Characteristics of Empagliflozin Initiators: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Characteristics of Empagliflozin Initiators (NCT03050619), a registered phase not stated study. The primary outcome is Baseline Characteristics of Adults (as Measured by Demographics) With a Recorded Diagnosis of T2DM Starting Index Prescriptions in the UK.

2. Clinical Question

This is a descriptive question about who initiated different therapies, rather than a question about comparative treatment efficacy. The registry rationale notes that people starting a newer drug might differ from people starting drugs that had been on the market longer. It gives poorer glucose control as a possible example, not as a reported study finding. Accordingly, observed differences in patient characteristics would describe the groups at treatment initiation; by themselves, they would not establish that a drug caused those differences or that the groups are exchangeable for treatment-effect comparisons. The registry record does not report numerical demographic findings or adjusted comparisons.

3. Trial Design

NCT03050619 is a completed registry study titled “Characteristics of Empagliflozin Initiators.” Its official title describes patients initiating empagliflozin or other non-insulin glucose-lowering drugs in the United Kingdom. The record lists Boehringer Ingelheim as sponsor, an enrollment of 31908, a study start of 2016-09-10, and primary completion on 2016-10-14. Those dates describe the study record; the population eligibility dates are separately stated as 01 August 2014 to 01 September 2015.

The design compares groups defined by the drug they initiated: empagliflozin, other SGLT-2 inhibitors, and other non-insulin glucose-lowering drugs. The non-insulin group includes DPP-4 inhibitors, metformin, sulfonylureas, and GLP-1 agonists. The registry does not state that treatment was randomly allocated; allocation is not reported. It characterizes the arms as not a like-for-like control comparison, so the groups should be understood as treatment-initiation cohorts rather than randomized arms.

The primary outcome is baseline demographics, reported as a percentage of participants. The description also refers to baseline concomitant medications and comorbidities. Secondary outcomes include extent of off-label use and demographic, lifestyle, BMI, blood-pressure, HbA1c, laboratory-test, and eGFR characteristics among people starting prescriptions classified as off-label. The registry lists 21 outcome measures and 0 analyses. These entries do not provide numerical results or explain the statistical procedures used.

The study aims to characterize prescribing populations and possible off-label use, not to test whether one drug improves clinical outcomes relative to another. The record does not report a protocol-level account of participant selection beyond the stated age, diagnosis, UK setting, and prescription dates. It also does not report group-specific sizes, follow-up duration, or the data source used to identify prescriptions and diagnoses.

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 study participants had at least 12 months of continuous registration in the United Kingdom (UK) Clinical Practice Research Datalink (CPRD) prior to the study initiation which is a data source comprising primary care electronic medical record (EMR).

Overall Study

Analysis population for the primary outcome, as defined in the registry: All patients (older than 18 years), with a recorded T2DM, who initiated treatment with empagliflozin, other SGLT-2i, or other commonly used non-insulin GLDs (comprising DPP-4i, metformin, SU and GLP-1 agonists) between 01 August 2014 and 01 September 2015 were included.

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: All patients (older than 18 years), with a recorded type 2 diabetes mellitus (T2DM), who initiated treatment with empagliflozin, other SGLT-2i, or other commonly used non-insulin glucose lowering drugs (GLDs) (comprising DPP-4i, metformin, SU and GLP-1 agonists) between 01 August 2014 and 01 September 2015 were included.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Baseline Characteristics of Adults (as Measured by Demographics) With a Recorded Diagnosis of T2DM Starting Index Prescriptions in the UK, measured over Baseline.

The primary outcome is baseline demographic characteristics among adults with a recorded diagnosis of type 2 diabetes who initiated an index prescription. The registry specifies a percentage-of-participants unit and describes the characteristics as measured by demographics. The index prescriptions include empagliflozin, other SGLT-2 inhibitors, and commonly used non-insulin glucose-lowering drugs comprising DPP-4 inhibitors, metformin, sulfonylureas, and GLP-1 agonists.

In general, demographic descriptions can show how the people starting different treatments differ in composition. This is useful when a study aims to characterize real-world treatment initiation, because comparisons of patient profiles can place later questions about prescribing patterns in context. Such a descriptive outcome is not itself a measure of glucose control, a clinical event, or treatment benefit. Nor does a percentage alone establish why a patient or prescriber selected a particular drug.

The registry record does not specify the individual demographic variables included in the primary outcome, their definitions, their measurement windows beyond the baseline label, or the denominator used for each percentage. It does not report the numerical percentages, uncertainty intervals, or comparisons. Therefore, the outcome's stated purpose and unit can be described, but the observed demographic profile cannot be reconstructed from the registry information. The endpoint should be read as a planned descriptive characterization rather than evidence of a treatment effect.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies the outcome format as a percentage of participants and labels the primary measure as baseline demographics. The statistical concepts below are general explanations; the registry record does not report the specific analytic procedures used for these outcomes.

What does a percentage describe?

