Doxycycline PEP for Prevention of Sexually Transmitted Infections Among Kenyan Women Using HIV PrEP: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Doxycycline PEP for Prevention of Sexually Transmitted Infections Among Kenyan Women Using HIV PrEP (NCT04050540), a randomized phase 4 study. The primary outcome is Incidence of N. Gonorrhoeae, C. Trachomatis, or Early Syphilis Infection by Laboratory-based Diagnosis.

2. Clinical Question

The clinical question is whether assignment to doxycycline post-exposure prophylaxis (dPEP), alongside standard of care, is associated with a different incidence rate of laboratory-diagnosed bacterial sexually transmitted infections than assignment to standard of care alone. The population in the registry record is Kenyan women taking HIV pre-exposure prophylaxis (PrEP). The primary outcome combines diagnoses of Neisseria gonorrhoeae, Chlamydia trachomatis, or early syphilis.

This question concerns an event rate, not simply the proportion of participants with at least one infection. A rate uses time at risk as its exposure basis, so it can reflect both the number of events and the amount of person-time observed. The registry record identifies a combined endpoint; it does not report separate primary results for each infection type.

The trial’s stated broader aims included assessing acceptability, cost, and effects on tetracycline resistance, as well as using questionnaires, focus group discussions, SMS, and in-depth interviews to study acceptability and changes in sexual behavior. Those aims do not alter the specified primary comparison. The registry record posts one outcome measure and no secondary outcomes.

3. Trial Design

This was a randomized, parallel-group, phase 4 trial with no masking. The registry record lists an enrollment of 449 participants: 224 in the dPEP intervention arm and 225 in the standard-of-care arm. The trial status is completed. The recorded start date is 2020-02-05, and the primary completion date is 2022-10-30.

Participants were assigned to dPEP plus standard of care or to standard of care alone. In the intervention arm, participants were instructed to take doxycycline 200 mg, described as two 100mg capsules, orally within 24 hours and up to 72 hours after each condomless sex act. The registry record does not provide further detail about the contents of standard of care, the randomization sequence, allocation concealment, or operational procedures for assignment.

The primary endpoint was assessed over 12 months post enrolment. It combines laboratory-based diagnoses of Neisseria gonorrhoeae, Chlamydia trachomatis, or early syphilis, with incidence expressed as events per 100 person-years. The endpoint description gives examples of diagnostic criteria: NAAT for gonorrhoea or chlamydia, and a four-fold increase in non-treponemal titers for syphilis.

The primary analysis population is described as intention to treat. In general, an intention-to-treat comparison retains participants in the groups to which they were assigned, regardless of whether they followed the assigned regimen; the registry record does not provide further detail on how that principle was applied in particular situations. The number of participants contributing outcome data, follow-up completeness, and handling of missing outcome or exposure time are not reported in the registry record.

The registry record identifies the University of Washington as sponsor. It gives no original sample-size justification or detailed account of the trial’s statistical plan. Those omissions mean that the posted comparison can be described, but the design rationale and prespecified analytic choices cannot be reconstructed from this record alone.

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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Intention to treat, STI incidence events per 100 person-years

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 Incidence of N. Gonorrhoeae, C. Trachomatis, or Early Syphilis Infection by Laboratory-based Diagnosis, measured over 12 months post enrolment.

The primary endpoint is the combined incidence of Neisseria gonorrhoeae, Chlamydia trachomatis, or early syphilis diagnosed by laboratory-based methods. The registry record specifies a 12-month period post enrolment and reports the outcome as events per 100 person-years. For gonorrhoea and chlamydia, the description gives a positive NAAT as an example of diagnosis; for syphilis, it gives a four-fold increase in non-treponemal titers.

An incidence rate relates the number of events to the time participants contribute while at risk. In general, this can be useful when participants contribute different amounts of observation time, and when an outcome may occur more than once. A rate is therefore not interchangeable with the percentage of participants who had any infection. The registry record does not report a separate count of participants with at least one endpoint event.

Combining three infections gives a single overall measure for the primary comparison, but it does not show whether the component infections had similar patterns. The registry record posts no separate primary estimates for gonorrhoea, chlamydia, or early syphilis, and no information about the distribution of the combined events across those diagnoses. The endpoint description states that laboratory-based diagnosis was used, but does not report further detail on testing schedules, blinded assessment, or adjudication.

