Evaluation of Diagnostic HIV Ag/Ab Combo Assay: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Evaluation of Diagnostic HIV Ag/Ab Combo Assay (NCT00933933), a non-randomized phase 3 study. The primary outcome is Architect HIV Combo Test Data for Clinical Sensitivity in HIV Positive Specimens.

2. Clinical Question

The central question is whether the ARCHITECT HIV Ag/Ab Combo assay detects HIV markers in specimens known to be positive, and how it behaves in specimens from people described as apparently healthy and at low risk or from individuals at risk for HIV infection. The registry record frames these as separate evaluations rather than as a randomized comparison between treatments or diagnostic strategies.

The primary outcome is therefore a count-based endpoint within a diagnostic evaluation, not a patient-centered outcome or a comparison of clinical treatments. The registry record does not report the number of specimens correctly identified, the number of false-negative results, or an estimate of uncertainty around assay sensitivity. The available description supports identifying the intended evaluation, but not judging the assay’s measured performance.

3. Trial Design

This completed study is registered as Phase 3 and is described as non-randomized, unmasked, and single-group. Those labels are important to the interpretation: participants or specimens were not randomly allocated to competing treatment strategies, and masking is listed as none. The study is better understood as a series of diagnostic specimen evaluations than as a controlled clinical comparison.

The registry lists overall enrollment of 635. The three described groups address different diagnostic questions. The specificity group used specimens from apparently healthy individuals considered at low risk for HIV infection and tested them with both the investigational assay and an FDA-licensed HIV test. The sensitivity group used confirmed positive specimens for HIV-1 antigen, HIV-1 antibody, or HIV-2 antibody. The reactivity group used specimens collected from individuals at risk for HIV infection. Because these groups have different specimen selection criteria and purposes, they are not interchangeable arms in a randomized comparison.

The sensitivity description names 63 samples, panels, or viral isolates for HIV-1 antigen, 1003 for HIV-1 antibody in a US population, and 201 for HIV-2 antibody in an endemic area of Ivory Coast. These are subgroup counts stated in the outcome description; they should not be treated as a single participant total or assumed to represent the same sampling frame. The record does not explain how these specimen counts relate to the overall enrollment figure.

The study summary says that specimens were tested using the investigational assay and that results were compared with an approved HIV antibody test, with supplemental testing to determine HIV status. It does not establish that every specimen received every test. The registry lists the study as beginning in 2009-06, with primary completion in 2009-10, and status as completed. Further operational details, including recruitment procedures, specimen handling, and the exact analysis populations, are not reported in the registry record.

Since the study is single-group and non-randomized, differences between the specimen groups cannot be interpreted as the effect of assignment to one test strategy rather than another. The design can describe assay results within selected groups, but direct comparative claims require results and methods that the registry record does not report.

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: All subjects were assigned to one group. There was no pre-screening criteria for this study.

General Population From 16 to 89 Years

Reasons for not completing this period:

Pregnant Females Between 16-44 Years

Reasons for not completing this period:

Pediatric From 2 to 21 Years

Analysis population for the primary outcome, as defined in the registry: Specificity determined from 6,164 specimens from apparently healthy individuals (low risk): includes 250 specimens from pregnant females in first trimester and 580 specimens from low risk population prospectively collected and tested fresh during study. Total of 37 confirmed HIV positives by supplemental testing were excluded from 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 Architect HIV Combo Test Data for Clinical Sensitivity in HIV Positive Specimens, measured over 3 months.

The primary endpoint is the number of blood specimens in the clinical sensitivity evaluation among confirmed HIV-positive specimens. The endpoint title identifies clinical sensitivity, while the registered parameter is a number. The population description specifies confirmed positive specimens or panels for HIV-1 antigen, HIV-1 antibody, and HIV-2 antibody, with confirmation based on supplemental testing.

Reference status was determined using HIV-1 Western blot, HIV-2 Western blot, and/or HIV-1 ribonucleic acid (RNA) testing. Such supplemental methods provide the basis for classifying specimens as positive in the sensitivity evaluation. In general, a sensitivity analysis asks how often an assay identifies specimens that are truly positive according to the chosen reference standard. The count of positive specimens tested establishes the denominator context, but the count alone is not a sensitivity estimate.

The registry record does not report how many confirmed positive specimens tested positive on the investigational assay, how many tested negative, or a percentage sensitivity. It also does not report confidence intervals or separate assay-performance estimates by HIV marker. Consequently, the endpoint description explains which positive specimen groups were intended for assessment and how positive status was established, but the posted record does not supply the numerical performance result.

