Study on Medical Records of Women Using an Intrauterine Device (IUD) to Analyze the Risks That the IUD Will be Expelled or Perforates the Womb in Relation to Breastfeeding, the Point in Time When the IUD Was Inserted After Childbirth and in Relation to Different Types of IUDs: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study on Medical Records of Women Using an Intrauterine Device (IUD) to Analyze the Risks That the IUD Will be Expelled or Perforates the Womb in Relation to Breastfeeding, the Point in Time When the IUD Was Inserted After Childbirth and in Relation to Different Types of IUDs (NCT03754556), a registered phase not stated study. The primary outcome is Crude Incidence Rate of Uterine Perforation Stratified by Breastfeeding Status Among Women Within 52 Weeks Postpartum-first Observed IUD Insertions.

2. Clinical Question

3. Trial Design

This was a completed, non-randomized study of electronic medical records, rather than a trial in which participants were assigned to different IUDs or postpartum insertion times. The registry describes women with IUDs and records from Kaiser Permanente Northern California, Kaiser Permanente Southern California, Kaiser Permanente Washington, and the Regenstrief Institute. Allocation is not stated in the record, and the comparisons should not be read as like-for-like randomized groups.

The study examined uterine perforation and IUD expulsion in relation to breastfeeding at insertion, timing of insertion after childbirth, and IUD type. Breastfeeding comparisons focused on insertions within 52 weeks postpartum. Timing comparisons included first observed insertions at six weeks or less, more than six through 14 weeks, and more than 14 through 52 weeks, compared with insertion more than 52 weeks after childbirth. The description also refers to a finer postpartum timing analysis, including insertion from 0 to 3 days and from 4 days through six weeks. Women without a recorded delivery within the past 52 weeks were included in the more-than-52-week reference group.

The registry lists a total enrollment of 326658 and a completed status. The listed study dates are 2018-12-03 for study start and 2019-11-29 for primary completion. The official title identifies the work as a postmarketing requirement for Mirena, while the study description addresses IUDs more broadly. The record does not report the protocol details needed to determine how each database contributed or how records were harmonized.

The primary outcome was a crude uterine perforation incidence rate stratified by breastfeeding status among women within 52 weeks postpartum at first observed insertion. Listed secondary outcomes include expulsion rates and adjusted hazard ratios, as well as analyses by postpartum timing. Numerical estimates and detailed analysis procedures are not reported in the registry record.

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: Between 01 January 2001 and 30 April 2018, 326658 participants were identified: 322898 participants with intrauterine device (IUD) type available and 94817 participants with breastfeeding status available (and IUD insertion within 12 months postpartum)

Overall Study

Analysis population for the primary outcome, as defined in the registry: Women who were within 52 weeks postpartum and were or were not breastfeeding at the time of IUD insertion

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.

6. Primary Endpoint

The registry lists the primary outcome as Crude Incidence Rate of Uterine Perforation Stratified by Breastfeeding Status Among Women Within 52 Weeks Postpartum-first Observed IUD Insertions, measured over Up to 11 years.

The primary endpoint is the crude incidence rate of uterine perforation, separated by breastfeeding status among women whose first observed IUD insertion occurred within 52 weeks postpartum. The registry defines the rate as the number of outcomes occurring during person-time at risk divided by the total person-time at risk, and expresses it per 1000 person-years. The stated timeframe is up to 11 years.

A rate uses time at risk in its denominator, rather than simply counting outcomes among all participants. In general, this can make comparisons more informative when people contribute different amounts of observed follow-up. A crude rate is not adjusted for differences between comparison groups, however. It describes the observed event frequency relative to observed person-time; by itself, it does not establish that breastfeeding status caused any difference.

The record identifies a 95% confidence interval as the measure of precision for this endpoint. In general, a confidence interval conveys uncertainty around an estimated rate, with wider intervals indicating less precision. The record does not report the rate estimates, interval limits, event counts, or person-time for the breastfeeding groups, so the endpoint's observed magnitude and precision cannot be assessed from the registry record.

The endpoint is clinically relevant to the study's stated question because it focuses on uterine perforation after IUD insertion in the postpartum population. But the listed primary endpoint alone does not answer every question in the study description: expulsion, postpartum timing, IUD type, and potential interactions are addressed through other listed outcomes or analyses.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The study is described as an analysis of electronic medical records, not a randomized comparison. The registry names four data sources: Kaiser Permanente Northern California, Kaiser Permanente Southern California, Kaiser Permanente Washington, and the Regenstrief Institute. It describes women with IUDs and electronic health records in those databases. In general, observational records allow researchers to study exposures and outcomes recorded in routine care, but group membership is not assigned randomly. The record does not report detailed eligibility rules, database-specific contributions, or methods for joining or harmonizing records.

What does a crude incidence rate represent?

