The three dose cohorts were not a like-for-like control comparison.

Primary question

What was the relative change in total GMFM-88 score from baseline to Week 52 during treatment?

The registry states that participants had previously received 26 weeks of treatment in a phase I trial. In this study, participants were to receive 13 additional infusions, with one infusion every other week. The primary measure is described as a change over time, not as a randomized estimate of the difference between Metazym and another treatment or no treatment.

The posted record does not report a numerical primary-endpoint estimate or its confidence interval. As a result, the clinical question can be described from the endpoint definition and study plan, but the size and precision of the observed motor-function change are not reported in the registry record.

3. Trial Design

This was a single-center, open-label, non-randomized, uncontrolled, multiple-dose, dose-escalation study. The registry classifies it as phase 1 and phase 2, with a parallel model and no masking. It reports an enrollment of 13 and a completed status. The study started on 2007-01-22, and its primary completion date was 2008-11-25.

Participants were assigned to one of three dose cohorts: 50 U/kg, 100 U/kg, or 200 U/kg of recombinant human arylsulfatase A. The record does not describe random allocation. Cohort membership therefore should not be interpreted as creating randomized, comparable groups. The study is explicitly described as uncontrolled, and the registry cautions that the arms are not a like-for-like control comparison.

Participants had previously been treated for 26 weeks in a phase I trial. The continuing-treatment study planned 13 infusions, one every other week; the record says participants would then have had 27 total infusions of Metazym. After a total of 52 weeks of treatment, participants were to continue treatment through a compassionate-use protocol. Safety was to be monitored at every visit.

Because the study was open-label, participants and study personnel were not masked to treatment, as reflected in the registry's masking classification. The record does not report further details about how participants entered each cohort, whether the doses were escalated according to specified rules, or how allocation decisions were made. It also does not report the protocol's sample-size justification, interim analyses, or a detailed statistical analysis plan.

This design can describe outcomes and safety observations during treatment in the enrolled participants. Without a control group and randomized allocation, changes over time cannot be separated confidently from other influences, including the course of the condition, participant differences, or other aspects of care. That limitation is central to interpreting both the motor-function endpoint and differences among dose cohorts.

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.

Before assignment: All participants that completed study recombinant human arylsulfatase A (rhASA-01) (NCT00418561) except 1 participant who did not complete (at week 18) the rhASA-01 (NCT00418561) participated in HGT-MLD-048/rhASA-03 (NCT00633139).

Overall Study

StageCohort 1Cohort 2Cohort 3
STARTED454
COMPLETED443
NOT COMPLETED011

Reasons for not completing this period:

ReasonCohort 1Cohort 2Cohort 3
Non compliance001
Withdrawal by Subject010

Analysis population for the primary outcome, as defined in the registry: Intent to Treat (ITT) population included all the participants in the study.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

CharacteristicUnitCohort 1Cohort 2Cohort 3Total
Participants analysed45413
Age, Categorical: <=18 yearsParticipants45413
Age, Categorical: Between 18 and 65 yearsParticipants0000
Age, Categorical: >=65 yearsParticipants0000
Age, Continuousmonths36.25 (SD 9.32)41.80 (SD 10.13)30.75 (SD 7.27)36.69 (SD 9.59)
Sex: Female, Male: FemaleParticipants2338
Sex: Female, Male: MaleParticipants2215
Region of Enrollment: DenmarkParticipants45413

6. Primary Endpoint

The registry lists the primary outcome as Relative Changes (%) in Gross Motor Function Measurement (GMFM), measured over Baseline, 52 Weeks.

ElementRegistry entry
Result typeMean
Measure of spread95% Confidence Interval
UnitRelative % change in total GMFM-88 SOT
Endpoint classificationContinuous Endpoint

The primary endpoint was relative change in Gross Motor Function Measurement (GMFM), measured with the GMFM-88, from baseline to the end of the study at Week 52. The registry describes the outcome as relative percentage change in the total GMFM-88 score over time. Each GMFM-88 item is scored from 0, minimal, to 3, maximum; item scores are summed to form a total score. The total scale runs from 0, minimal, to 264, maximum.

The registry states that the GMFM score decreases over time when the disease worsens. It defines relative change as percentage change from baseline divided by the age difference in months between the first and last visit. It also notes that the mean reported for score over time represents an adjusted mean. The outcome is classified as a continuous measure, with the mean and a 95% confidence interval specified as the summary format.

