The Long-Term Treatment of Drug Addiction and Unemployment: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Long-Term Treatment of Drug Addiction and Unemployment (NCT02487745), a randomized na study. The primary outcome is The Percentage of Opiate- and Cocaine-negative Urine Samples.

2. Clinical Question

The trial addressed whether adding wage supplements contingent on abstinence to supported employment was associated with a different primary outcome than supported employment alone. The intervention and comparator were both behavioral approaches; the registry describes the wage supplements as part of a model intended to arrange long-term exposure to employment-based abstinence reinforcement.

The primary endpoint was defined as the number of opiate- and cocaine-negative urine samples divided by the possible number of samples, multiplied by 100. The reported group values were 44.9 for IPS Only and 64.6 for IPS Plus Wage Supplement. These are group-level percentages, not a count of participants who remained abstinent. The registry specifies an intent-to-treat population for this endpoint, but details about how missing urine samples were handled are not reported in the registry record.

The record also identifies economic benefits and several behavioral and employment outcomes as interests of the trial. Those aims should be distinguished from the primary endpoint: the primary result concerns urine samples, and the posted information does not establish an economic effect. The question is comparative, but interpreting the comparison requires attention to the reported logistic-regression analysis, the absence of a confidence interval, and the limited detail about the statistical plan.

3. Trial Design

This was a randomized, parallel-group trial with no masking recorded. One group received IPS supported employment alone; the other received IPS supported employment plus abstinence-contingent wage supplements. The registry lists 91 participants enrolled: 47 in IPS Only and 44 in IPS Plus Wage Supplement. The record describes the allocation as randomized, but the method used to generate or conceal the allocation sequence is not reported in the registry record.

The study is marked completed. Its start is recorded as 2015-11 and its primary completion as 2019-09. The phase is listed as NA. The sponsor is Johns Hopkins University. The registry describes the purpose as evaluating the effectiveness and economic benefits of a wage supplement model intended to arrange long-term exposure to employment-based abstinence reinforcement. That statement describes the study aim, not a finding about effectiveness or economic benefit.

The primary endpoint was the percentage of opiate- and cocaine-negative urine samples over 1 year, in the intent-to-treat population. The record lists 12 outcome measures and reports 9 secondary outcomes. It also records 2 analyses. The complete analysis plan, including any planned adjustments or hierarchy among outcomes, is not reported in the registry record.

Randomization supports a between-group comparison, but the lack of masking is relevant context for a behavioral intervention and outcomes involving employment or self-report. Urine results are also subject to the collection schedule and handling of missing samples; those operational details are not reported in the registry record. The record does not state a sample-size justification, the intended enrollment target’s basis, or whether interim analyses were planned. The enrollment count alone cannot establish what level of precision the study was designed to achieve.

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: Participants were enrolled in a 3-month abstinence initiation phase (Phase 1) prior to random assignment.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat

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 The Percentage of Opiate- and Cocaine-negative Urine Samples, measured over 1 year.

The primary endpoint is the percentage of urine samples that were negative for both opiates and cocaine during 1 year. The registry defines it as the number of samples negative for both substances divided by the possible number of samples, multiplied by 100. The unit is therefore urine samples expressed as a percentage, rather than participants classified as abstinent or not abstinent.

This measure captures repeated biological test results across the assessment period. In general, using a proportion of samples can describe how often the collected samples met the specified negative-result criterion. It does not, by itself, describe the timing or continuity of negative results, nor does it show how many individual participants had a particular pattern. Those interpretations would require participant-level or more detailed longitudinal results, which are not reported in the registry record.

The registry identifies the analysis population as intent-to-treat. In general, an intent-to-treat approach aims to analyze participants according to their randomized group, regardless of intervention adherence. How missing or unavailable samples were incorporated into the denominator is not reported in the registry record. That detail matters because a percentage depends on both the number of qualifying samples and the number treated as possible samples.

The reported group values are 44.9 for IPS Only and 64.6 for IPS Plus Wage Supplement. The registry does not provide an arm-level dispersion value or a confidence interval for the group percentages. The endpoint is relevant to the stated abstinence-reinforcement aim, but it is only one of the trial’s outcomes and does not, on its own, answer the broader question about economic benefits.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Comparative clinical-trial results commonly include descriptive summaries and an inferential analysis. The descriptive summaries report what was observed in each group. The inferential analysis uses a statistical model to estimate a comparison and assess its compatibility with a specified null hypothesis. These are related but distinct kinds of information, and they need not use the same scale.