The combined rate is relevant to the stated trial question about bacterial STI incidence. Interpretation should remain specific to that definition, its diagnostic criteria, and the 12-month post-enrolment timeframe. It is not a measure of every possible sexually transmitted infection or of other outcomes mentioned among the trial’s broader aims.

7. Results Posted in the Registry

• dPEP Intervention Arm: 25.10 versus 29.00 (Events per 100 person-years) on control; rate ratio 0.87 (95% CI 0.61–1.24, approximate, Poisson model); p ≈ 0.427.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is an event rate, and the reported unit is events per 100 person-years. General explanation: an event rate divides events by accumulated time at risk, then expresses the result on a chosen exposure scale. This differs from a risk or proportion, which would use the number of participants meeting an outcome over a specified period as its basis. When follow-up time differs among participants, an exposure-based rate can use that variation rather than treating all participants as if they contributed identical observation time.

The registry reports rates of 25.1 and 29.0 events per 100 person-years in the dPEP and standard-of-care arms, respectively. The reported comparative measure is a rate ratio. General explanation: a rate ratio compares the event rate in one group with the event rate in another. A value of 1 represents equal rates; values below or above 1 represent lower or higher rates in the numerator group, respectively. The posted estimate was below 1, but its confidence interval included 1. That is why the point estimate should not be read alone as establishing a group difference.

The registry comparison is identified as Poisson-based. General explanation: Poisson methods are commonly used for event counts, often with exposure time accounted for when estimating rates. A model of this kind relies on assumptions about the relationship between event counts and exposure. Over-dispersion is a general concern when observed variability is greater than a simple Poisson model expects; if not accounted for, uncertainty estimates can be too small. The registry record does not report a dispersion estimate or describe any adjustment for over-dispersion, so no conclusion about whether it was present or addressed can be made.

The reported confidence interval for the rate ratio is an interval estimate of uncertainty around the comparative rate. General explanation: a confidence interval is constructed by a method that, under its assumptions and repeated use, has a stated coverage property. It is not a range that guarantees the true effect lies within its endpoints, nor does it describe the range of individual participant outcomes. In this trial, the interval crosses the null rate ratio of 1. The record’s lower endpoint is printed with more than three decimal places, so it is not repeated; the substantive interpretation is that the interval includes values on both sides of 1.

The unadjusted p-value is 0.427. General explanation: a p-value measures how incompatible the observed data, or data more extreme, are with a specified null model, assuming the model and its assumptions. It is not the probability that the null hypothesis is true, the probability that the result occurred by chance, or a measure of effect size. The registry identifies this value as unadjusted; an adjusted p-value is not reported. The record does not explain whether the comparison was prespecified in a detailed analysis plan or whether other planned analyses were considered.

The primary population is labeled intention to treat. General explanation: intention-to-treat analysis aims to compare participants according to randomized assignment, preserving the allocation basis of the trial. The record does not explain how participants with missing laboratory results, incomplete follow-up, or deviations from assigned treatment were handled. Nor does it report the amount of person-time by arm or the count of participants contributing endpoint data. These details matter because rate estimation depends on both observed events and exposure time.

The combined endpoint includes three specified diagnoses. Combining outcomes can produce a broader measure and may increase the number of events considered in a primary comparison, but it can conceal differences among components. The registry does not post component-specific rates, so the overall estimate cannot establish whether each infection type followed the same pattern. It also does not report separate estimates for repeated events versus participants with any event.

The study is described as randomized, parallel, and unmasked. General explanation: randomization is intended to make assigned groups comparable on average at the start of a trial, while a parallel design follows distinct groups under their assigned conditions. Lack of masking can be relevant to behavior, reporting, and care, but the primary endpoint is described as laboratory-based. The record does not explain whether laboratory staff were masked or how testing and outcome ascertainment were organized. Consequently, the possible influence of unmasked assignment on outcome ascertainment cannot be assessed from the registry record.

The posted analysis is a result summary rather than a full statistical report. The record does not report model covariates, a detailed variance method, treatment of recurrent events, or sensitivity analyses. These are not details that can be inferred from the rate ratio, confidence interval, or p-value. The numerical result should therefore be interpreted as the posted, unadjusted primary comparison, with those analytic details unresolved.

10. Confidence Intervals and Precision

The rate ratio estimate was below 1, and its reported confidence interval extended from below 1 to above 1. Since 1 is the null value for a rate ratio, the interval includes no relative difference between groups. It also includes rate ratios indicating a lower rate in the dPEP arm and values indicating a higher rate. This breadth means the point estimate does not precisely settle the direction of the underlying relative effect.