General statistical context: a continuous motor-function measure can represent differences in degree of function rather than simply classifying participants into categories. Repeated measurement at baseline and follow-up makes it possible to describe change over time. The adjustment described in the registry is part of the endpoint's definition, but the specific adjustment model and covariates are not reported in the registry record.

The endpoint may be useful for describing the observed course of motor scores in this group, but the registry record does not report the numerical estimate, confidence interval, or participant-level data for this outcome. Its magnitude and precision therefore cannot be assessed from the outcome details in the registry record.

7. Why No Re-analysis Is Shown

Reported as posted, not recalculated. Not a randomized comparison (allocation: non-randomized); arms are not a like-for-like control comparison. A result is recalculated only when the posted data support a clean two-arm comparison, so no result is shown beyond what the sponsor posted.

8. Statistical Interpretation

Clinical Biostats interpretation

The registry identifies relative change in total GMFM-88 score as the primary outcome and specifies a mean with a 95% confidence interval. It does not report a numerical mean or confidence interval for that outcome. Consequently, no conclusion about the observed size, direction, or precision of the primary-endpoint change can be drawn from the endpoint information in the registry record. The record's general note that GMFM scores decrease over time as disease worsens describes the scale's interpretation; it is not a numerical study result.

General statistical context: a confidence interval describes the uncertainty around an estimated quantity under the assumptions of the analysis. A narrower interval indicates greater statistical precision, while a wider interval indicates less precision. An interval for a within-group change would describe uncertainty in that change; it would not by itself establish what would have happened without treatment. No interval limits for the primary outcome are reported in the registry record.

The design further limits interpretation. The study was non-randomized and uncontrolled. Thus, even if a numerical change were available, it would describe the enrolled cohort over time rather than isolate a treatment effect against a concurrent comparison group. The three dose cohorts do not supply a randomized control comparison, and their results should not be treated as though differences between them were caused only by dose. The record does not report the cohort-level primary-endpoint estimates needed to examine those patterns in any case.

The registry describes the participants as having received 26 weeks of treatment before entering this continuing-treatment study. This context matters because the study did not begin with treatment-naive participants, according to the record. It also planned additional infusions and subsequent compassionate-use treatment after 52 weeks. The available record does not state how observations after the planned endpoint were handled in the primary analysis.

The posted safety counts describe participants with at least one serious adverse event and participants with other adverse events within each cohort. They are counts, not comparisons against an untreated group. The registry also lists selected serious-event terms with percentages, but does not report the denominators for those percentages in the information available. These observations describe events recorded during the study; they do not, on their own, establish that the study treatment caused or did not cause an event.

The enrollment of 13 is small in absolute terms, and the three cohorts contain 4, 5, and 4 participants at risk for the reported safety categories. General statistical context: estimates from small groups tend to be sensitive to the outcome of individual participants and can be imprecise. The record does not provide uncertainty intervals for the adverse-event counts. Together with the lack of randomization and a control group, this means that descriptive findings should not be read as reliable estimates of comparative benefit or risk.

The registry lists a secondary outcome, change in cerebrospinal fluid (CSF) sulfatide, and reports three outcome measures and three analyses overall. Numerical findings for the primary endpoint and the detailed analysis methods are not reported in the registry record. The available information therefore supports a description of the study design, outcome definition, and recorded safety counts, but not a quantitative assessment of treatment-related change.

9. Statistical Methods Explained

What does a mean describe?

The primary outcome is continuous and is summarized using a mean. In general, a mean represents the average of the analyzed observations on the outcome scale. It can provide a concise description of a group's results, but it does not show the full distribution of participant outcomes. In a small cohort, individual observations can have a substantial influence on the mean. The registry identifies the mean for the primary measure as adjusted, but the model, adjustment variables, and exact estimation procedure are not reported in the registry record.

The outcome is described as relative change in total GMFM-88 score. The registry's definition combines percentage change from baseline with the age difference in months between first and last visit. This is the stated construction of the endpoint; it should not be replaced with an alternative calculation. The registry also gives the direction of the scale: a decreasing GMFM score over time indicates worsening over time. The interpretation of a relative-change estimate would therefore depend on its sign and the precise endpoint construction, neither of which can be evaluated numerically without the reported estimate.

What does a 95% confidence interval convey?

General statistical context: a confidence interval gives a range of values compatible with the observed data and the statistical model at a stated confidence level. It communicates uncertainty around an estimate rather than a guarantee that the true value lies in a particular interval. Its width depends on factors such as the variability of observations and the amount of information available. A confidence interval can be useful alongside a mean because it shows the estimate's precision more directly than a point estimate alone.