What does the endpoint percentage describe?

The primary outcome is a percentage of urine samples negative for opiates and cocaine. In general, a percentage is a ratio: qualifying observations form the numerator, while the observations counted as possible form the denominator. The registry defines the endpoint using the number of negative samples and the possible number of samples. The resulting group means of 44.9 and 64.6 summarize the endpoint for the two arms, but do not reveal the distribution of individual participants’ values.

Repeated samples from one participant are not necessarily statistically independent. In general, samples from the same person may be more alike than samples from different people. A method that treats every sample as independent could misstate uncertainty if this dependence is not accounted for. The registry names logistic regression but does not report whether the model accounted for repeated observations, what observation was modeled, or what correlation structure was used. The posted record therefore does not permit a detailed assessment of this issue.

What does an odds ratio mean?

Odds compare the probability of an event with the probability of its alternative. An odds ratio compares those odds between groups. It is not the same quantity as a difference in group percentages, and it is not generally appropriate to interpret it as a direct percentage-point increase. The reported odds ratio is 2.29. However, the registry does not give the outcome coding or fuller model specification, so the direction and precise interpretation of that odds ratio cannot be fully reconstructed from the record alone.

Logistic regression is commonly used when the modeled response is binary. The registry’s endpoint is expressed as a percentage of samples, but it does not state whether the regression operated on sample-level results, participant-level summaries, or another representation. This distinction matters because the unit of analysis affects both the meaning of the odds ratio and the uncertainty estimate. That model detail is not reported in the registry record.

How should the p-value be read?

The reported p-value is .05. In general, a p-value measures how incompatible the observed data, or more extreme data, would be with a specified null model under the assumptions of the analysis. It is not the probability that the null hypothesis is correct, and it does not measure the size or importance of an effect. Interpretation depends on the model, the analysis plan, and whether other outcomes or analyses were considered.

The registry classifies the reported analysis as superiority and lists a 95% confidence level, but no confidence limits are reported. A confidence interval is generally a range of values generated by a method designed to have a stated long-run coverage under its assumptions. It helps convey both the estimated effect and its precision. Without the limits, the reported p-value and odds ratio provide only a partial picture of uncertainty.

Why does over-dispersion matter?

In general, over-dispersion occurs when observed variability is greater than a statistical model’s assumptions allow. For binary or count-like observations, extra variability can arise when observations are clustered, when participants differ substantially, or when repeated observations within a participant are correlated. If a model underestimates variability, its standard errors may be too small and its inferential results too optimistic. Appropriate modeling can account for such structure, but the registry does not report enough detail to determine whether over-dispersion was assessed or how within-participant dependence was handled in this trial.

What do the posted summaries leave unresolved?

The arm summaries report the average percentage and group size, but no arm-level dispersion estimate is posted. The registry identifies full range as the dispersion measure, yet the arm results do not give a dispersion value. Without information about how individual participants’ percentages are distributed, it is not possible to judge whether the group means reflect broadly similar patterns or a more uneven distribution.

The intent-to-treat label indicates the stated analysis population, but it does not by itself describe missing-data handling. In general, missing outcomes can affect both the denominator and the composition of an analysis. The registry does not report how unavailable urine samples or participant withdrawals were treated, whether imputation was used, or whether sensitivity analyses were conducted. These details are especially relevant when the endpoint is built from repeated samples over time.

Finally, the record does not provide a detailed statistical analysis plan, covariate list, model diagnostics, or interim-analysis information. General statistical principles can explain why those details matter, but they cannot establish what the investigators did. The reported logistic-regression result should therefore be interpreted as the model-based result recorded in the registry, with its limitations in specification and precision kept in view.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s comparison as a logistic regression with an odds ratio of 2.29 and p = .05. It classifies the analysis as superiority and lists a 95% confidence level, but the confidence interval limits are not reported in the registry record. The analysis is distinct from the arm-level mean percentages of 44.9 and 64.6: those summaries are descriptive, while the odds ratio is a model-based measure.

