Metformin, Vitamin D, and Depression in Polycystic Ovary Syndrome (PCOS) Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Metformin, Vitamin D, and Depression in Polycystic Ovary Syndrome (PCOS) Trial (NCT03380091), a randomized phase 4 study. The primary outcome is Change From Baseline in Depression Score on the Beck Depression Inventory (BDI-II) at 12 Weeks.

2. Clinical Question

The BDI-II is a self-reported measure of depressive symptoms. Its scores range from 0-63, with higher totals indicating more severe symptoms. The outcome was defined as the week 12 total score minus the baseline score, so a negative change represents a lower score at week 12 than at baseline, while a positive change represents a higher score.

3. Trial Design

The MINDD Trial was a Phase 4, randomized, parallel-arm, open-label clinical trial sponsored by the University of California, San Francisco. Its stated purpose was to assess metformin versus vitamin D for improving symptoms of depressed mood in people with PCOS. The two listed interventions were metformin and cholecalciferol, a form of vitamin D. The registry record describes metformin as 1000 mg PO bid and cholecalciferol as 5,000 IU PO daily.

The record lists enrollment of 5 and gives a start date of 2017-12-12 and a primary completion date of 2019-09-01. The trial status is TERMINATED. The reason for termination is not reported in the registry record. The record identifies randomization but does not report the randomization sequence method, allocation concealment, or the allocation ratio. Masking is listed as NONE, meaning the study is described as open-label.

The primary outcome was change from baseline in BDI-II score at 12 weeks. The listed timeframe is baseline and 12 weeks. The record also identifies outcomes concerning insulin resistance score, State-Trait Anxiety Inventory score, and vitamin D 25-hydroxy levels at 12 weeks. Details of the measurement procedures beyond the outcome descriptions are not reported in the registry record.

Although enrollment was 5, only one participant in each arm contributed a calculable primary-outcome change. The registry record says the other 3 participants in each arm had no week 12 score available. The record does not explain why these participants did not have week 12 scores. It also does not report a protocol-level sample-size justification or a detailed analysis plan.

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: Of the 5 participants that were enrolled in the trial, all were randomized and assigned to treatment arms. 3 were assigned to the Vitamin D supplement arm and 2 were assigned to the Metformin arm.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: In both the Vitamin D and the Metformin arms, only 1 participant in each arm was analyzed for Change in Depression from Baseline as only 1 participant in each arm participated through to week 12. A Change could not be calculated for the other 3 participants as no week 12 scores were available.

5. Baseline Characteristics

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

Population: Although 3 participants were assigned to Vitamin D, one dropped out just after enrollment and arm assigment and another dropped out after completing baseline time point measures, thus only one participant in the Vitamin D arm had data available for outcome measures analysis. Although 2 participants were assigned to Metformin, one dropped out just after enrollment and arm assignment, thus only one participant in the Metformin arm had data available for outcome measures analysis.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Depression Score on the Beck Depression Inventory (BDI-II) at 12 Weeks, measured over Baseline & 12 weeks.

The primary endpoint is change in BDI-II score from baseline to 12 weeks. The BDI-II is described as a validated, self-reported instrument consisting of 21 questions, each scored 0-3. Total scores range from 0-63, and higher totals indicate more severe depressive symptoms. The reported change is calculated as the week 12 total score minus the baseline total score.

Change scores are used to describe how a participant’s measured outcome differs between two time points. A negative value means the later score is lower than the baseline score; a positive value means it is higher. This direction is important when interpreting the posted values, but a change score by itself does not establish why a score changed or whether the change was caused by treatment.

The BDI-II was selected as the primary measure of depressed mood in this trial. The registry record reports that only one participant in each arm had both the information needed to calculate a change and a week 12 score. The posted arm values are therefore based on very limited endpoint data. At this sample size, the outcome cannot characterize the range of participant experiences within either arm. The registry record does not report an established threshold for interpreting the clinical importance of the observed changes.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a score change measured on a numerical scale. The registry record describes the arm summaries as means with standard deviations. Statistical methods for continuous outcomes can describe the center of observed values and how dispersed those values are. In this trial, the mean change is reported as -8.0 in the Vitamin D (Cholecalciferol) arm and 7.0 in the Metformin arm. Each arm has one analyzed participant, and each reported standard deviation is 0.0.

What does a mean describe?

In general, a mean is an average of observed values. It can be a useful summary when multiple observations are available, but it does not show the full distribution of those observations. With one observation, the mean is simply that participant’s observed value. It cannot show whether other participants in the same arm would have had similar or very different changes. The posted arm means should therefore be read as individual observed changes, rather than as reliable estimates of typical changes in the trial population.

What does the standard deviation add?

