Naltrexone for Opioid Dependent Released Human Immunodeficiency Virus Positive (HIV+) Criminal Justice Populations: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Naltrexone for Opioid Dependent Released Human Immunodeficiency Virus Positive (HIV+) Criminal Justice Populations (NCT01246401), a randomized phase 1, phase 2 study. The primary outcome is Participants Who Had Undetectable HIV-1 RNA Levels at Less Than 400 Copies/mL at Six Month.

2. Clinical Question

The clinical question addresses a transition period in which participants move from the structure of incarceration into community life. The trial description frames monthly medication as one possible part of that transition and compares it with placebo. Random allocation and masking were intended to support a group comparison, while the primary endpoint focuses on a laboratory measure rather than a self-reported outcome.

The endpoint is defined using a specific viral-load threshold: fewer than 400 copies/mL at six months. That question is narrower than whether treatment changes overall HIV care, opioid use, or community outcomes. Those broader topics appear among the stated hypotheses and secondary outcomes, but they should not be treated as interchangeable with the primary endpoint.

3. Trial Design

This was a randomized, placebo-controlled, quadruple-masked, parallel-group trial listed as phase 1 and phase 2. The registry reports enrollment of 151 participants and a completed study. The recorded start was 2011-03, and primary completion was 2016-03. Yale University is listed as the sponsor.

The study compared extended-release naltrexone with placebo. Participants in both groups were to receive intramuscular injections once monthly for six months, with the first injection before release. The registry describes the target population as HIV-infected prisoners meeting DSM-IV criteria for opioid dependence who were transitioning from a correctional setting to the community. Further details about eligibility criteria, recruitment sites, or the allocation ratio are not reported in the registry record.

The primary outcome was the number of participants with HIV-1 RNA below 400 copies/mL at six months. The outcome description says baseline laboratory tests were drawn in prison one to three months before release and that laboratory tests were planned every three months for one year. It identifies the first six months as the treatment period used for the primary outcome. The primary endpoint counts are posted for 66 participants in the extended-release naltrexone group and 27 in the placebo group, rather than for all 151 enrolled participants.

The registry lists several other intended outcomes, including viral suppression at a lower threshold, CD4 cell count, time to opioid relapse, addiction severity, craving, antiretroviral adherence, urine toxicology results, and opioid abstinence. The posted secondary results cover only selected outcomes. Details about the trial’s original sample-size justification, the randomization procedure, allocation concealment, and the full statistical analysis plan 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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: A total of 80 participants had viral load data at 6 months (56 in XR-NTX group and 24 in Placebo group). The remaining 13 participants (10 in XR-NTX group and 3 in Placebo group) with missing viral load data at 6 months were considered as failure - meaning - having a viral load of 400 or more. Thus included into the final analysis.

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 Participants Who Had Undetectable HIV-1 RNA Levels at Less Than 400 Copies/mL at Six Month, measured over 6 months.

The primary endpoint is the number of participants whose HIV-1 RNA level was below 400 copies/mL at six months. Viral load is a laboratory measure of HIV-1 RNA in blood; the endpoint turns that measurement into a binary outcome by classifying each participant as either below the stated threshold or not below it. The registry reports 45 such participants among 66 in the extended-release naltrexone group and 16 among 27 in the placebo group.

A threshold-based endpoint offers a direct way to compare the proportion of participants meeting a specified virologic target at a common follow-up time. It is different from comparing average viral-load changes over time: the primary outcome does not present the size of each participant’s change, but instead counts participants on one side of the threshold. The study summary describes lower changes in HIV-1 RNA as one of its hypotheses, while the posted primary outcome is the count below 400 copies/mL.

The outcome description also states that 13 participants without six-month viral-load data were treated as failures, meaning they were counted as having a viral load of 400 or more. That rule makes missing primary measurements count against meeting the endpoint, rather than excluding those participants from the primary outcome count. The registry does not report the reason those measurements were missing or a sensitivity analysis using a different missing-data assumption.