A simple unadjusted re-calculation from posted summary statistics cannot reproduce the sponsor’s logistic-regression estimate. The registry does not provide the individual observations, model coding, or full model specification needed to do so. No separate recalculated comparative estimate is reported in the registry record. Since the record does not identify a discrepancy between the sponsor’s analysis and a recalculation, there is no basis to replace the reported analysis with another estimate.

The reported odds ratio should be interpreted with the limits of the record in mind. The regression’s covariates, handling of repeated samples, and missing-data methods are not reported. The group percentages provide useful descriptive context but cannot supply the missing model details or a confidence interval. The sponsor’s analysis remains the posted inferential result, and its precision cannot be fully judged from the available information.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 12 outcome measures, reports 9 secondary outcomes, and records 2 analyses. The primary outcome is identified, but a complete hierarchy of primary and secondary hypotheses, or a plan for controlling the chance of false-positive findings across analyses, is not reported in the registry record. A multiplicity adjustment is also not reported.

In general, examining multiple outcomes or conducting multiple analyses can increase the chance that at least one result appears statistically notable by chance. Whether an adjustment is appropriate depends on the trial’s prespecified goals and analysis plan. Without that plan, the p-value of .05 cannot be interpreted in the context of the full family of intended comparisons. The record does not establish that the reported primary result was adjusted for other analyses, nor that it was not; the adjustment approach is simply not reported.

The primary endpoint is labeled intent-to-treat. In general, intent-to-treat analysis aims to preserve the randomized-group comparison by retaining participants in their assigned groups. The label does not specify how missing outcomes were handled. The registry does not report the number of missing urine samples, the rules for defining possible samples, any imputation procedure, or sensitivity analyses examining alternative missing-data assumptions.

Missingness can matter for an endpoint expressed as a proportion of samples. If some planned observations are unavailable, the numerator and denominator may depend on the handling rule. Different rules can affect both the group summaries and their uncertainty. The record does not say whether participants with incomplete follow-up contributed all available samples, were handled using another method, or were excluded from particular calculations.

The registry also does not report the statistical analysis plan, details about interim analyses, or a sample-size justification. These omissions prevent assessment of whether the analysis and outcome emphasis followed prespecified decisions. They do not demonstrate that the analysis was inappropriate; they mean that the record does not contain the detail needed to evaluate these aspects independently.

12. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: death (11.4%), leg amputated (2.1%), kidney failure (2.1%), stroke (2.1%), upper respiratory problem (2.1%).

The registry reports serious adverse events in 8 of 47 participants in IPS Only and 5 of 44 participants in IPS Plus Wage Supplement. Other adverse events are reported in 11 of 47 and 8 of 44 participants, respectively. Deaths are listed for 4 participants in IPS Only and 5 in IPS Plus Wage Supplement. These are counts within the recorded groups and should be read as descriptive safety information.

The serious-event terms listed include Death at 11.4%, followed by leg amputated, kidney failure, stroke, upper respiratory problem, and abdominal pain, each at 2.1%. The registry record does not provide enough context to infer that an event was caused by either intervention. Event counts do not, by themselves, establish comparative risk, especially without detail on event timing, recurrence, severity definitions, or adjudication.

These figures should not be treated as a complete account of safety or as a direct measure of treatment effect. The registry does not report a statistical comparison for the adverse-event counts in the information recorded here, and the counts alone do not establish whether groups differed beyond chance. They also do not provide individual-level clinical context. The safety information is therefore best understood as the events recorded for each group, not as evidence of causation.

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

The posted results describe a randomized comparison of IPS supported employment alone with IPS plus abstinence-contingent wage supplements. For the primary endpoint, the reported mean percentage of opiate- and cocaine-negative urine samples over 1 year was 44.9 in IPS Only and 64.6 in IPS Plus Wage Supplement. The sponsor’s logistic-regression analysis reports an odds ratio of 2.29 and p = .05, with a superiority classification.

These results answer a specific statistical question about the endpoint and the analysis recorded in the registry. They do not establish the size of a difference in percentage points, because the odds ratio is on a different scale and the record does not provide a confidence interval. They also do not establish how much individual participants’ results varied, since arm-level dispersion is not reported. The reported means should not be read as the percentage of participants who were abstinent throughout the year.

The result does not, on its own, demonstrate an economic benefit. Although economic benefits are part of the stated trial purpose, the primary endpoint reported is the proportion of urine samples negative for two substances. The posted information does not report an economic-effect estimate. Nor does the primary endpoint resolve the trial’s other questions about employment, income, or self-reported behavior.

