Reducing Assessment Barriers for Patients With Low Literacy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reducing Assessment Barriers for Patients With Low Literacy (NCT03584490), a randomized na study. The primary outcome is The Degree of Differential Item Functioning (DIF) in NIH Patient-Reported Outcomes Measurement Information System (PROMIS Profile 57 v 2.0) Anxiety Subscale.

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group study titled “Reducing Assessment Barriers for Patients With Low Literacy.” Its stated purpose is to determine the effects of health literacy on questionnaire-based measurement. The record lists health literacy as the condition and Northwestern University as the sponsor. Study status is completed. The recorded study start date is 2018-09-14, and the primary completion date is 2021-09-09.

The posted primary outcome compares differential item functioning (DIF) across adequate and low health-literacy groups for the NIH PROMIS Profile 57 v 2.0 anxiety subscale. The outcome timeframe is 6 months. The record presents results for a pen-and-paper format and a computerized talking touchscreen format. It describes the former as traditional paper questionnaires and the latter as a computerized version that allows questions and answer choices to be read aloud. Randomization is reported, masking is listed as none, and the model is parallel.

Enrollment is listed as 729. The arm records give 362.0 participants for pen-and-paper and 367.0 for the computerized talking touchscreen. The registry also lists phone administration of questionnaires as an intervention, but it does not provide a corresponding arm description or a result for phone administration in the posted primary outcome data. The phase is recorded as NA. Further details about the randomization procedure, allocation concealment, recruitment, eligibility criteria, and conduct of the study 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

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 Degree of Differential Item Functioning (DIF) in NIH Patient-Reported Outcomes Measurement Information System (PROMIS Profile 57 v 2.0) Anxiety Subscale, measured over 6 months.

The primary endpoint is the degree of differential item functioning in the anxiety subscale of the NIH PROMIS Profile 57 v 2.0. DIF describes whether response patterns for an item differ according to health-literacy group. The record measures this with McFadden pseudo-R-square, with the maximum value across items reported by condition. A higher value corresponds to higher DIF, while a lower value corresponds to lower DIF.

The stated rationale is to assess whether health literacy, or its interaction with anxiety, determines the probability of a response type. This matters for questionnaire measurement because different response patterns across literacy groups may indicate that an item functions differently between those groups. The record notes that the scale is normalized to a theta metric, with mean 0 and standard deviation 1 by definition, and that PROMIS is scaled by T-scores referenced against the US population mean. Item response theory is used to aggregate items, but the registry says aggregation is not relevant to this endpoint because the pseudo-R-square is evaluated for each item.

The registry reports the maximum DIF values as 0.06 for pen-and-paper and 0.02 for the computerized talking touchscreen. It does not provide item-level results, uncertainty intervals, or an inferential test for the difference between formats. These reported values therefore describe the endpoint as posted; they do not by themselves establish whether the formats differ statistically.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry uses McFadden pseudo-R-square to express the degree of DIF. In general, a pseudo-R-square is a measure based on a model’s fit, but it is not the same quantity as the R-square commonly used in linear regression. The registry describes it as an analogue to R-square and says it captures the degree to which health-literacy group determines the probability of response type for an item. It also notes that health literacy and/or its interaction with anxiety may determine that probability. The number is therefore a model-based measure of response-pattern differences, not a direct percentage of participants affected.

What does the maximum across items represent?

The outcome is the maximum DIF value across items in the anxiety subscale, reported by condition. A maximum emphasizes the item with the largest observed value rather than describing a typical item or an overall average. This can be useful when the analysis seeks to identify the strongest item-level signal, but it gives no information by itself about how the remaining items behaved. The registry record does not identify the item that yielded the maximum in either arm or provide the item-level values.

The record says item response theory is used to aggregate items, but that aggregation is not relevant to the reported pseudo-R-square because the measure is evaluated for each item. It also states that each scale is normalized to a theta metric, with mean 0 and standard deviation 1 by definition. PROMIS is described as scaled by T-scores referenced against the US population mean. These scale details provide context for the measure; they do not supply an uncertainty estimate for the posted arm results.

How should the arm values be read?

The record reports 0.06 for the pen-and-paper format and 0.02 for the computerized talking touchscreen. In the registry’s description, higher values indicate higher DIF and lower values indicate lower DIF. These are descriptive endpoint values. They are not accompanied by confidence intervals, standard errors, or a reported hypothesis test. A confidence interval, in general, represents a range of values compatible with the data and the method used to calculate it; without one, readers cannot assess the precision of these arm estimates from the posted results.