The endpoint is one marker within a broader set of clinical and behavioral questions. A result for this laboratory threshold alone does not establish what happened on other outcomes, including longer-term viral-load patterns, retention in care, opioid relapse, or reincarceration.

7. Results Posted in the Registry

• Extended-Release Naltrexone: 68.2% (45/66) versus 59.3% (16/27) on control; risk difference 8.9 percentage points (95% CI -12.8 to 30.6), risk ratio 1.15 (0.81–1.64); unadjusted two-proportion z-test p = 0.411.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Particpants Who Had Undetectable HIV-1 RNA Levels at Less Than 50 Copies/mL

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Extended-Release Naltrexone: 60.6% (40/66) versus 40.7% (11/27) on control; risk difference 19.9 percentage points (95% CI -2.1 to 41.8), risk ratio 1.49 (0.91–2.44); unadjusted two-proportion z-test p = 0.081.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Extended-Release Naltrexone and the control. The interval gives the range of effects that remain compatible with the data.

Antiretroviral Therapy (ART) Adherence 100%

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Extended-Release Naltrexone: 37.9% (25/66) versus 44.4% (12/27) on control; risk difference -6.6 percentage points (95% CI -28.7 to 15.5), risk ratio 0.85 (0.51–1.44); unadjusted two-proportion z-test p = 0.557.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Extended-Release Naltrexone and the control. The interval gives the range of effects that remain compatible with the data.

Opioid Abstinence at 6 Months for Those With More Than 4 Injections

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • 4 or More XR-NTX Injections: 81.2% (13/16) versus 31.0% (18/58) on control; risk difference 50.2 percentage points (95% CI 27.7 to 72.7), risk ratio 2.62 (1.67–4.11); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for 4 or More XR-NTX Injections than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes address several different domains, including a lower viral-load threshold, ART adherence, and opioid abstinence. Their numerical pattern is mixed rather than uniformly favoring one group across outcomes. Some of the posted intervals include no difference, while the injection-based abstinence comparison is more separated from no difference. These outcomes are not all measuring the same construct, and their contrasts are not all defined in the same way.

The lower-threshold viral-suppression result is conceptually related to the primary endpoint because both use HIV-1 RNA at six months, but it applies a different threshold. ART adherence is a care-process outcome rather than a laboratory measure. The injection-based abstinence result is a behavioral outcome and compares participants according to injection exposure, combining placebo participants with those who received three or fewer injections as its reference group. It therefore does not answer the same randomized question as the primary analysis.

The primary result remains the most direct posted answer to the trial’s designated primary question. Secondary results can add context, but they do not overturn or resolve the uncertainty in the primary comparison. A result on a different outcome, time point, or comparison group should not be treated as confirmation of the primary endpoint.

Unadjusted secondary tests carry less weight when many outcomes are examined. Without a reported multiplicity plan, an isolated p-value may be more likely to appear notable by chance across the collection of comparisons. The registry does not report a multiplicity adjustment or an ordering that identifies which secondary findings were intended to be confirmatory. Results for the other outcomes named in the trial description, including several opioid and community outcomes, are not reported in the registry record.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Particpants Who Had Undetectable HIV-1 RNA Levels at Less Than 50 Copies/mL
• CD4 Cell Count (Cells/mL)
• Time to Opioid Relapse or End of Intervention
• Addiction Severity
• Craving for Opioids
• Antiretroviral Therapy (ART) Adherence 100%
• Participants With Opiate Abstinence Via By Doing Urine Toxicology Test
• Opioid Abstinence at 6 Months for Those With More Than 4 Injections

9. Statistical Interpretation

10. Statistical Methods Explained

The results involve two distinct kinds of statistical statements: estimates of group differences and tests of a null hypothesis. An estimate describes the magnitude and direction of a contrast, while a p-value describes how compatible the data are with a particular null model under the assumptions of the test. Neither one, by itself, describes the importance of an outcome or the quality of the trial.

What is a risk difference?