In general, using separate marker-specific groups can be useful because an assay may be evaluated against more than one biological target. The results still need to be interpreted in relation to how each specimen group was selected, the applicable reference testing, and the number of specimens with interpretable assay results. Those result details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical interpretation of a diagnostic study depends on the outcome definition, the reference standard, the sample selection, and the way results are tabulated. The registry record gives some information about the first two elements: the primary outcome is a count of confirmed positive specimens, and positive status was determined using HIV-1 Western blot, HIV-2 Western blot, and/or HIV-1 RNA testing. It does not report the assay outcome counts or the statistical methods used to estimate performance.

What does a sensitivity estimate represent?

In general, sensitivity is the proportion of reference-positive specimens that the test identifies as positive. The numerator is the number of reference-positive specimens with a positive assay result, and the denominator is the number of reference-positive specimens with evaluable assay results. The registry record identifies confirmed positive specimen groups but does not report either the assay-positive count or the evaluable denominator for each group. A specimen count alone cannot be converted into sensitivity without those result details.

In general, sensitivity is often reported with a confidence interval. A confidence interval describes the statistical uncertainty around an estimate under the assumptions of the method used; it does not remove uncertainty from specimen selection or errors in the reference classification. No sensitivity estimate or confidence interval is reported in the registry record, and the method for any such interval is not reported in the registry record.

How is specificity different?

In general, specificity is the proportion of reference-negative specimens that the test classifies as negative. It is distinct from sensitivity because it uses a different reference-status group and answers a different question. The registry describes a low-risk specificity group and says those specimens were tested with the investigational assay and an FDA-licensed HIV test. It does not report the reference-negative denominator, the investigational assay’s negative count, or the number of discordant results. Therefore, a specificity estimate cannot be recovered from the registered group description.

For both sensitivity and specificity, a result table would ordinarily distinguish concordant and discordant classifications relative to the reference method. In general, discordance can arise because the investigational test or the reference method classifies a specimen incorrectly, and the interpretation depends on the reference standard and testing procedures. The registry identifies supplemental tests for positive status but does not detail all procedures for resolving discordant or indeterminate results. Those procedures are not reported in the registry record.

Why do the separate specimen groups matter?

The sensitivity, specificity, and reactivity groups have different eligibility descriptions. The confirmed-positive sensitivity specimens were selected to assess detection among positive specimens; the low-risk specificity specimens were used to assess results among a group expected to have low HIV prevalence; and the reactivity group involved people at risk for HIV infection. General statistical reasoning cautions against combining such groups as though they were a single random sample. Their composition and the purpose of sampling differ, and the record does not provide a common sampling framework or a method for pooling results.

The registry describes the sensitivity specimens as 63 samples, panels, or viral isolates for HIV-1 antigen, 1003 US population for HIV-1 antibody, and 201 in an endemic area of Ivory Coast for HIV-2 antibody. The descriptions indicate distinct specimen sources and targets. They do not establish a single denominator or a direct comparison among the marker groups. The registry does not describe any adjustment for differences in specimen source, population, or test target.

What does uncertainty depend on?

In general, the precision of an estimated proportion depends in part on how many evaluable specimens contribute to its denominator and on the observed proportion. Larger denominators can reduce sampling uncertainty under suitable assumptions, while small or highly selected groups can yield estimates that are less precise or less generalizable. This general point does not establish the precision of this assay’s results: the registry does not report the outcome counts needed to calculate estimates or uncertainty intervals.

In general, diagnostic performance estimates from selected specimen sets may not equal performance in routine testing populations. Selection criteria, disease spectrum, specimen handling, and the reference method all affect how well an estimate applies beyond the specimens studied. The registry record describes low-risk, confirmed-positive, and at-risk specimen groups, but does not provide enough detail to characterize their full clinical spectrum or to infer population-level predictive values.

What can planning calculations say?

Any sample-size planning illustrations associated with these analyses are planning calculations, not observed trial results and not observed power. The stated planning framework uses a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. The registry record does not report the study’s own sample-size justification or say that the trial was designed using those assumptions. Moreover, the study is registered as non-randomized and single-group, so a hypothetical equal-allocation comparison does not describe its actual design.

In general, a planning illustration asks how much information might be needed to detect an assumed effect under specified assumptions. It does not show that the effect was observed or that the study had a particular probability of detecting it. The assumptions and the underlying effect used for any calculation need to be distinguished from study results. The registry record does not report the protocol’s design assumptions, interim analyses, or a statistical analysis plan.