The registry specifies a 95% confidence interval for the primary outcome but does not report its bounds or the mean estimate. Therefore, the precision of the endpoint estimate cannot be assessed. The confidence interval should also be distinguished from a comparison between groups: a within-group interval does not substitute for a controlled contrast between treatment and comparator.

Why does the design matter for statistical interpretation?

General statistical context: randomization is used to make treatment groups comparable, on average, with respect to measured and unmeasured factors. An uncontrolled, non-randomized study lacks that protection. Differences across dose cohorts may reflect who entered each cohort, their starting characteristics, or other factors, as well as any dose-related pattern. Since this study is described as uncontrolled and non-randomized, comparisons among its cohorts are descriptive rather than equivalent to randomized treatment comparisons.

Open-label status means treatment assignment was not masked according to the registry. General statistical context: knowledge of treatment can affect reporting, assessment, and decisions about follow-up, depending on the outcome and study procedures. The registry does not report masking of outcome assessors or other safeguards, so those details cannot be characterized. It also does not report the protocol's allocation procedures or a detailed statistical plan.

How should the dose cohorts be read?

The cohorts received 50 U/kg, 100 U/kg, and 200 U/kg. Those doses identify the groups described in the registry, but the existence of multiple doses does not establish a dose-response relationship. General statistical context: a dose-response assessment requires outcome estimates across dose levels and a method that accounts for uncertainty and relevant differences between participants. Numerical outcome estimates by cohort and the method for assessing a dose pattern are not reported in the registry record.

The arms are not a like-for-like control comparison. It would therefore be inappropriate to interpret a difference in cohort outcomes as a clean estimate of the effect of one dose versus another. The record does not state whether dose assignment followed a prespecified escalation rule or whether the cohorts differed systematically at baseline.

What do planning sample-size illustrations mean?

General statistical context: a sample-size calculation can illustrate how many participants might be needed under stated assumptions to detect a specified effect with a chosen error rate and power. Such an illustration is distinct from the study's enrollment and from evidence about the study's ability to detect its observed results. Any sample-size figures accompanying the record are planning illustrations computed from the observed effect, not the trial's own design. They use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For non-time-to-event outcomes, the stated convention is participants per arm. These planning assumptions do not turn this uncontrolled study into a two-arm comparison, and the illustrated sample sizes are not observed power.

The registry record does not report a prospective sample-size justification for this study. It is therefore not possible to say from the record whether enrollment of 13 was selected to meet a particular precision or power target. Planning illustrations based on an observed effect should be interpreted as hypothetical design calculations, not as evidence that a treatment effect has been established.

What analysis details are unavailable?

The registry names a mean and 95% confidence interval for the primary outcome and states that the adjusted mean is the mean represented for score over time. It does not report the adjustment model, the covariates, the exact analysis dataset for the primary endpoint, or how repeated measurements were incorporated. It also does not report a method for handling missing observations, a protocol for multiplicity, or interim analyses. These omissions prevent a full assessment of how the stated endpoint would be estimated and how uncertainty would be quantified.

General statistical context: analysis choices can affect estimates, confidence intervals, and the interpretation of multiple outcomes. Transparent reporting of the analysis population, missing-data methods, and prespecified comparisons helps readers distinguish planned analyses from exploratory ones. In this record, the intent-to-treat population is defined as including all participants, but further details about endpoint-specific inclusion and missing measurements are not reported.

10. Multiplicity, Missing Data and Analysis Populations

The registry identifies the GMFM relative-change measure as the primary outcome and lists change in CSF sulfatide as a secondary outcome. It also reports three outcome measures and three analyses. The record does not describe how the three analyses relate to the outcomes, which comparisons were prespecified, or whether any multiplicity adjustment was planned. A statistical plan addressing multiple outcomes or analyses is not reported in the registry record.

General statistical context: when several outcomes or comparisons are examined, the chance of finding an apparently notable result by chance can increase if each is assessed separately. Methods for addressing multiplicity may adjust the interpretation or thresholds for inference. Without the prespecified analysis details, it is not possible to determine whether or how multiplicity was handled. The numerical findings for the primary and secondary outcomes are also not reported in the registry record.

The primary population is described as the intent-to-treat population, which the registry defines as including all participants in the study. The record does not report how many participants contributed data to each endpoint at each visit, whether all enrolled participants had Week 52 measurements, or how missing observations were handled. It also does not report imputation methods, sensitivity analyses, or reasons for missing data. The general term intent-to-treat does not, by itself, reveal how incomplete outcome measurements were treated in a particular analysis.