The primary endpoint is a crude incidence rate of uterine perforation, stratified by breastfeeding status among women within 52 weeks postpartum at first observed insertion. The registry defines the calculation as the number of outcomes occurring during person-time at risk divided by total person-time at risk, and gives the unit as per 1000 person-years. Person-time represents the period during which individuals are considered at risk and under observation for the outcome. In general, using person-time accounts for differing lengths of observation in the denominator. A crude rate remains unadjusted for differences in participant characteristics.

The endpoint specifies a 95% confidence interval. In general, an interval describes uncertainty around an estimate under the statistical model used to construct it. It is not a guarantee that the true value falls within the interval, and its interpretation depends on the method and assumptions. The registry record does not report the interval calculation method, the actual interval limits, the rate estimates, the event counts, or accumulated person-time.

Why are adjusted hazard ratios listed?

The outcome list includes adjusted hazard ratios for breastfeeding status and expulsion, postpartum timing and expulsion, and postpartum timing analyses using different cut points. In general, a hazard ratio compares event rates over time between groups, conditional on the model and any covariates included. An adjusted estimate attempts to account for specified measured characteristics. It is not equivalent to a crude incidence rate, and it does not by itself show that an exposure caused an outcome. The registry record does not report the hazard-ratio model, its covariates, the handling of time-varying information, or checks of model assumptions.

Other listed measures include crude incidence rates and cumulative incidence for expulsion. In general, cumulative incidence describes the probability of an event over follow-up, while a rate relates event occurrence to person-time. Their values answer different questions and depend on how follow-up and competing events are treated. The registry does not report those methodological details or numerical results.

How are the exposure groups defined?

The breastfeeding comparison concerns status at insertion among women within 52 weeks postpartum. Postpartum timing comparisons use first observed insertion in intervals after childbirth and compare them with an insertion more than 52 weeks after childbirth; the description includes women with no recorded delivery in the past 52 weeks in that reference group. The registry also describes a finer timing analysis that includes insertion from 0 to 3 days and from 4 days through six weeks, as well as later postpartum intervals. It does not report how missing or uncertain dates were classified.

The study also considers different IUD types and describes interaction questions involving LNG-releasing versus copper IUDs in relation to breastfeeding or insertion timing. In general, an interaction analysis asks whether an association differs across levels of another exposure. The registry description states that these questions were assessed, but does not report the corresponding model specifications or findings.

What can the planning illustrations tell readers?

The accompanying sample-size illustrations are planning calculations based on the observed effect, not the study's own design or observed power. The stated assumptions are 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 time-to-event calculations, the illustrations give the total number of events; other illustrations give participants per arm. In general, such calculations show how planning requirements depend on assumptions about the target effect and statistical procedure. They do not show that the study had a particular probability of detecting an effect, and no such probability should be inferred from them.

The registry does not report its own sample-size justification, interim analyses, or statistical analysis plan. It also lists 39 outcome measures and 0 analyses, but does not provide numerical findings for the endpoints described. These record-level entries do not supply enough information to reconstruct the analysis or evaluate the precision of specific comparisons.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 39 outcome measures, covering uterine perforation, expulsion, breastfeeding status, postpartum timing, and IUD type, among other described comparisons. It records 0 analyses, and numerical outcome estimates are not reported in the registry record. The record does not state whether a multiplicity adjustment was planned or applied, how primary and secondary comparisons were prioritized beyond the named primary endpoint, or whether any analyses were exploratory.

In general, examining many outcomes or comparisons creates multiple opportunities for chance findings. Procedures such as prespecifying a primary comparison or adjusting for multiplicity can influence how evidence is interpreted. No specific procedure should be assumed for this study: its multiplicity strategy is not reported in the registry record. The number of listed outcomes alone does not establish that all were tested or that any particular adjustment was necessary or used.

The registry identifies the population for the primary endpoint and the electronic record databases, but does not describe analysis populations in detail. It does not report whether participants with incomplete records were excluded, how duplicate or inconsistent records were handled, or the number contributing data to each endpoint. Missing-data methods, missingness by exposure group, and sensitivity analyses are not reported in the registry record.

The record also does not specify definitions for censoring, follow-up completion, or competing events. In general, these choices can affect person-time rates, cumulative incidence, and time-to-event estimates. Without these details and numerical results, readers cannot evaluate how missing records or incomplete follow-up may have influenced the reported comparisons.

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

The registry record documents a study designed to examine uterine perforation and IUD expulsion in relation to breastfeeding, postpartum insertion timing, and IUD type. It gives the primary endpoint's definition and unit, identifies the electronic record databases, and lists the broad outcome families. It does not report numerical estimates for the primary or secondary outcomes. As a result, it cannot establish the observed direction, magnitude, or precision of an association.

The primary outcome is a crude rate per 1000 person-years, not an adjusted comparison or a direct measure of an individual's probability of an event. In general, crude rates describe observed outcomes relative to observed time at risk. They may reflect differences between groups as well as the exposure of interest. The listed adjusted hazard ratios may address measured factors, but the record does not provide their estimates or explain the adjustment models.