In general, a standard deviation describes the spread of observations around their mean. It is not a measure of uncertainty in the same way as a confidence interval. A standard deviation of 0.0 alongside one analyzed participant does not mean the treatment has no variability or that the result is precisely known; there is no set of multiple observed changes within that arm from which to assess participant-to-participant spread. The record’s value should be interpreted in light of the one-person outcome sample.

What would a confidence interval show?

In general, a confidence interval gives a range of values compatible with the observed data under a specified statistical method. It communicates uncertainty around an estimate, but it does not guarantee that the true value lies within the interval or remove limitations such as missing outcomes. The registry record does not report confidence intervals for the primary outcome. With only one analyzed participant in each arm, an estimate of between-arm uncertainty would be particularly difficult to support from the posted summary information.

Why is a formal comparison absent?

The registry record lists 0 analyses and indicates that values are missing or not estimable for a two-arm comparison of this continuous endpoint. A between-arm analysis would ordinarily aim to estimate how the outcome differs between randomized groups and how uncertain that estimate is. The posted individual arm summaries do not provide a reported inferential comparison. No p-value or adjusted model result is reported in the registry record, and none should be inferred from the arm means alone.

Randomization is intended, in general, to create groups that can be compared without systematic assignment based on participant characteristics. However, randomization does not ensure balance in a very small realized sample, and it cannot compensate for the absence of endpoint data. The registry record says the allocation was randomized but does not report the sequence-generation or concealment methods. Because masking is listed as NONE, participants and study personnel were not masked according to the record. Open-label design may matter for self-reported outcomes in general, since knowledge of treatment can influence reporting or other aspects of study conduct; the registry record does not report whether such influences occurred in this trial.

How should planning illustrations be understood?

General sample-size illustrations use assumptions about the outcome, the target effect, the type I error rate, allocation, and the statistical approximation to describe how many participants might be needed for a specified design. The posted planning approach is described as a two-sided significance level of 0.05, equal allocation, and a normal approximation, with target effects expressed as fractions of the observed effect. These are planning illustrations based on the observed effect, not the trial’s own design justification. The protocol’s sample-size justification is not reported in the registry record. Such illustrations are not observed power and cannot recover information missing from the week 12 data.

In general, power is a design concept: it refers to the chance that a specified study and analysis would detect a chosen effect under stated assumptions. It is not a measure of whether a result is true. A small sample-size illustration can be sensitive to the observed effect used as its starting point; when the observed endpoint is based on one participant per arm, it is especially uncertain as a basis for planning. The record does not report that the trial had adequate power for its primary question.

Over-dispersion is a general concern in count outcomes when observed variability exceeds what a simple count model assumes. It is not the central issue for this primary endpoint, which is a continuous score change. The principal statistical limitations are the very small analyzed endpoint sample, the lack of a reported comparative analysis, and missing week 12 outcomes. The registry record does not report interim analyses, a detailed statistical analysis plan, or adjustments for baseline values and other covariates.

10. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary outcome and lists three secondary outcomes: insulin resistance score, State-Trait Anxiety Inventory score, and vitamin D 25-hydroxy levels, each at 12 weeks. It reports four outcome measures in total and three posted secondary outcomes. A plan for handling multiplicity across outcomes or multiple statistical tests is not reported in the registry record. In general, testing several outcomes can increase the chance of finding at least one apparently notable result by chance; methods such as prespecifying a testing hierarchy or adjusting significance thresholds can address that concern. No such method is reported, and no secondary-outcome analyses are listed.

Missing primary-outcome data are a major limitation. The record says only one participant in each arm was analyzed for change from baseline because only one participant in each arm participated through week 12. It says the other 3 participants in each arm had no week 12 score, so a change could not be calculated. The reasons for missing scores, any efforts to obtain them, and any method for addressing missing outcomes are not reported in the registry record.

In general, an analysis limited to participants with observed outcomes can be biased if the availability of those outcomes is related to the outcome itself or to treatment. The record does not provide information to assess whether that occurred. It also does not report a defined intention-to-treat analysis population, a per-protocol population, or a sensitivity analysis for missing data. The primary outcome results consequently describe the two participants with calculable changes and should not be assumed to represent all randomized participants.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry record lists serious adverse events for 0 participants in the Vitamin D (Cholecalciferol) arm, with 3 participants at risk, and 0 participants in the Metformin arm, with 2 participants at risk. It also lists 0 other adverse events in each arm and 0 deaths in each arm. These figures describe the recorded events and the corresponding at-risk counts; they do not establish that no adverse events could occur.

The at-risk counts are the safety denominators reported in the record. They are not the same as the number analyzed for the primary outcome: the primary-outcome description identifies one participant with a week 12 score in each arm. The registry record does not explain the relationship between the safety populations and the primary-outcome populations.