These gaps make it difficult to assess whether the analyzed data represent all enrolled participants as intended and whether the results might be sensitive to missingness or multiple testing. No assumptions about the missing-data approach or multiplicity procedure can be inferred from the endpoint summary alone.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

ArmParticipants with serious adverse eventsParticipants at riskPercentage
Cohort 12450.0%
Cohort 22540.0%
Cohort 33475.0%
Participants with at least one serious adverse event
Serious adverse events by armCohort 150.0%Cohort 240.0%Cohort 375.0%% of participants with at least one serious adverse event

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

ArmParticipants affected of at risk
Cohort 14 of 4
Cohort 25 of 5
Cohort 34 of 4

Most frequent serious adverse event terms

Event termOrgan systemCohort 1Cohort 2Cohort 3
malnutritionMetabolism and nutrition disorders1 of 42 of 52 of 4
pneumoniaInfections and infestations1 of 40 of 53 of 4
pneumonia aspirationRespiratory, thoracic and mediastinal disorders0 of 41 of 51 of 4
bronchitis acuteInfections and infestations1 of 40 of 50 of 4
vomitingGastrointestinal disorders0 of 41 of 50 of 4
blindnessEye disorders0 of 40 of 51 of 4
post procedural vomitingInjury, poisoning and procedural complications0 of 40 of 51 of 4
peritonitisGastrointestinal disorders0 of 40 of 51 of 4

Most frequent non-serious adverse event terms

Event termOrgan systemCohort 1Cohort 2Cohort 3
HypotoniaNervous system disorders2 of 44 of 53 of 4
Muscle spasticityNervous system disorders4 of 44 of 51 of 4
PyrexiaGeneral disorders3 of 43 of 53 of 4
Infusion related reactionGeneral disorders2 of 42 of 53 of 4
GastroenteritisInfections and infestations3 of 42 of 51 of 4
Cognitive disorderNervous system disorders2 of 42 of 51 of 4
Muscle spasmsMusculoskeletal and connective tissue disorders1 of 43 of 51 of 4
VomitingGastrointestinal disorders0 of 42 of 52 of 4
Pharyngolaryngeal painRespiratory, thoracic and mediastinal disorders2 of 41 of 50 of 4
Speech disorder developmentalNervous system disorders1 of 42 of 50 of 4

Serious events with the highest arm-level rate: pneumonia (75.0%), malnutrition (50.0%), bronchitis acute (25.0%), pneumonia aspiration (25.0%), blindness (25.0%).

Do not overextend the safety comparison. Serious-adverse-event counts are safety observations, not efficacy outcomes, and should not be combined with the primary result into a single benefit-risk number.

The registry reports serious adverse events by cohort as participants at risk and participants affected. In Cohort 1, 4 participants were at risk and 2 were affected; in Cohort 2, 5 were at risk and 2 were affected; in Cohort 3, 4 were at risk and 3 were affected. For other adverse events, the corresponding figures were 4 at risk and 4 affected in Cohort 1, 5 at risk and 5 affected in Cohort 2, and 4 at risk and 4 affected in Cohort 3.

These are participant counts for the categories in the registry. They should not be read as event counts, because the record does not state how many events each affected participant experienced. Nor do the cohort counts establish a comparative safety difference: allocation was non-randomized, there was no control group, and the cohorts were small. The record says safety was monitored at every visit.

The registry lists pneumonia at 75.0%, malnutrition at 50.0%, and bronchitis acute, pneumonia aspiration, blindness, and post procedural vomiting at 25.0% each among the selected serious-event terms. The denominators used for these percentages are not reported in the registry record. Death counts are not reported in the registry record. These entries describe reported events; they do not establish that treatment caused them, and they do not provide a controlled estimate of risk.

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

The record establishes that a small, single-center cohort of patients with late infantile MLD was followed during continuing Metazym treatment, with three dose cohorts and a planned Week 52 motor-function endpoint. It also defines how GMFM-88 scores were constructed and states that decreasing scores over time indicate worsening. The reported safety summaries give counts of participants affected by serious and other adverse events within each cohort.

However, the record does not report numerical results for the primary GMFM outcome, including its mean estimate or 95% confidence interval. It does not report numerical results for the secondary CSF sulfatide outcome. Without those outcome estimates, the direction, magnitude, and precision of observed changes cannot be assessed from the registry record. The endpoint description is not itself evidence of an observed improvement or worsening in this study.