The registry’s skip note says values are missing or not estimable for a two-arm comparison for this count-other endpoint. This is consistent with the absence of a posted inferential comparison. The arm values can be stated as recorded, but a difference between them should not be assigned a p-value or described as statistically significant on the basis of this record. The registry does not report an adjusted or unadjusted statistical test for the primary endpoint.

Why does over-dispersion matter in general?

Over-dispersion is a general concern for count outcomes when variability exceeds what a statistical model assumes. If it is not accounted for, standard errors and uncertainty estimates may be poorly calibrated. This primary endpoint is described as a pseudo-R-square for item-level response probabilities, not as a count outcome. The registry record does not report an over-dispersion assessment or a statistical model specification for the posted arm values. No claim about whether over-dispersion was present or addressed can therefore be made.

What can sample-size calculations establish?

In general, a sample-size calculation can illustrate how many participants per arm might be needed under specified assumptions to detect a chosen effect with a chosen power. Such planning calculations are not observed results and do not show the power actually achieved in a completed trial. The registry record does not report the trial’s own sample-size justification or provide a statistical plan explaining the assumptions used for its design. It also does not report an interim-analysis plan. These details are not reported in the registry record, so the observed enrollment cannot be used to infer what effect the trial was designed to detect or the precision it was expected to achieve.

The record lists zero posted analyses. It therefore does not provide a fuller account of model specification, covariate adjustment, handling of multiple item-level assessments, or analysis populations. The posted endpoint definition and arm values support a descriptive account, but not a reconstruction of the underlying analysis procedure beyond what the registry explicitly states.

10. Multiplicity, Missing Data and Analysis Populations

The outcome is based on DIF evaluated for each anxiety-subscale item, with the maximum value across items reported by condition. Considering multiple items can raise multiplicity questions: when many item-level quantities are examined, the chance of observing an unusually large value may differ from the chance associated with a single pre-specified quantity. The registry record does not report a multiplicity adjustment, a procedure for selecting the maximum, or details of any item-level testing. It is therefore not possible to determine how multiplicity was handled or whether any formal tests were conducted.

The record lists three outcome measures, but only the primary endpoint is described in the registry entries. It reports zero secondary outcomes and zero analyses. The registry record does not report a full statistical analysis plan, analysis populations, or which participants contributed to the primary endpoint. Missing-data methods, reasons for missing observations, and the handling of incomplete questionnaires are not reported in the registry record.

Although the record lists enrollment and arm counts, it does not state whether those counts match the population analyzed for the primary outcome. It also does not report exclusions, protocol deviations, or whether any analysis followed the randomized assignment regardless of questionnaire completion. Thus, the posted values should not be assumed to represent a particular intention-to-treat or per-protocol analysis population. These omissions limit assessment of selection, missingness, and the reproducibility of the analysis.

11. 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.

The adverse-event record lists zero serious adverse events and zero deaths for the pen-and-paper format and for the computerized talking touchscreen. It also lists zero other adverse events for both groups. These are the posted counts in the registry record; they should be read as reported event counts rather than as proof that no events occurred outside the record’s collection or reporting process.

The record does not provide descriptions of event surveillance, definitions, follow-up procedures, or adjudication. It does not report whether adverse events were actively collected, how missing safety information was handled, or whether the listed counts cover the same follow-up period as the primary outcome. The available counts therefore offer limited detail about safety reporting and cannot establish the completeness of ascertainment.

The listed serious adverse-event counts are distinct from the primary endpoint, which concerns questionnaire DIF. The registry does not report a formal safety comparison between formats. These counts alone should not be interpreted as evidence that either format is safer, and they do not indicate whether participants found the formats acceptable or accessible.

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

The posted primary results describe maximum DIF values of 0.06 for pen-and-paper and 0.02 for the computerized talking touchscreen. According to the registry’s definition, a higher value corresponds to higher DIF and a lower value to lower DIF. These observations address the specified anxiety-subscale endpoint at 6 months. They provide a descriptive comparison of the recorded endpoint values, but the registry record does not report an inferential test or uncertainty interval for that comparison.