For a binary outcome, the risk difference is the event proportion in one group minus the event proportion in the comparison group. Here, the event is having HIV-1 RNA below the stated threshold at six months. A risk difference of zero corresponds to equal proportions. A positive value means the event proportion is higher in the extended-release naltrexone group; a negative value means it is lower. The recalculated primary risk difference is 0.089. This is an absolute contrast, not a relative measure such as a risk ratio.

The risk difference is useful because it stays on the scale of the outcome proportions and expresses the contrast directly. Other measures, such as risk ratios and odds ratios, express relative comparisons and can give a different impression of magnitude. They answer related but not identical questions. No one measure should be read without its uncertainty and the clinical context of the outcome.

What does a confidence interval communicate?

A confidence interval describes the uncertainty around an estimate under a statistical procedure. In repeated use of that procedure under its assumptions, intervals constructed at a given confidence level would cover the target quantity at the stated rate. It is not a probability statement that the particular interval contains the true value. For the primary endpoint, the 95% interval for the recalculated risk difference runs from -0.128 to 0.306. Because it includes zero and extends across positive and negative values, it does not tightly distinguish among a range of possible group contrasts.

The endpoints of the interval should not be confused with observed group proportions. They describe uncertainty in the estimated contrast, not additional participants or outcomes. A broad interval can reflect limited information, including the size of the groups and the variability inherent in the outcome. General statistical interpretation also depends on whether the analysis assumptions are reasonable and whether the outcome and missing-data rules were specified appropriately.

How should p-values be read?

A p-value is calculated under a specified null hypothesis and statistical model. It indicates how unusual data at least as incompatible with that null would be under the model; it does not give the probability that the null hypothesis is true. Nor does a larger or smaller p-value measure the clinical importance of the observed difference. The recalculated unadjusted primary p-value is 0.411, while the registry-reported analysis lists p = 0.431 from a Welch’s T Test. Those are distinct entries and should be reported as such.

For a binary endpoint, an analysis can be framed around event counts or proportions. The registry’s reported primary method is labeled Welch’s T Test, but the method parameters and test statistic are not listed. A Welch test is generally used to compare means without assuming equal variances. The record does not explain how that method was applied to this primary count outcome. The recalculated comparison is an unadjusted comparison based on the posted summary counts; it is not a reconstruction of an unavailable individual-participant analysis.

Why do analysis populations and missing outcomes matter?

The primary endpoint description says that participants without six-month viral-load data were treated as failures, meaning they were classified as not below 400 copies/mL. Such a rule retains participants in the endpoint count under a conservative classification for achieving the threshold, but it does not reveal their actual viral loads. A different rule could produce a different estimate. The reasons for missing data, the analysis population definition, and any sensitivity analyses are not reported in the registry record.

More generally, missingness can affect a comparison if the probability of missing data is related to treatment group or outcome status. Treating every missing value as failure is an assumption about how those values enter the endpoint; it is not a measurement of the missing outcomes. A sensitivity analysis can examine how conclusions vary under alternative assumptions, but such analyses are not reported in the registry record.

Why distinguish randomized and exposure-based comparisons?

Randomization aims to make groups comparable at the start of a trial, on average, so that an appropriately conducted intention-to-treat comparison can support a causal interpretation. A comparison defined after randomization by the number of injections received does not preserve that same protection automatically. Participants who received more injections may differ from those who received fewer for reasons related to adherence, opportunity, health, or other factors. The registry does not report an adjusted causal analysis for the injection-based secondary comparison. It should therefore be read as an association in the posted groups, not as an equivalent randomized treatment contrast.

Finally, the analyses are unadjusted for multiplicity unless otherwise stated. When many outcomes or comparisons are examined, the chance of observing an apparently notable result by chance can increase. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting secondary outcomes. That context is important when comparing isolated p-values across a broad outcome set.