In general, a percentage expresses the number of people meeting a defined condition relative to a stated denominator. For baseline demographics, that might describe the composition of a treatment-initiation group for a particular characteristic. Interpretation depends on clear categories, a consistent time point, and an explicit denominator. If some demographic fields are missing, the denominator may vary across characteristics; whether it does so in this study is not reported in the registry record. A percentage should not be interpreted as a treatment effect simply because it is reported separately for different treatment groups.

Why does group definition matter?

This study defines groups by the treatment initiated, rather than by random assignment. In general, people receiving different treatments can differ before initiation in clinical history, treatment history, or other characteristics related to prescribing. Such differences can confound comparisons: an apparent association between treatment group and an outcome may partly reflect who was prescribed each treatment rather than an effect of the drug. This concern is particularly relevant to a descriptive study whose stated rationale anticipates that initiators of a newer medicine may differ from people initiating drugs that have been available longer. The registry does not report a method for controlling confounding or a causal analysis.

Randomization, when used, aims to make treatment groups comparable on average at baseline. The registry does not report random allocation for this study, and its comparison is explicitly not described as like-for-like. In general, statistical adjustment can account for measured differences only under assumptions about measurement and model specification; it cannot automatically remove differences that were not measured. No adjusted estimates or adjustment variables are reported in the registry record.

What would uncertainty measures add?

In general, a confidence interval describes the range of estimates compatible with a statistical model and the observed data, under that model's assumptions. It communicates precision as well as the point estimate. A narrow interval indicates greater precision than a wide interval, but neither precision nor a small p-value resolves bias from non-random group formation. The registry record does not report confidence intervals, standard errors, or p-values for the primary outcome. As a result, the precision of any demographic percentage or group difference cannot be assessed from the record.

For a binary outcome, the estimated percentage is based on the number meeting a specified definition and the applicable denominator. General binomial methods often rely on an appropriate model for independent observations and correctly specified outcome categories. If observations are clustered or otherwise dependent, standard errors may need to account for that dependence. The record does not describe whether any such structure was present or how uncertainty was handled. The outcome label alone is not enough to identify the estimation method.

How should planning illustrations be read?

The record describes planning illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says that the effect to detect is a stated fraction of the observed effect. In general, a sample-size calculation uses assumptions about the effect, variability, error rates, and design to estimate how many observations a future study may need. Such a calculation is a planning aid, not an observed result and not a measure of the probability that the completed study detected an effect. The record cautions that these illustrations are not the study's own design. No numerical illustration values are reported in the registry record.

General power calculations depend on the assumptions used, and changing the effect targeted changes the required sample size. Equal allocation is an assumption for the stated two-arm illustration; it should not be taken to mean the actual treatment groups were equal in size. The record does not report the observed effect used for the illustration, the allocation in the actual cohorts, or a sample-size justification for the study. The record also does not report whether the calculations were based on a particular endpoint or how the assumptions were selected.

What can be concluded from the recorded analysis information?

The registry lists 0 analyses. It does not specify a statistical model, covariate set, handling of missing data, multiplicity adjustment, subgroup strategy, or interim analysis. These omissions prevent a detailed reconstruction of the analysis. They are not a basis for assuming that a particular method was or was not used outside the information recorded. The registry's completed status and its total enrollment figure do not replace outcome estimates or analysis details.

For this study, the available methodological description supports a broad account: observational groups were defined by initiated drug; the primary measure was a baseline demographic percentage; and the aims also included off-label use and other baseline characteristics. General statistical principles explain why denominators, uncertainty estimates, confounding control, and missing-data procedures matter, but the registry record does not report those details or numerical findings. Any interpretation should keep those design and reporting limits in view.

10. Multiplicity, Missing Data and Analysis Populations

The study record lists 21 outcome measures, including the primary demographic outcome and secondary measures concerning off-label use and baseline characteristics. When a study examines multiple outcomes or many categories within an outcome, general statistical practice considers multiplicity: repeated tests can increase the chance of at least one apparently notable result arising by chance. Possible approaches include defining a limited set of primary hypotheses or adjusting inferential thresholds, but the appropriate choice depends on the study aims and analysis plan. The registry record does not report a multiplicity strategy, a hierarchy among all outcome measures, or whether formal hypothesis tests were used.

Missing-data handling is also not reported in the registry record. For descriptive percentages, missing observations can affect both the denominator and how representative an estimate is of the eligible group. A complete-case estimate, for example, describes participants with observed data for that characteristic; other approaches make different assumptions. This is a general methodological point, not a statement about the method used in this study. The registry does not give missingness counts, reasons for missing values, imputation methods, or outcome-specific denominators.

The record does not define analysis populations beyond the stated population of adults with recorded type 2 diabetes who initiated the relevant prescriptions within the eligibility dates. It does not describe exclusions after identification, rules for repeated prescriptions, or whether the analysis population differed by outcome. Although the enrollment is listed as 31908, group-specific analysis counts are not reported. These gaps limit assessment of how closely any analyzed population would represent the eligible population and how comparable denominators are across groups or outcomes.