What remains unknown about the analysis?

The registry lists five outcome measures and eight analyses, but does not explain how the analyses were structured, which were prespecified, or how their results were interpreted. It does not report the statistical tests, confidence-interval methods, handling of repeated or related specimens, or procedures for missing and indeterminate results. These omissions limit assessment of reproducibility and of the uncertainty around any unposted performance estimates. In general, diagnostic accuracy should be evaluated using clear definitions of reference status, analysis denominators, and result classification; the registry record does not report those details sufficiently to reconstruct such estimates.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists five outcome measures and eight analyses. It does not describe which analyses were primary, secondary, or exploratory beyond identifying the clinical sensitivity outcome as primary and listing reactivity as a secondary outcome. The analysis hierarchy, any prespecified testing sequence, and any adjustment for multiple statistical comparisons are not reported in the registry record.

In general, multiplicity arises when several outcomes, subgroups, or statistical tests are examined. If each test is interpreted independently without accounting for the number of tests, the chance of at least one apparently notable result can increase. Adjustment methods or a prespecified hierarchy can address that issue, but no such method is described in the registry record. The number of listed analyses alone does not establish which results were tested or whether any adjustment was used.

The record also does not report the number of specimens with missing, invalid, or indeterminate assay results, or how such specimens were handled. It does not describe whether the primary analysis included all specimens enrolled, only specimens with a completed assay, or another defined analysis population. Such details matter because exclusions can change the denominator used for diagnostic performance estimates. The registry record does not report the analysis populations or missing-data methods.

In general, missing diagnostic results may be informative if the chance of an unavailable or indeterminate result is related to specimen characteristics or assay performance. Whether that concern applies cannot be determined from the registry record. No imputation, sensitivity analysis, or other missing-data procedure is reported. As a result, the posted information does not permit a complete assessment of the multiplicity strategy, missingness, or the population contributing to each analysis.

11. 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.

The adverse-event listing reports zero serious events among 586 at risk in the specificity group and zero among 55 at risk in the reactivity group. For the sensitivity group, the safety fields list zero at risk and zero affected. These entries should be read as registry safety counts for the named groups, not as diagnostic performance results.

The safety listing also records zero other events among 586 at risk in the specificity group and zero among 55 at risk in the reactivity group. The record does not provide event terms in the serious-event listing. It does not report deaths for these groups, and the death fields are not populated with a count. No serious-event terms are listed.

In general, zero reported events does not establish that the risk of an event is absent. Interpretation depends on the number and type of people or specimens monitored, the duration and completeness of observation, event definitions, and the way events were collected. The registry states a primary outcome timeframe of 3 months, but does not establish that this timeframe applies to all safety surveillance. The safety entries do not provide exposure duration, completeness of follow-up, or an attribution assessment for the assay.

Accordingly, these counts are best described narrowly: no serious events are recorded for the groups with nonzero safety denominators in the registry listing. They do not independently demonstrate an absence of harm or provide a comprehensive safety evaluation.

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

The registry record establishes that the ARCHITECT HIV Ag/Ab Combo assay was evaluated in a completed, non-randomized study involving different specimen groups for sensitivity, specificity, and reactivity. It identifies the primary outcome as a count of confirmed HIV-positive specimens and describes supplemental methods used to determine positive status. These details explain the study’s diagnostic purpose and the broad design used to pursue it.

The record does not report the primary assay-result counts, sensitivity estimate, specificity estimate, or confidence intervals. It also does not report cross-classified results against the FDA-licensed or approved test, or the results for the at-risk reactivity group. Thus, the posted information does not establish how frequently the investigational assay detected confirmed positive specimens or how often it produced negative results in the low-risk group.

The overall enrollment of 635 should not be read as the denominator for every outcome. The record identifies distinct specimen groups and marker-specific specimen counts, but does not explain how those counts relate to overall enrollment or to the number with evaluable assay results. Without analysis-specific denominators and outcome counts, the precision and completeness of any performance estimate cannot be assessed.

In general, diagnostic results are most interpretable when the reference standard, specimen selection, assay-result categories, and analysis denominators are clearly stated. The registry supplies some context about reference testing and specimen sources, but leaves the actual test results and several analytic details unreported. This does not establish that the assay performed well or poorly; it means that the posted record is insufficient to quantify that performance.

The record also does not support a direct treatment-effect interpretation. This was not a randomized clinical comparison, and the described groups have different functions rather than serving as equivalent control and intervention arms. The comparator test is mentioned for the specificity evaluation, but comparative results are not reported.