11. Confidence Intervals and Precision

The recalculated primary risk difference is 0.089, and its 95% confidence interval extends from -0.128 to 0.306. The interval crosses zero, so the posted summary does not locate the underlying difference exclusively on one side of no difference. Its width also signals limited precision: a range of positive and negative contrasts remains compatible with the data and the calculation’s assumptions.

The primary groups contain 66 and 27 participants, respectively. Unequal group sizes and modest counts can leave an estimate uncertain, especially for a binary outcome. The confidence interval quantifies uncertainty in the group contrast; it does not account for every possible source of uncertainty, such as unreported design details, missing-data assumptions, or departures from the statistical model.

The secondary viral-suppression comparison at the lower threshold likewise has a confidence interval that includes zero. The ART adherence interval also includes zero. By contrast, the injection-based opioid-abstinence comparison has an interval that remains above zero, but its exposure-defined groups do not constitute the same direct randomized comparison. Precision and study design must be considered together: an interval can be relatively separated from zero while still applying to a comparison with different interpretive limitations.

Confidence intervals are not guarantees about the truth of a particular endpoint value, and they do not measure the probability that a treatment effect lies inside the displayed range. They are best read as an estimate plus a range of values compatible with the data under the method. Here, the primary estimate has substantial uncertainty, so the point estimate alone should not be treated as a precise description of the group difference.

12. 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 figures are illustrations based on the observed effect, not the trial’s original design assumptions. For these binary outcome planning illustrations, the reported quantities are participants per arm. The stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each scenario uses the stated fraction of the observed effect as the effect to detect. The trial’s own original sample-size justification and design assumptions are not reported in the registry record.

The 80% power column describes the number of participants per arm estimated to give an 80% chance of detecting the assumed effect under the planning assumptions, if that effect were the true one. The 90% power column gives the corresponding planning target for a 90% chance. These are design probabilities under a specified model and effect assumption, not probabilities that the result is true. They are not observed power.

Each row changes the assumed effect size. Planning for 75% of the observed effect or 50% of the observed effect means targeting an effect smaller than the observed estimate. Smaller effects are harder to distinguish from random variation, so they require more participants per arm for the same power target. The displayed participant counts are planning illustrations and should not be confused with the number enrolled or the number contributing to the primary outcome.

The stated rule distinguishes binary outcomes from time-to-event outcomes: the binary-outcome figures are participants per arm, while time-to-event figures would be total events. The planning illustrations attached to the posted binary comparisons therefore refer to participants per arm. They do not show that the completed trial had that level of power, and they do not replace the trial’s unreported original sample-size rationale.

13. 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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.593; treatment proportion p₁ = 0.682; average p̄ = 0.637.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6800 = 1.3327.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6770 = 0.5698.
4. n = (1.3327 + 0.5698)² / (0.0892)² = 454.6, rounded up to 455 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
p0 <- 0.592592592592593                          # control proportion
p1 <- 0.681818181818182                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1246401)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 455 (80% power), 609 (90% power)
 75% of observed effect: n per arm = 820 (80% power), 1,097 (90% power)
 50% of observed effect: n per arm = 1,867 (80% power), 2,499 (90% power)
power.prop.test: 454.6 
Simulated power at n = 455 per arm: 0.804 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry-reported primary analysis lists a Welch’s T Test and a p-value of 0.431 for the comparison between the extended-release naltrexone and placebo groups. The recalculated unadjusted comparison from the posted summary counts gives a risk difference of 0.089, a 95% confidence interval from -0.128 to 0.306, and a p-value of 0.411. These are not identical results: the registry entry reports a test and p-value, while the recalculation expresses the contrast as a risk difference with an interval based on the posted counts.

The two p-values are close, but they are outputs of different analyses and should not be merged or substituted for one another. The registry record does not list the Welch test’s parameters, test statistic, or additional details that would allow the calculation to be reproduced from the entry. The recalculation is a simple unadjusted analysis of summary statistics and does not recreate an analysis of individual participant data.