Interpretation is limited by missing details about the model, the handling of repeated and missing samples, and the role of the other outcomes and analyses. The reported p-value is not a measure of clinical importance or certainty that a treatment effect exists. Without confidence limits and fuller methods, the precision and robustness of the comparative estimate cannot be fully assessed.

The trial record therefore supports a careful description of the observed group summaries and the sponsor-reported analysis, while leaving important questions about effect precision, modeling, missingness, multiplicity, and broader outcomes unresolved. It does not support a broader claim beyond the outcomes and analysis recorded.

14. Why This Trial Matters Statistically

This trial illustrates why endpoint definition matters. Its primary outcome is a proportion of repeated urine samples, not a participant-level classification. That distinction changes the unit being summarized and the interpretation of the reported means. A reader who treats the result as the percentage of participants abstinent would be answering a different question from the one specified in the registry.

It also illustrates the difference between descriptive and model-based results. The group means are 44.9 and 64.6, while the reported logistic-regression comparison is an odds ratio of 2.29. Those figures are not interchangeable: the odds ratio is not a direct difference between group percentages. A statistical report is easier to evaluate when its model specification and uncertainty estimates are available alongside descriptive summaries.

The primary result also shows why p-values should not be read in isolation. The recorded p-value is .05, but the confidence interval is not reported. Without an interval, the registry offers limited information about the precision of the model-based estimate. The number of outcome measures and analyses also makes the unreported multiplicity plan relevant to interpretation.

Finally, repeated samples and missing observations raise questions about dependence and denominator rules. General methods can address within-person correlation and incomplete follow-up, but the registry does not identify the methods used. The case demonstrates that a named endpoint and a reported model are not always enough to reproduce or fully assess an analysis; details about observations, missingness, and model assumptions are also important.

15. 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.

Several trial-specific details needed to evaluate the primary comparison are not reported in the registry record. The model’s outcome coding, covariates, handling of repeated samples, and diagnostic checks are not specified. The record also does not state how missing urine samples were treated or how the possible number of samples was determined for each participant. These details affect interpretation of both the mean percentages and the logistic-regression estimate.

The odds ratio is reported without confidence limits, limiting assessment of precision. No arm-level dispersion values are given for the primary endpoint, so the spread of participant-level percentages cannot be judged from the posted summaries. The primary endpoint is an intent-to-treat outcome, but the record does not describe whether all randomized participants contributed data or how incomplete follow-up affected the analysis.

The trial was randomized and parallel-group, but masking was recorded as none. The record does not describe allocation concealment or the sequence-generation method. The study’s stated interest in economic benefits is not accompanied by an economic result among the posted primary endpoint information. The record also does not provide a statistical analysis plan, sample-size justification, or details on interim analyses.

The registry reports 12 outcome measures, 9 secondary outcomes, and 2 analyses, but does not report a multiplicity strategy. This leaves uncertainty about how the primary result fits into the full planned analysis. The safety counts are descriptive and lack the timing and adjudication information needed to attribute events or make a detailed between-group safety comparison. These are limitations of the information recorded, not evidence that unreported methods or analyses were absent from the trial itself.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: The Long-Term Treatment of Drug Addiction and Unemployment — NCT02487745.
• Registry record last updated 2022-08-11; status last verified 2022-07.
• Results first posted 2022-08-11.
• 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.

19. Record Summary

NCT02487745 is a completed, randomized, parallel-group trial sponsored by Johns Hopkins University. The registry records 91 participants enrolled, with 47 assigned to IPS Only and 44 to IPS Plus Wage Supplement. Masking is listed as none. The recorded condition is intravenous substance abuse.

The primary endpoint was the percentage of opiate- and cocaine-negative urine samples over 1 year in the intent-to-treat population. The reported group means were 44.9 for IPS Only and 64.6 for IPS Plus Wage Supplement. The registry reports a logistic-regression odds ratio of 2.29 and p = .05; confidence limits are not reported.

The record lists 12 outcome measures, reports 9 secondary outcomes, and records 2 analyses. It does not report the full statistical plan, missing-data methods, or a multiplicity approach. The primary result is therefore best read as the posted endpoint summaries and sponsor-reported model comparison, with important limits on assessing precision and reproducibility.