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

This study addresses a useful descriptive question: what characteristics are recorded among people initiating empagliflozin and other glucose-lowering drugs in the UK? Its stated aims also include the extent of off-label use and the baseline profile of people whose index prescriptions were classified as off-label. Such characterization can help distinguish a description of prescribing populations from a randomized evaluation of treatment outcomes.

The registry information does not include numerical results for the demographic percentages or secondary outcomes. It therefore cannot establish what characteristics were more or less common in any initiation group, how frequently off-label use occurred, or how precise those estimates were. Nor does the absence of reported results establish that groups were similar. It indicates that the registry record does not report the relevant values.

Even if group characteristics were reported, a descriptive comparison would not by itself show that treatment caused those characteristics or that one treatment changed a clinical outcome. Drug initiation was not described as randomized, allocation is not reported, and the registry cautions that the arms are not a like-for-like control comparison. Differences could reflect patient selection, prior treatment, clinical circumstances, or other factors. A causal interpretation would require an appropriate design and analysis, with assumptions and methods made explicit.

The enrollment figure of 31908 describes the registry's enrollment entry, not the number in every treatment group or the denominator for each outcome. Group-specific counts, missing-data information, statistical estimates, and uncertainty intervals are not reported. Accordingly, the findings that can be described from this record are limited to the study's objectives, population definition, treatment-initiation groups, and outcome topics—not the magnitude, precision, or clinical implications of any observed difference.

12. Why This Trial Matters Statistically

Statistically, this study matters because treatment-initiation populations are not necessarily comparable by default. When an observational study contrasts people starting different medicines, the composition of those groups is part of the context needed to interpret later comparisons. Describing baseline characteristics can identify potential differences in who receives each treatment and can help readers avoid treating drug groups as if they had been formed by random allocation.

The registry rationale makes that issue explicit: initiators of empagliflozin may differ from people starting other glucose-lowering drugs that had been on the market longer. It gives poorer glucose control as a possible example, rather than a reported finding. The study's demographic and other baseline aims therefore fit a broader statistical need to understand selection into treatment groups before attempting to interpret outcome contrasts.

The study also illustrates the distinction between a study objective and its reported evidence. A registry record can define a population, outcomes, and intended comparisons without presenting the numerical results needed to evaluate observed patterns. Here, the record lists 21 outcome measures but 0 analyses, and does not provide demographic estimates or their uncertainty. Transparent reporting of denominators, missingness, analysis methods, and results is necessary for readers to judge both descriptive representativeness and statistical precision.

Finally, the record's planning assumptions should remain distinct from evidence generated by the study. A sample-size illustration based on a stated fraction of an observed effect is a planning calculation, not observed power and not a substitute for outcome results. The central statistical contribution that can be assessed from the record is the planned characterization of treatment initiators; the actual characteristics and their comparative distributions are not reported.

13. Limitations

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

A central limitation for interpretation is that the registry record does not report the numerical primary or secondary results. Without estimates, denominators, or uncertainty measures, the magnitude and precision of the intended characterizations cannot be assessed. The record lists enrollment as 31908 but does not provide group-specific counts or outcome-specific analysis populations.

The population and comparison details are also limited. The eligibility description specifies adults older than 18 years with a recorded type 2 diabetes diagnosis in the UK and prescription initiation between 01 August 2014 and 01 September 2015. It does not report the data source, participant identification procedures, handling of multiple prescriptions, or criteria used to classify a prescription as off-label. These omissions prevent assessment of selection and classification processes.

The registry does not report demographic variables in detail, their operational definitions, or their outcome-specific denominators. It also does not describe missing-data methods, adjustment for differences between treatment groups, inferential procedures, or multiplicity handling. Since the study is not described as randomized and the groups are not characterized as like-for-like controls, confounding is a particular concern for any comparative interpretation.

Finally, the registry's adverse-event fields list 0 for serious and other events at risk and affected across the named groups, deaths are null, and no serious-event terms are listed. The record does not explain these entries or establish how adverse events were collected. They cannot be interpreted as a complete safety assessment or as evidence that no events occurred.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Characteristics of Empagliflozin Initiators — NCT03050619.
• Registry record last updated 2019-01-25; status last verified 2018-08.
• Results first posted 2019-01-25.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

17. Record Summary

NCT03050619 is a completed UK study of characteristics among people initiating empagliflozin or other glucose-lowering drugs. It lists an enrollment of 31908, Boehringer Ingelheim as sponsor, a study start of 2016-09-10, and primary completion on 2016-10-14. The eligible population is described as adults older than 18 years with recorded type 2 diabetes who initiated the specified prescriptions between 01 August 2014 and 01 September 2015.

The primary outcome concerns baseline demographics, expressed as a percentage of participants. Secondary aims include off-label use and other baseline characteristics. The registry lists 21 outcome measures and 0 analyses, but does not report numerical outcome results, group-specific denominators, or statistical analysis details. Treatment groups are observational initiation groups, not a like-for-like randomized comparison. The record therefore supports description of the study's aims and design, but not conclusions about observed differences, treatment effects, or comparative safety.