Interventions for Unemployed Hazardous Drinkers: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Interventions for Unemployed Hazardous Drinkers (NCT02559609), a randomized na study. The primary outcome is Longest Duration of Alcohol Abstinence (Samples) (Physiological Measure).

2. Clinical Question

The question has two related components: whether reinforcement for alcohol-negative samples, reinforcement for job activities, or their combination is associated with the endpoint; and how the observed endpoint values compare across the assigned conditions. The factorial arrangement is intended to examine these components in combination, but the posted primary analysis does not report separate main effects or an interaction estimate. The primary endpoint is based on submitted breath tests, not directly on days of alcohol abstinence. Employment and longest duration of alcohol abstinence in days are listed as secondary outcomes, but their results are not described in the primary endpoint record.

3. Trial Design

The registry identifies this as a randomized, non-masked, factorial study. The brief summary describes a 2 x 2 design with four conditions: standard care, reinforcement for negative breathalyzer samples, reinforcement for job-seeking activities, and reinforcement for both. The arm descriptions provide more detail: all conditions included standard services, job activity contracting, and alcohol monitoring, with the reinforcement components varying by condition.

The interventions are behavioral. The registry lists contingency management for alcohol abstinence, contingency management for completing job-related activities, and job-related activity contracting. The brief summary describes a partnership with CT United Labor Agency and a remote breath alcohol monitoring procedure. It states that study interventions were in effect for three months and that participants were followed for one year. The primary endpoint timeframe is 12 weeks.

The registry enrollment field lists 240. Separately, the brief summary says N = 280 would be randomly assigned. The record does not reconcile these figures or explain which reflects the final enrolled population. The primary endpoint description says 234 were analyzed and identifies six people with no endpoint data. These are distinct registry statements and should not be treated as interchangeable measures of enrollment, allocation, or analysis.

The registry marks the phase as NA and the study status as completed. The recorded start date is 2016-04-19, and the primary completion date is 2022-06-01. The sponsor is UConn Health. Details such as sequence generation, allocation concealment, stratification, and the precise schedule of study contacts 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: N=234 analyzed. n=6 with no data: n=1 Group B withdrawal; n=5 had no samples submitted (n=1 Group A, n=1 Group B, n=1 group D) or none submitted on-time (n=2 Group B).

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 Longest Duration of Alcohol Abstinence (Samples) (Physiological Measure), measured over 12 weeks.

The primary endpoint is the longest duration of alcohol abstinence measured through breath-test samples. More precisely, it counts the greatest number of consecutive breath tests that were submitted on time—within one hour of the prompt—and read alcohol negative, defined in the record as BAL < 0.02. The unit is the number of alcohol-negative breath tests, and the registry reports the median and inter-quartile range for each group over 12 weeks.

This measure captures a sustained sequence of recorded negative samples under the study’s monitoring procedure. It is not the same as the longest duration of abstinence expressed in days: that is listed separately as a secondary outcome. Nor does the primary measure count every day without alcohol use. It depends on participants submitting samples, submitting them on time, and the readings meeting the negative threshold. Missing or late samples can therefore affect whether a consecutive sequence is recorded.

Using a physiological measure provides an objective reading for submitted tests, while the sequence length focuses on continuity rather than a single test. At the same time, the endpoint describes observed monitoring records rather than every aspect of alcohol use. It does not, on its own, establish why groups differed or characterize the full pattern of alcohol consumption between tests.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is summarized with medians and inter-quartile ranges. The median is a measure of the center of a distribution that is less sensitive to extreme observations than the arithmetic mean. The inter-quartile range describes the spread of the middle portion of observations. These summaries are a natural way to describe a count-like endpoint when values may be unevenly distributed, although the registry does not report the full distribution or explain the reason for choosing these summaries.

What does the reported rank test assess?

The registry names the Wilcoxon (Mann-Whitney) method. In general, a rank-based test compares the ordering of observed values rather than relying on the raw values having a particular normal distribution. Such tests can be useful when data are skewed or contain outlying observations. A rank test is not automatically a test of medians alone: its interpretation depends on the distributions being compared and the design of the analysis.

There is an important reporting ambiguity. The registry lists all four groups for the analysis but names a method commonly used for a two-group comparison. It does not state whether the test was part of a broader procedure, which groups were contrasted, or how the four-group result was formed. The posted p-value should therefore be reported as the registry reports it, without assigning it to a particular pair of groups or to a specific factorial effect.

What does a p-value communicate?

In general, a p-value describes how unusual data at least as incompatible with a null model as the observed data would be, if that model and the test assumptions held. It is not an effect size, a measure of clinical importance, or the probability that a result is due to chance. The registry reports p = .028 for the primary analysis, but does not report a confidence interval or a corresponding contrast estimate. Without those estimates, the size and precision of a particular difference cannot be judged from the primary analysis alone.

Why does the factorial structure matter?

In general, a factorial design can distinguish the contribution of separate intervention components and whether their combination behaves differently from what would be expected from considering each component separately. That requires analyses that estimate the relevant component effects and their interaction, with enough detail to understand the contrasts. Although this trial is described as factorial, the reported primary analysis does not give separate estimates for reinforcement of negative samples, reinforcement of job activities, or an interaction between them. The four group summaries do not substitute for those estimates.