The serious-event counts are limited registry entries and should not be used to infer overall safety. Likewise, any planning illustrations based on assumed fractions of an observed effect are not results from the study and are not observed power. The distinction between study design, outcome definitions, and posted results is essential: the first two are described, while the numerical diagnostic findings needed for an assessment are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically important as an example of why diagnostic evaluations need to be read differently from randomized treatment comparisons. The study is registered as single-group and non-randomized, with separate groups for confirmed positive specimens, low-risk specimens, and specimens from individuals at risk for HIV infection. Those groups address different questions. Treating them as interchangeable arms would obscure the sampling purpose and could lead to invalid comparisons.

The primary outcome is described as a number of positive specimens tested. That is a useful descriptor of the evidence base, but it is not itself a sensitivity estimate. To calculate sensitivity, the assay’s positive-result count and the denominator of evaluable reference-positive specimens are needed. The registry record does not report those results. Similarly, a specificity estimate needs results among reference-negative specimens, which are not reported. This distinction between denominator information and performance results is a basic but important statistical issue.

The record also illustrates the role of reference standards. Positive status was determined using supplemental Western blot and/or HIV-1 RNA testing. In general, a diagnostic performance estimate depends on how reference status is assigned, what happens when methods disagree, and whether all specimens receive a consistent verification process. The available description does not report enough procedural detail to assess those matters fully.

Finally, the record separates the existence of a completed study from the availability of interpretable results. It names the outcome, study groups, overall enrollment, and timeframe, but does not report the numerical diagnostic findings or their uncertainty. This distinction matters for evidence synthesis: a registry record may document the intended evaluation without containing the outcome data required to estimate accuracy. Clear reporting of counts, denominators, uncertainty, exclusions, and analysis populations is necessary for readers to understand both the magnitude and precision of diagnostic performance.

14. 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 central limitation is that the registry record does not report the primary endpoint result. It gives the endpoint definition and describes confirmed positive specimen categories, but provides no count of specimens detected by the investigational assay, no sensitivity estimate, and no confidence interval. The absence of these results prevents quantitative interpretation of the primary outcome.

Specificity and reactivity results are also not reported. Although the specificity group is described as being tested with both the investigational assay and an FDA-licensed test, there are no diagnostic cross-classification counts or comparative estimates. The at-risk reactivity group is described, but its assay outcomes and reference-status results are not given.

The relationship between the overall enrollment of 635 and the sensitivity subgroup specimen counts is not explained in the registry record. Those subgroup descriptions refer to samples, panels, or viral isolates from different populations and infection markers. The record does not report whether specimens overlapped, how many were evaluable, or how missing and indeterminate results were handled. Therefore, the listed counts cannot be treated as a unified analysis denominator.

The non-randomized, single-group design is appropriate to understand as a specimen-based diagnostic evaluation, but it does not support the logic of a randomized comparison across the described groups. The groups differ by purpose and specimen selection. In addition, the record does not report detailed selection procedures, the full reference-testing algorithm, handling of discordant results, assay failure rates, or analysis populations.

Further limitations include the absence of a reported statistical analysis plan, test methods, multiplicity strategy, and sample-size justification. The registry reports a primary completion timeframe but does not explain the observation period for all safety outcomes. These gaps do not prove that the study lacked such procedures; they mean those details are not reported in the registry record and cannot be assessed from it.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Evaluation of Diagnostic HIV Ag/Ab Combo Assay — NCT00933933.
• Registry record last updated 2011-07-12; status last verified 2011-07.
• Results first posted 2010-08-19.
• 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.

18. Record Summary

The registry record describes a completed Phase 3, non-randomized, unmasked, single-group evaluation of the ARCHITECT HIV Ag/Ab Combo assay, with overall enrollment of 635. The study evaluated confirmed positive specimens for clinical sensitivity, low-risk specimens for specificity, and specimens from individuals at risk for HIV infection for reactivity. The primary outcome is the number of confirmed HIV-positive specimens tested, with positive status determined using HIV-1 Western blot, HIV-2 Western blot, and/or HIV-1 RNA testing. The record does not report the assay-result counts, sensitivity or specificity estimates, confidence intervals, or analysis denominators. Serious-event listings report zero affected among 586 at risk in the specificity group and among 55 at risk in the reactivity group. Missing-data procedures, analysis populations, multiplicity methods, and the study’s own sample-size justification are not reported in the registry record.