The design also does not support a controlled causal comparison. Participants were not randomized, there was no control group, and the dose cohorts are not described as like-for-like comparisons. Any within-cohort change over time would need to be interpreted as an observation during treatment, not as a direct estimate of what treatment caused compared with an alternative condition.

The adverse-event figures should likewise be understood descriptively. They report how many participants in each cohort were affected, but do not provide a control-group rate or establish attribution. The listed serious-event terms and percentages do not supply the missing denominator information needed to interpret those percentages fully.

General statistical context: a study can contribute useful descriptive information even when it cannot answer a controlled comparative question. The strength of an inference depends on the design, the completeness of outcome reporting, the precision of estimates, and the analysis methods. Here, the registry record provides design and endpoint definitions, but does not report the primary numerical result, detailed analysis methods, missing-data handling, or multiplicity approach. The result therefore cannot be characterized quantitatively from the information recorded.

13. Why This Trial Matters Statistically

This study is statistically notable as a continuing-treatment, dose-cohort study in a specific patient population, with a defined motor-function outcome and repeated treatment over a planned 52-week period. The registry supplies useful context about the GMFM-88 scale, including the range of item and total scores, the direction of worsening, and the stated adjustment for age difference in months. These details help clarify what the primary endpoint was intended to represent.

The record also illustrates why study design and outcome reporting must be read together. The study was open-label, non-randomized, and uncontrolled. Those features mean that observed changes cannot be separated from the natural course of the condition or other influences using a concurrent randomized comparison. The three dose groups do not automatically provide a valid substitute for a control group.

For biostatistical interpretation, a specified mean and confidence interval are not enough unless their numerical values and estimation methods are reported. In this record, the endpoint summary format is stated, but the primary estimate and its interval are not reported. The absence of those results prevents readers from assessing the estimate's size and precision. Likewise, the registry reports three outcome measures and three analyses but does not describe any multiplicity strategy or missing-data method.

The safety counts offer descriptive information on participants affected in each cohort, while also showing the limits of small, uncontrolled groups for comparative safety inference. Overall, the record is a useful example of how a clearly named endpoint and structured safety reporting can coexist with substantial uncertainty about the study's quantitative findings and causal interpretation.

14. Limitations

The principal trial-specific limitation is the absence of numerical primary-endpoint results in the registry record. Although the outcome, time points, summary measure, and confidence-interval format are described, the mean and interval are not reported. Numerical results for the listed secondary CSF sulfatide outcome are also not reported. This prevents assessment of the observed changes and their precision.

The study was single-center, open-label, non-randomized, and uncontrolled. The registry itself notes that its arms are not a like-for-like control comparison. The three dose cohorts therefore cannot be assumed to be comparable, and any differences between them cannot be attributed to dose alone based on the design described.

Enrollment was 13, with 4, 5, and 4 participants at risk for the reported safety categories in the three cohorts. Small cohort counts limit the detail that can be drawn from descriptive safety summaries. The record does not report confidence intervals for these counts or explain the denominators for the percentages attached to selected serious-event terms. Death counts are not reported in the registry record.

Participants had already received 26 weeks of treatment in a phase I trial before this study. This continuing-treatment context differs from initiating treatment in a group without prior exposure. The registry does not report endpoint-specific follow-up completeness, missing-data handling, the adjusted-mean model, or the covariates used. It also does not report the statistical plan for multiplicity, interim analyses, or the study's sample-size justification.

Finally, although treatment was planned for 52 weeks before continued access through a compassionate-use protocol, the record does not describe how later observations were handled in the primary analysis. These reporting gaps and design constraints limit both reproducibility and the strength of conclusions that can be drawn from the registry information.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

Continue through the Clinical Biostats statistical pathway

Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.

18. Record Summary

The registry record describes a completed, single-center, open-label, non-randomized and uncontrolled phase 1/phase 2 study of recombinant human arylsulfatase A (Metazym) in late infantile MLD. It reports enrollment of 13 and three dose cohorts: 50 U/kg, 100 U/kg, and 200 U/kg. Participants had prior treatment in a phase I trial and were planned to receive 13 further infusions, one every other week, with a primary follow-up point at Week 52.

The primary outcome was relative change in GMFM-88 from baseline to Week 52, summarized as a mean with a 95% confidence interval. The registry does not report the numerical primary-endpoint estimate or interval. It lists change in CSF sulfatide as a secondary outcome and provides descriptive adverse-event counts by cohort. Because there was no randomized control group and the cohorts were not like-for-like comparisons, the record supports descriptive interpretation rather than a controlled estimate of treatment effect.

Clinical Biostats