The confidence interval accompanies the recalculated estimate and helps convey the uncertainty and direction of the primary contrast. The sponsor’s reported p-value does not supply that same absolute-effect range. The registry-reported result remains the sponsor’s own analysis; the recalculated result is a separate descriptive check based on the posted summaries, not a replacement for the trial’s original analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcome measures and analyses, including a primary viral-load endpoint and several secondary outcomes spanning HIV laboratory measures, care, opioid-related measures, risk behaviors, and reincarceration. Testing or estimating many outcomes raises a multiplicity issue: when many comparisons are made, some may appear noteworthy by chance even if there is no underlying difference for those outcomes. The registry record does not report a multiplicity adjustment, a prespecified hierarchy among outcomes, or which secondary analyses were intended as confirmatory.

The primary endpoint description gives a specific missing-data rule. Participants without a six-month viral-load measurement were counted as failures, meaning as having a viral load of 400 or more. This identifies how those missing primary observations entered the posted count, but does not establish why measurements were missing or whether missingness differed by group. The registry does not report sensitivity analyses under alternative missing-data assumptions.

The analysis population is not reported in enough detail to determine whether all randomized participants were included in every outcome analysis or whether a different population definition was used. The posted primary endpoint counts include 66 participants in the extended-release naltrexone group and 27 in the placebo group, while the overall trial enrollment is 151. The record does not explain the full flow from enrollment to each outcome denominator.

For the primary endpoint, missing values being counted as failures is a defined analytic choice, not evidence that the missing participants actually had viral loads at or above the threshold. Other outcomes may have different missingness and denominators, but those details are not reported in the registry record. Without a fuller account of participant flow, missingness, analysis populations, and multiplicity handling, the posted results should be interpreted as the recorded outcome comparisons rather than a complete account of every planned analysis.

16. Safety Results

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

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 among 66 participants at risk in the extended-release naltrexone group and zero among 27 participants at risk in the placebo group. The serious adverse-event terms list is empty. These entries describe what was posted for the reported risk groups; they do not establish that no unreported events occurred, nor do they rule out uncommon harms.

Other adverse events are listed for 13 participants in the extended-release naltrexone group and 4 in the placebo group. The registry does not provide event types or severity details in the adverse-event summary included in the record. Death counts are not reported in the registry record. A blank or unreported death value should not be read as a count of zero.

Adverse-event counts are not a stand-alone comparison of safety. Group sizes differ, and the summary does not give event-level detail, timing, duration, or attribution. The figures should be read as the registry’s posted counts within the stated at-risk groups, not as proof of an absence of risk or as a basis for individual treatment decisions.

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

The primary result describes a six-month comparison of the proportion meeting a specific viral-load threshold. The posted counts are 45 of 66 in the extended-release naltrexone group and 16 of 27 in the placebo group. The recalculated risk difference is positive, but its 95% confidence interval ranges from -0.128 to 0.306 and includes zero. The data therefore leave substantial uncertainty about the size and direction of the underlying difference in this endpoint.

This result does not establish what would happen for every person in the trial population or for people outside the registry’s described population. It does not answer every question about HIV care, opioid dependence, or transition from incarceration. It is a comparison of the posted six-month endpoint under the registry’s stated rule for participants with missing viral-load data.

The endpoint is a threshold measure, not a complete description of viral-load trajectories. The study description refers to laboratory monitoring over one year, but the primary outcome is based on the first six months. The posted result does not by itself describe longer-term changes, CD4 counts, retention in care, relapse timing, addiction severity, craving, risk behavior, or reincarceration. Results for many of those outcomes are not reported in the registry record.

Secondary findings should be interpreted in light of their different outcomes, comparison groups, and the absence of a reported multiplicity plan. In particular, a comparison based on the number of injections received does not have the same interpretation as the randomized primary contrast. The result also should not be used as personalized medical advice. Its meaning is limited to statistical description of the groups, endpoint, follow-up time, and analyses recorded in the trial registry.