The record also identifies the trial as non-masked. In general, lack of masking can influence behavior, reporting, or other parts of a study, although the endpoint uses recorded breath tests. The registry does not report whether outcome processing or analysis was masked, so no conclusion about those procedures is warranted.

How should planning calculations be read?

In general, a sample-size planning calculation asks what number of participants or events would be needed under stated assumptions to detect a chosen effect at a specified error threshold and power. Such a calculation is not the same as the power achieved by a completed trial. The registry’s planning assumptions describe two-sided significance level 0.05, equal allocation between two arms, a normal approximation, and effects defined as a stated fraction of the observed effect. They are planning illustrations based on the observed effect, not the trial’s original design justification. For time-to-event calculations, the stated convention is to give total events; other calculations are expressed as participants per arm. No additional claim about the study’s planned or achieved power follows from those assumptions.

Interpretation of the primary result also depends on the analysis population and endpoint availability. The registry says 234 were analyzed and identifies six with no data. It does not describe an imputation method, a sensitivity analysis, or a detailed rule for including participants with partially submitted or late samples. These details matter because the endpoint requires a sequence of on-time negative tests. The registry record does not report the full statistical analysis plan, the handling of multiple outcomes, or any interim analysis procedures.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a Wilcoxon (Mann-Whitney) result across the four listed groups, with p = .028. The group summaries are medians and inter-quartile ranges. The record does not provide a corresponding confidence interval or an effect estimate for a defined contrast, so there is no precise sponsor-reported difference estimate to compare with a recalculated one.

No discrepancy is flagged between a recalculation and the sponsor’s analysis in the registry record. However, the record does not give a separate recalculated estimate or enough detail to reconstruct a clearly specified four-group contrast. The sponsor’s reported analysis therefore remains the inferential result to cite. It should not be expanded into a claim about a particular pair of groups or a factorial component effect that the record does not report.

The analysis label deserves care: it names a Wilcoxon (Mann-Whitney) method while listing four groups, without explaining the exact contrast or test structure. That uncertainty affects interpretation, not the accuracy of the reported p-value as a registry entry. A p-value alone also does not show the magnitude or precision of an effect. The medians and inter-quartile ranges describe the observed group values, but do not resolve which differences the reported test represents.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures and six analyses. The primary endpoint is the longest duration of alcohol abstinence measured by breath-test samples. Two secondary outcomes are identified: percent of days worked for pay over an 84-day intervention-period denominator, and longest duration of alcohol abstinence in days measured physiologically. The record does not report a multiplicity adjustment, a hierarchy for interpreting outcomes, or a prespecified family of tests. Without those details, the chance of at least one apparently unusual result across multiple analyses cannot be assessed from the registry record.

The primary endpoint description states that 234 participants were analyzed. It identifies six with no data: one Group B withdrawal; and five with no submitted samples or no samples submitted on time. The description further says that no samples were submitted by one person in Group A, one in Group B, and one in Group D, while two in Group B had none submitted on time. These categories are reported as written; the registry does not give a fuller account of endpoint completeness or explain how partially observed sequences were handled.

The registry does not report whether missing endpoint values were imputed, whether an analysis was conducted under a missing-at-random assumption, or whether sensitivity analyses explored other assumptions. It also does not provide the number of randomized participants by group for the primary analysis beyond the arm summaries, or explain how the analyzed population was defined. The difference between the enrollment field of 240 and the brief summary’s N = 280 is not reconciled in the record.

These reporting gaps limit assessment of robustness. Missing tests are particularly relevant because a run requires consecutive on-time negative readings: an absent or late test may interrupt the recorded sequence even when the person’s alcohol use is unknown at that time. The posted endpoint therefore reflects both readings and submission patterns. No further conclusion about the missing data or multiplicity strategy is justified by the registry record.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: inpatient substance use treatment (9.0%), inpatient medical treatment (6.9%), inpatient psychiatric treatment (3.4%).

The registry reports serious adverse events by group as affected participants among those listed as at risk: Group A, 6 among 50; Group B, 8 among 65; Group C, 9 among 58; and Group D, 9 among 67. Other adverse events are reported as 5 among 50, 5 among 65, 1 among 58, and 6 among 67, respectively. Deaths are reported as 0 in Groups A, B, and D, and 1 in Group C. These are descriptive counts; they do not establish that an event was caused by an intervention.

The serious adverse event terms with the largest reported percentages are inpatient substance use treatment at 9.0%, inpatient medical treatment at 6.9%, and inpatient psychiatric treatment at 3.4%. The registry does not provide a group breakdown for these terms in the information, nor does it state whether a participant could contribute to more than one term. The terms and percentages should therefore not be treated as mutually exclusive categories or as a complete account of every event.

Comparisons of safety counts require caution. The groups have different numbers listed as at risk, and the registry does not provide event timing, exposure-adjusted rates, or a statistical comparison. Small counts can also vary substantially without providing a stable estimate of risk. The figures describe what the registry reports during the study; they do not, on their own, quantify comparative safety or establish causality.