The interval is the most direct posted indication of precision for the comparative estimate. General explanation: a confidence interval expresses statistical uncertainty around an estimate under the method’s assumptions. A wider interval indicates less precision than a narrower interval on the same scale. It does not mean that every value inside is equally likely, and it does not describe variation in individual participants’ experiences.

The registry’s lower confidence limit is printed with more than three decimal places, so its digits are not reproduced. The conclusion relevant to interpretation is unaffected: the interval spans the null value. The upper limit is 1.236. The record does not provide confidence intervals for the two arm-specific rates, so precision for those rates cannot be assessed from the posted results.

The unadjusted p-value of 0.427 is consistent with substantial uncertainty about whether the observed rate ratio differs from 1 under the analysis used. A p-value does not replace the interval: it gives a test-based summary, while the interval displays a range of effect values compatible with the data and method. Neither statistic establishes equivalence, and the record does not report an equivalence margin or a non-inferiority framework.

The total enrollment was 449, but enrollment alone does not determine precision. Event frequency, time at risk, endpoint measurement, and the analysis method also matter. The registry record does not report the achieved person-time, participant-level outcome completeness, or an original precision target. Therefore, the posted interval can be described, but it cannot be compared with an unreported design target.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning estimates are illustrative calculations based on the observed effect, not the original trial design. They give participants per arm, not total events. The calculation assumptions listed in the registry are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect.

For the full observed effect, the planning estimates are 2,797 participants per arm for 80% power and 3,744 participants per arm for 90% power. For planning at 75% of the observed effect, the corresponding estimates are 4,878 and 6,530 participants per arm. For planning at 50% of the observed effect, they are 10,769 and 14,417 participants per arm.

In general, an 80% power target means that, under the calculation assumptions and for the effect specified, the design would have an 80% chance of reaching the stated statistical significance criterion if that effect were the true one. The 90% column uses a higher target chance, which requires more participants under the same assumptions. Smaller effects are harder to distinguish from random variation, so planning to detect a smaller fraction of the observed effect requires more participants.

These figures are not observed power. They are not evidence about the probability that the trial’s result is correct, and they do not describe the probability that a future study will reproduce the observed estimate. The registry does not report the original sample-size justification or the assumptions used to choose the enrolled sample size. The planning calculations use the observed effect as their reference, so their values should be read as hypothetical planning illustrations rather than as a reconstruction of the trial’s design rationale.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, λ₀ and λ₁ are event rates per patient-year, and T is follow-up per participant in years. This assumes Poisson counts and equal follow-up.

Worked calculation for this trial

1. Control rate λ₀ = 0.2900; treatment rate λ₁ = 0.2510 events per patient-year; follow-up T = 1.00 years.
2. n = (1.96 + 0.842)² × (1/0.2900 + 1/0.2510) / (1.00 × [ln(0.866)]²) = 2,796.6, rounded up to 2,797 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
lambda0 <- 0.29                      # control rate, events per patient-year
lambda1 <- 0.251                      # treatment rate
tt      <- 1                       # follow-up per participant, years
n_per_arm <- function(power, frac = 1) {
  l1 <- lambda0 * (lambda1 / lambda0)^frac   # smaller effect = fraction of the observed log rate ratio
  ceiling(round((z_a + qnorm(power))^2 * (1 / lambda0 + 1 / l1) / (tt * log(l1 / lambda0)^2), 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Simulation 1: Poisson counts. Simulation 2: over-dispersed counts (variance = 1.5 x mean), analysed with the matching variance.
set.seed(4050540)
n    <- n_per_arm(0.80)
reps <- 20000
m0 <- lambda0 * n * tt; m1 <- lambda1 * n * tt
x0 <- rpois(reps, m0); x1 <- rpois(reps, m1)
z  <- log(x1 / x0) / sqrt(1 / x0 + 1 / x1)
cat(sprintf("Simulated power, Poisson counts, n = %s per arm: %.3f\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE)))
phi <- 1.5
y0 <- rnbinom(reps, mu = m0, size = m0 / (phi - 1)); y1 <- rnbinom(reps, mu = m1, size = m1 / (phi - 1))
z2 <- log(y1 / y0) / sqrt(phi * (1 / y0 + 1 / y1))
cat(sprintf("Simulated power, over-dispersed counts (phi = 1.5), same n: %.3f\n", mean(abs(z2) > z_a, na.rm = TRUE)))