18. Why This Trial Matters Statistically

This trial raises statistical questions that are common in clinical research involving participants transitioning between settings. A randomized placebo comparison can support a direct group contrast, but the usefulness of that contrast depends on outcome definitions, follow-up, missing-data handling, and clear reporting of the analysis population. The primary endpoint is a binary classification at a specified viral-load threshold, which makes the result interpretable as a difference in proportions but does not capture the full range of viral-load change.

The primary comparison illustrates why effect estimates and uncertainty should be read together. The risk-difference point estimate is positive, yet the confidence interval is broad and crosses zero. Reporting the point estimate alone would obscure the range of contrasts compatible with the data. Conversely, a p-value alone would not communicate the absolute scale of the estimate or its uncertainty.

The record also illustrates the distinction between the randomized treatment comparison and a secondary comparison defined by treatment exposure. Participants receiving more injections may not be comparable to those receiving fewer injections for reasons that arise after randomization. That distinction matters when interpreting an apparent difference on an opioid-related outcome.

Finally, the broad set of trial hypotheses makes transparent reporting especially important. Outcomes span laboratory measures, care, opioid use, and community outcomes. Without a reported multiplicity strategy, analysis populations, and results for all planned outcomes, readers have limited ability to distinguish confirmatory evidence from exploratory patterns. The registry entry is informative about the posted results, but does not supply all the detail needed to reconstruct the full analysis strategy.

19. 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 endpoint analysis has smaller outcome denominators than the overall enrollment, and the registry does not explain the complete participant flow into those denominators. The primary endpoint description does state that 13 participants without six-month viral-load data were counted as failures. Reasons for missing measurements and sensitivity analyses using other assumptions are not reported in the registry record.

The primary endpoint measures whether viral load was below 400 copies/mL at one time point. It does not provide a complete account of changes in HIV-1 RNA across follow-up. The study description mentions laboratory testing every three months for one year, but the posted primary result is for six months. The registry does not report the posted results for every outcome named in the study description.

Although the trial is described as randomized and quadruple-masked, details about the randomization procedure, allocation concealment, participant flow, adherence, and masking assessment are not reported in the registry record. The original sample-size justification and the trial’s own design assumptions are also not reported. The sample-size illustrations based on observed effects cannot substitute for those original planning details.

The reported primary test is labeled Welch’s T Test, while the primary outcome is presented as counts of participants. The registry does not provide enough test details to reproduce the sponsor-reported analysis from the record. The unadjusted recalculation is based on the posted summary counts and is not an individual-participant reanalysis.

Several secondary outcomes involve multiple comparisons, with no multiplicity adjustment reported. The opioid-abstinence comparison based on injection exposure is not the same as an intention-to-treat randomized contrast. These design and reporting limitations constrain how strongly secondary patterns can be interpreted, even when a posted interval is separated from zero.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Naltrexone for Opioid Dependent Released Human Immunodeficiency Virus Positive (HIV+) Criminal Justice Populations — NCT01246401.
• Registry record last updated 2020-03-19; status last verified 2017-06.
• Results first posted 2017-07-17.
• 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.

23. Record Summary

The registry record describes a completed randomized, placebo-controlled, quadruple-masked, parallel-group phase 1 and phase 2 trial of extended-release naltrexone for HIV-positive participants with opioid dependence transitioning from prison to the community. Enrollment was 151. Monthly intramuscular extended-release naltrexone or placebo was planned for six months, beginning before release.

The primary endpoint was the number of participants with HIV-1 RNA below 400 copies/mL at six months. The posted counts were 45 of 66 in the extended-release naltrexone group and 16 of 27 in the placebo group. The recalculated risk difference was 0.089, with a 95% confidence interval from -0.128 to 0.306. The registry-reported primary analysis lists a Welch’s T Test with p = 0.431.

Selected secondary outcomes and adverse-event counts are also posted. Missing-data handling is specified for the primary endpoint, but reasons for missingness, the full analysis populations, multiplicity handling, the original sample-size justification, and several other planned results are not reported in the registry record.