With small at-risk groups, zero recorded events provide little information about the frequency of uncommon events or about differences between treatments. In general, absence of a recorded event in a small group is not evidence that the risk is zero. The record lists no top serious adverse-event terms and does not provide further event descriptions in the posted results. These safety counts should be read as limited descriptive reporting, not as a basis for comparing safety profiles.

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

The posted primary outcome records opposite directions of change in the two analyzed participants: the Vitamin D (Cholecalciferol) participant had a change of -8.0, and the Metformin participant had a change of 7.0. These values document what was observed in those individuals. They do not establish the average effect of either intervention, nor do they establish a difference between interventions.

The most important reason is that only one participant per arm contributed a week 12 change. The other 3 participants in each arm had no week 12 score available. There is therefore very little information about outcomes among participants who did not contribute endpoint data, or about the range of possible outcomes among those who did. No inferential comparison, confidence interval, or primary-outcome analysis is reported in the registry record.

Randomized allocation and a parallel design describe the intended comparison, but they do not make sparse observed data conclusive. The trial is open-label, and the primary outcome is self-reported. In general, these design features can affect interpretation; the registry record does not show whether they influenced the reported scores. The reason the trial was terminated is also not reported, so termination should not be assigned a cause based on these results.

The findings do not show that metformin or vitamin D does, or does not, improve depressive symptoms. They do not establish comparative efficacy, clinical importance, or safety. The registry record supports a narrower statement: it reports two calculable BDI-II changes at 12 weeks, one in each arm, alongside substantial missing endpoint data and no reported comparative analysis. Any broader claim would go beyond the information recorded.

13. Why This Trial Matters Statistically

This trial illustrates why the number randomized and the number with an observed primary outcome need to be kept distinct. Total enrollment was 5, but only one participant in each arm had a calculable BDI-II change at week 12. A reader who focuses only on the trial’s randomized design may overlook how little endpoint information is available for the comparison.

It also illustrates the difference between describing observations and estimating a treatment contrast. Arm-level mean changes are reported, but with one observation in each arm they provide no meaningful picture of within-arm variability. The registry record does not report a comparative estimate or its uncertainty. The individual values can be described, but they cannot support a dependable conclusion about how the interventions compare.

The open-label design and self-reported primary outcome are relevant considerations for interpreting the measurements, while the missing week 12 scores raise questions about how the observed participants relate to the full enrolled group. The reasons for missingness and the planned handling of missing outcomes are not reported in the registry record. These are not merely technical details: they shape what population the posted results describe and how broadly the observations can be interpreted.

Finally, the record’s terminated status and its small observed outcome population demonstrate why registry results should be read with attention to both design and data completeness. The statistical lesson is not a treatment conclusion. It is that a randomized label, a numerical endpoint, and observed arm values do not by themselves provide a precise comparative answer.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary-outcome evidence is limited to one analyzed participant in each arm, despite total enrollment of 5. The record reports that the other 3 participants in each arm had no week 12 score available. Reasons for the missing scores and methods for addressing missing data are not reported in the registry record. Consequently, the endpoint results represent only participants with calculable change scores, and the extent to which they represent the randomized groups cannot be assessed.

The trial is open-label, and the primary endpoint is self-reported. These features can matter in general when interpreting symptom scores, although the record does not report evidence that they affected these results. The registry record does not describe allocation concealment, the randomization sequence, or a detailed statistical analysis plan. It also does not report a primary-outcome confidence interval, formal comparison, or analysis of the three listed secondary outcomes.

The trial is marked TERMINATED, but the reason is not reported in the registry record. The record does not report a protocol-based sample-size justification or explain whether interim analyses were planned or conducted. Safety information is also based on small at-risk groups, and the record does not provide further detail about the relationship between safety and primary-outcome populations. These limitations prevent strong conclusions about comparative outcomes, precision, generalizability, or the causes of missing data.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Metformin, Vitamin D, and Depression in Polycystic Ovary Syndrome (PCOS) Trial — NCT03380091.
• Registry record last updated 2021-02-09; status last verified 2021-01.
• Results first posted 2021-02-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The MINDD Trial was a Phase 4, randomized, parallel, open-label comparison of metformin and cholecalciferol in people with PCOS. Total enrollment was 5, and the registry record lists the status as TERMINATED. The primary endpoint was change from baseline in BDI-II score at 12 weeks.

Only one participant in each arm had a calculable primary-outcome change. The reported mean changes were -8.0 for Vitamin D (Cholecalciferol) and 7.0 for Metformin, with a standard deviation of 0.0 in each arm. The registry record reports no primary-outcome analyses or between-arm inference. It lists 0 serious adverse events in each arm and 0 deaths in each arm, with 3 and 2 participants at risk, respectively.

Missing week 12 scores, very small analyzed groups, and the absence of a reported comparative analysis sharply limit interpretation. The reason for termination and the protocol-level statistical plan are not reported in the registry record.