Output from running the code

100% of observed effect: n per arm = 2,797 (80% power), 3,744 (90% power)
 75% of observed effect: n per arm = 4,878 (80% power), 6,530 (90% power)
 50% of observed effect: n per arm = 10,769 (80% power), 14,417 (90% power)
Simulated power, Poisson counts, n = 2,797 per arm: 0.799
Simulated power, over-dispersed counts (phi = 1.5), same n: 0.627

The first simulation draws Poisson counts and should land close to the target power. The second allows extra variability (variance 1.5 times the mean, a purely illustrative value), which lowers the power at the same sample size, so rate trials usually need more participants than the Poisson formula suggests.

13. Multiplicity, Missing Data and Analysis Populations

The registry record posts one outcome measure and no secondary outcomes. It reports a primary comparison but does not provide a detailed statistical analysis plan or describe whether additional outcomes or subgroup analyses were planned. Because the record does not identify a family of multiple confirmatory tests, the extent of any multiplicity issue cannot be assessed. An adjustment for multiple comparisons is not reported; the posted p-value is explicitly unadjusted.

The primary population is identified as intention to treat. That label gives an important indication of the intended comparison population, but the registry does not describe how the analysis handled participants who did not follow assigned treatment, had missing laboratory tests, had incomplete follow-up, or contributed limited person-time. The number with available endpoint data in each arm and the amount of missing data are not reported in the registry record.

General explanation: missing outcome information can affect a comparison if missingness is related to assignment, infection risk, or other factors. For a rate endpoint, incomplete observation time also matters because both events and time at risk contribute to the rate. Analytic approaches can make different assumptions about missingness, but the record does not identify an approach or provide enough information to evaluate one.

The registry gives no detailed account of sensitivity analyses, imputation, censoring rules, or handling of recurrent infections. It also does not report a protocol-specific definition of the analysis set beyond the intention-to-treat description. These omissions do not demonstrate that such procedures were absent from the full trial conduct; they mean the procedures are not reported in the registry record. Interpretation should therefore stay with the posted primary comparison and avoid assumptions about unreported analyses.

14. 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.

Most frequent non-serious adverse event terms

The registry reports zero serious adverse events in the dPEP arm, among 224 participants at risk, and zero in the standard-of-care arm, among 225 participants at risk. It also reports zero deaths in each arm. No serious adverse event terms are listed. These are the serious-event counts recorded in the registry; they should not be read as evidence that no event of any kind occurred.

For other adverse events, the record reports 40 affected participants among 224 at risk in the dPEP arm and 12 among 225 at risk in the standard-of-care arm. The registry does not provide further detail in this record about the event types, severity, timing, relatedness, or how events were collected. The counts alone do not establish that assignment caused an event or explain why the groups differ in their reported counts.

These figures are descriptive safety information, separate from the primary STI incidence endpoint. They do not provide a complete assessment of safety, and the record does not report comparative statistical analyses for adverse events. Interpretation is limited to the counts and denominators posted.

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

The primary result describes a lower observed incidence rate in the dPEP group than in the standard-of-care group: 25.1 versus 29.0 events per 100 person-years. The rate ratio estimate was below 1, but the confidence interval included 1, and the unadjusted p-value was 0.427. The posted data therefore show a numerical difference, while retaining uncertainty about the direction and size of the underlying relative difference.

The result does not establish that the groups had equal rates. A confidence interval that includes the null is not proof of equivalence. The registry does not report an equivalence or non-inferiority design, nor an equivalence margin. Conversely, the lower point estimate is not by itself proof of a difference. The estimate and its uncertainty need to be considered together.

The endpoint applies to the combined rate of laboratory-diagnosed gonorrhoea, chlamydia, or early syphilis during 12 months post enrolment. It does not give separate conclusions for each infection, the proportion of participants with any infection, or outcomes outside the stated endpoint. Nor does the primary comparison resolve the trial’s broader questions about acceptability, cost, sexual behavior, or tetracycline resistance; secondary outcomes are not posted in the registry record.

Interpretation is also bounded by the information recorded about the analysis. The primary population is described as intention to treat, but missing-data procedures, follow-up completeness, and person-time details are not reported. The result should be understood as the registry’s posted comparison, rather than as a complete account of all analyses or trial findings.