The study's observational design also limits causal interpretation. In general, differences in recorded breastfeeding status, insertion timing, or IUD type may coexist with other characteristics that relate to the outcome. Statistical adjustment can account for measured factors under suitable assumptions, but does not automatically remove confounding from unmeasured or poorly measured factors. The registry record does not provide enough methodological detail to assess these issues for this analysis.

The enrollment figure of 326658 is not a substitute for outcome-specific sample sizes, event counts, or person-time. Those quantities are not reported in the registry record. Likewise, the stated follow-up timeframe of up to 11 years does not reveal how much follow-up was available for each comparison.

Accordingly, the record supports a description of the research question and planned outcome measures, not a conclusion about whether breastfeeding, postpartum timing, or IUD type was associated with a higher or lower observed risk. It also does not support a treatment recommendation or a claim that any IUD type works better or worse. Such conclusions would require outcome estimates, uncertainty measures, and fuller methods information.

12. Why This Trial Matters Statistically

This study addresses a statistical challenge common to research based on routine medical records: comparing outcomes across groups defined by recorded exposures rather than randomized assignment. Its questions involve breastfeeding at insertion, how long after childbirth insertion occurred, and IUD type. These comparisons may be relevant to understanding how observed outcomes vary across clinical circumstances, while still requiring careful attention to confounding and measurement.

The primary endpoint illustrates why the choice of scale matters. A crude incidence rate relates uterine perforations to person-time at risk, while other listed outcomes include cumulative incidence and adjusted hazard ratios. In general, these measures represent different summaries of event occurrence and rely on different assumptions. Reading them together can be informative only when definitions, follow-up, analysis populations, and uncertainty are clearly reported.

The record also describes comparisons across postpartum timing intervals and potential interaction questions involving LNG-releasing and copper IUDs. In general, interaction analyses ask whether an association varies by another factor, rather than merely whether groups differ overall. Such questions can require careful modeling and sufficiently detailed reporting. The registry record does not report the estimates or the methods needed to assess those analyses.

Its statistical importance is therefore twofold: it sets out a broad observational research question, and it highlights the need for transparent reporting of outcome-specific denominators, event counts, estimates, confidence intervals, adjustment methods, and missing-data handling. The registry record does not include those numerical findings or several of those methods. Readers can identify what the study set out to examine, but cannot use this record alone to assess the evidence for the stated associations.

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.

The registry record leaves key details needed to assess the analysis unreported. It does not give numerical values for the primary uterine perforation rates, their 95% confidence intervals, event counts, or person-time. It also does not report the results of the listed expulsion, cumulative-incidence, or adjusted hazard-ratio outcomes. Without these quantities, the magnitude and precision of the comparisons cannot be evaluated.

Although the record identifies several electronic medical record databases, it does not explain how exposure and outcome information was defined or validated, how records from different databases were combined, or whether all databases contributed to every endpoint. The definitions of missing or uncertain breastfeeding status, delivery dates, insertion dates, and outcomes are not reported in the registry record.

The record does not specify the covariates used in adjusted analyses, the model forms, the treatment of censoring or competing events, or whether assumptions were checked. It also does not report analysis-specific populations, outcome-specific denominators, or sensitivity analyses. These omissions limit assessment of potential confounding, follow-up, and robustness.

The study is observational, and the registry does not state an allocation mechanism. The comparisons therefore should not be interpreted as randomized contrasts. The record also does not report its own sample-size justification or statistical analysis plan, and it does not describe a multiplicity strategy for the many listed outcomes and comparisons. The enrollment figure alone cannot resolve these limitations or establish the precision of any particular analysis.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Study on Medical Records of Women Using an Intrauterine Device (IUD) to Analyze the Risks That the IUD Will be Expelled or Perforates the Womb in Relation to Breastfeeding, the Point in Time When the IUD Was Inserted After Childbirth and in Relation to Different Types of IUDs — NCT03754556.
• Registry record last updated 2020-12-04; status last verified 2020-11.
• Results first posted 2020-12-04.
• 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

The registry describes a completed, observational electronic medical record study of IUD use, with a listed enrollment of 326658. The study examined uterine perforation and expulsion in relation to breastfeeding status at insertion, postpartum insertion timing, and IUD type, using records from Kaiser Permanente Northern California, Kaiser Permanente Southern California, Kaiser Permanente Washington, and the Regenstrief Institute.

The primary endpoint is the crude uterine perforation incidence rate per 1000 person-years among women within 52 weeks postpartum, stratified by breastfeeding status at first observed IUD insertion. Its stated timeframe is up to 11 years, and a 95% confidence interval is specified. The registry also lists expulsion measures and adjusted hazard ratios.

Numerical outcome estimates, event counts, person-time, detailed statistical methods, missing-data procedures, and multiplicity methods are not reported in the registry record. The record therefore describes the study questions and outcome measures, but does not permit interpretation of the observed associations.