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

The primary result describes the longest recorded sequence of on-time negative breath tests during the 12-week intervention period. The observed median was 6.0 in the standard job activity contracting and alcohol monitoring group, 11.0 in the group with reinforcement for negative alcohol samples, 7.0 in the group with reinforcement for completing job activities, and 7.5 in the group with reinforcement for both. The inter-quartile ranges show broad variation within every group.

The registry reports p = .028 for an analysis listing all four groups. This indicates that the registry’s reported analysis found evidence of variation under its statistical test, but the record does not specify a pairwise contrast, provide an effect estimate with a confidence interval, or report separate factorial main effects and interaction. It therefore does not identify which condition differs from which, or how precisely any particular difference is estimated.

The result does not measure days of abstinence, all alcohol use, employment outcomes, or overall health. It is based on breath tests that were submitted on time and read as negative. A sequence measure can be influenced by submission patterns as well as by the readings themselves, and missing or late tests can interrupt a recorded run. The registry identifies six participants with no primary endpoint data but does not describe a missing-data method or sensitivity analysis.

The result also does not establish that a particular intervention caused an observed difference or that the combination of components adds benefit beyond either component alone. Although the trial used a factorial design, the reported primary analysis does not provide the contrasts needed to answer those component-specific questions. The employment and days-based abstinence outcomes are listed as secondary measures, but their results are not part of the primary result.

In short, the posted data describe differences in group summaries and a reported four-group p-value. Their interpretation is constrained by the unspecified test contrast, absent effect precision, missing endpoint data, and the distinction between a run of negative tests and broader measures of alcohol use or employment.

14. Why This Trial Matters Statistically

This trial illustrates why endpoint definition matters. “Longest duration of alcohol abstinence” can refer to different measures, and the primary endpoint is specifically a count of consecutive on-time negative breath tests. A separate secondary outcome measures the longest duration in days. Treating these as interchangeable would change the question being answered.

The study also highlights the interpretive value—and reporting requirements—of a factorial design. Four conditions represent the presence or absence of two reinforcement components, but a useful factorial interpretation depends on reported component effects and their interaction. The posted primary analysis names four groups but does not provide those estimates or specify the exact contrast. Group summaries alone cannot establish how each component contributed.

There is also a practical statistical distinction between an observed group summary and an inferential result. Medians and inter-quartile ranges describe the recorded data; the reported p-value is attached to a stated analysis method. Without a clear contrast and an effect estimate with uncertainty, the p-value cannot convey the size, direction, and precision of a specific comparison. This is a reason to read statistical reporting in context rather than relying on a single test result.

Finally, the endpoint depends on both test readings and timely submission. The registry identifies participants with missing or untimely samples, which makes data completeness directly relevant to interpretation. The mismatch between the enrollment field and the brief summary’s planned population further demonstrates why registry entries should be read as separate reported details when the record does not reconcile them. These issues are important for understanding what the trial’s statistical result does—and does not—describe.

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 limitations in the registry record constrain interpretation. First, the primary analysis is listed across four groups but described with the Wilcoxon (Mann-Whitney) label, and the record does not explain the test structure or identify the comparison represented by p = .028. No confidence interval or effect estimate for a defined contrast is reported. The result therefore cannot establish the size or precision of a particular group difference.

Second, the endpoint is the longest run of on-time negative breath tests, not a direct count of alcohol-free days. It depends on breath-test submission and timing, and six people are described as having no primary endpoint data. The record does not report imputation, sensitivity analyses, or a fuller account of how incomplete test sequences were handled.

Third, the factorial design is not accompanied in the posted primary result by separate estimates for the two intervention components or their interaction. The record also does not report a multiplicity strategy despite listing six outcome measures and six analyses. The implications of considering multiple outcomes and analyses are therefore unclear.

Finally, the registry enrollment field lists 240, while the brief summary describes N = 280 as the planned randomized population. The record does not reconcile these figures. It also does not report the full statistical analysis plan, allocation procedures, or masking of outcome processing. These are limitations of the information reported in the registry, and should not be filled in by assumption.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Interventions for Unemployed Hazardous Drinkers — NCT02559609.
• Registry record last updated 2024-04-23; status last verified 2024-04.
• Results first posted 2023-06-22.
• 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

NCT02559609 is a completed randomized, non-masked factorial study sponsored by UConn Health. It concerns unemployed people with hazardous drinking and compares standard services with reinforcement for negative breath alcohol samples, job-related activities, or both. The registry enrollment field is 240, while the brief summary describes N = 280 as the planned population; the record does not explain the difference.

The primary endpoint is the longest run of consecutive, on-time negative breath tests over 12 weeks. The registry reports group medians and inter-quartile ranges and a four-group Wilcoxon (Mann-Whitney) analysis with p = .028. It states that 234 were analyzed and identifies six with no endpoint data. A specific contrast, confidence interval, missing-data method, and factorial component estimates are not reported in the registry record.

The recorded study start date is 2016-04-19, and the primary completion date is 2022-06-01. The registry lists percent of days worked for pay and longest duration of alcohol abstinence in days as secondary outcomes.