In general, a randomized comparison estimates the difference between assigned groups in the study setting. Its relevance beyond that setting depends on population, intervention instructions, comparator, outcome definition, and follow-up. The record concerns Kenyan women using HIV PrEP and a specific doxycycline post-exposure regimen; it does not justify extending the estimate to other populations or regimens. These statistical results describe the posted data and are not medical advice.

16. Why This Trial Matters Statistically

This trial illustrates how the choice of outcome scale shapes interpretation. Its primary endpoint is an incidence rate per 100 person-years, not a simple proportion of participants with an infection. That distinction matters when the scientific question concerns events over follow-up and when exposure time is part of the outcome definition. The rate ratio then expresses the comparison between group rates, while the confidence interval shows uncertainty on the relative scale.

It also shows why a point estimate should not be treated as a stand-alone conclusion. The observed rates were 25.1 and 29.0 events per 100 person-years, and the estimated rate ratio was below 1. Yet the reported confidence interval crossed 1 and the unadjusted p-value was 0.427. A reader can distinguish the direction of the point estimate from the strength and precision of evidence about a group difference.

The combined endpoint has a practical statistical trade-off. It provides one primary measure spanning three bacterial infections, but it does not reveal the separate pattern for each diagnosis. Without component-specific results, the overall rate cannot show whether one component accounts for more of the observed events or whether the components differ in their relative estimates.

Finally, this record illustrates the difference between a posted result and a full analysis report. It identifies the randomized design, outcome, population, rate comparison, and uncertainty measure, but does not report details such as missing-data procedures, dispersion assessment, or an original sample-size rationale. Recognizing those boundaries is part of responsible statistical interpretation: the available estimate can be described precisely without filling gaps with assumptions.

17. 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.
• Rate comparisons use a Poisson approximation unless the sponsor posted confidence limits; real event counts are often over-dispersed.

The registry record leaves several details unresolved that are specific to interpreting this trial. It does not report the amount of person-time accrued in either arm, the number of participants with endpoint data, or follow-up completeness. Since the primary outcome is an event rate, those details would help readers understand the exposure underlying the posted rates and the extent of available outcome information.

The endpoint combines gonorrhoea, chlamydia, and early syphilis, but the registry provides no component-specific event rates or participant counts. The posted result therefore cannot indicate whether the combined estimate reflects similar patterns across diagnoses. The record also does not report the number of participants experiencing at least one infection, so a risk-based interpretation cannot be derived from the primary rate results.

The trial was unmasked, but the record does not report whether laboratory staff or outcome assessors were masked, how often testing was conducted, or how diagnostic results were adjudicated. It describes laboratory-based diagnosis and gives example criteria, but no further operational detail. The degree to which measurement procedures might affect the comparison cannot be determined from this record.

Further analytic limitations include no reported details on over-dispersion, covariate adjustment, recurrent-event handling, missing-data methods, or sensitivity analyses. The registry does not report an original sample-size justification or a detailed statistical plan. The sample-size illustrations based on the observed effect do not fill that gap.

Finally, the registry posts no secondary outcomes despite broader aims concerning acceptability, cost, sexual behavior, and tetracycline resistance. Those questions cannot be evaluated from the primary estimate. These limitations describe what is not reported in the registry record and do not establish that the information was absent from other trial documents.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Doxycycline PEP for Prevention of Sexually Transmitted Infections Among Kenyan Women Using HIV PrEP — NCT04050540.
• Registry record last updated 2024-07-10; status last verified 2024-07.
• Results first posted 2024-07-10.
• 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.

21. Record Summary

The registry describes a completed, randomized, parallel, unmasked phase 4 trial sponsored by the University of Washington. Enrollment was 449 Kenyan women using HIV PrEP, assigned to dPEP plus standard of care or standard of care alone. The dPEP instructions specified doxycycline 200 mg orally within 24 hours and up to 72 hours after each condomless sex act.

The primary outcome was the combined incidence of laboratory-diagnosed gonorrhoea, chlamydia, or early syphilis over 12 months post enrolment, expressed per 100 person-years. Reported rates were 25.1 in the dPEP arm and 29.0 in the standard-of-care arm. The rate ratio estimate was below 1, its confidence interval included 1, and the unadjusted p-value was 0.427.

The primary analysis population is described as intention to treat. The registry does not report secondary outcomes, detailed missing-data procedures, or the original sample-size justification. Serious adverse events and deaths were recorded as zero in both arms; other adverse-event counts were 40 and 12, respectively.