The result does not establish that the computerized format reduces DIF in a statistically reliable or clinically important way. No confidence interval, standard error, p-value, or reported comparison estimate is available. The record also does not identify which item generated each maximum, so the two values cannot be interpreted as a complete account of item-level behavior. A maximum does not show whether most items had similar values or whether the largest value was an exception.

The endpoint is not a direct measure of questionnaire completion, comprehension, user preference, access, or health outcomes. The touchscreen description indicates that questions and answer choices can be read aloud, but the posted primary outcome does not quantify the effect of that feature on those broader outcomes. The record includes phone administration as an intervention but gives no result for a phone group in the primary outcome data; no conclusion about phone administration follows from the posted arm values.

More broadly, the result is specific to the stated PROMIS anxiety subscale, the listed formats, and the recorded follow-up timeframe. The registry does not report enough detail about participant characteristics, missing data, or the analysis population to establish how broadly the results apply. The posted values can inform a focused discussion of DIF as defined in the record, but they do not support claims about all questionnaires, all populations with low health literacy, or the overall effectiveness of a format.

13. Why This Trial Matters Statistically

This trial addresses a measurement question: whether health-literacy group is related to response patterns on questionnaire items, and whether the recorded DIF differs across questionnaire formats. That distinction is statistically important. A score comparison may be difficult to interpret if an item’s response behavior varies across groups for reasons connected with how the item is understood or answered. Examining DIF can therefore help researchers evaluate the properties of patient-reported measures in groups with different health-literacy levels.

The registry’s outcome focuses on item-level response probabilities and summarizes them with a maximum McFadden pseudo-R-square. This approach makes the item-level nature of the question explicit. It also illustrates why the endpoint definition matters: a maximum across items is not the same as an overall score difference, a mean DIF value, or a direct measure of accessibility. Statistical interpretation depends on matching the conclusion to the quantity actually analyzed.

The study also provides a setting for considering how administration mode and measurement properties intersect. The registry describes a touchscreen that can read questions and answer choices aloud, alongside traditional pen-and-paper questionnaires. The posted maximum DIF values offer descriptive information for these formats. However, the absence of reported inferential results, uncertainty estimates, item-level values, and analysis details limits what can be concluded from those figures.

For clinical researchers, the record is a reminder that measurement outcomes require transparent reporting of estimands, item-level procedures, missing-data handling, and uncertainty. Those details are especially useful when a result is defined by the largest value across multiple items. Without them, readers can understand the endpoint’s stated meaning and the values recorded, but cannot fully evaluate the precision or robustness of the comparison.

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 registry record does not report confidence intervals, standard errors, p-values, or a statistical test comparing the arm values. It also states that values are missing or not estimable for a two-arm comparison for this count-other endpoint, and it lists zero analyses. The result is consequently descriptive, with limited information about precision and no posted inferential assessment.

The outcome is the maximum DIF value across items, but item-level results and the identity of the item producing each maximum are not reported. The record does not describe how item-level values were selected, whether multiplicity was addressed, or whether any item-level testing took place. Without these details, the maximum cannot be evaluated in the context of the full distribution of item results.

Important design and analysis details are also absent. Eligibility criteria, participant characteristics, randomization procedures, allocation concealment, analysis populations, missing-data methods, protocol deviations, and the statistical analysis plan are not reported in the registry record. The registry does not explain how the listed enrollment and arm counts relate to the population contributing to the endpoint. It also does not report the trial’s sample-size justification or any interim-analysis plan.

The intervention information is incomplete for interpreting all listed formats. Phone administration appears in the interventions list, but a corresponding arm description and primary outcome result are not reported. The posted endpoint is limited to the PROMIS anxiety subscale and does not provide results for completion, comprehension, preference, or other aspects of questionnaire use. These limitations constrain interpretation to the recorded endpoint values and the specific formats with posted results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Reducing Assessment Barriers for Patients With Low Literacy — NCT03584490.
• Registry record last updated 2024-10-15; status last verified 2024-10.
• Results first posted 2023-12-05.
• 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

NCT03584490 is a completed randomized parallel-group study sponsored by Northwestern University. It examines health literacy and questionnaire-based measurement. The registry lists enrollment of 729, no masking, and a primary completion date of 2021-09-09. The primary endpoint is the maximum McFadden pseudo-R-square for DIF across items in the NIH PROMIS Profile 57 v 2.0 anxiety subscale at 6 months.

The posted endpoint values are 0.06 for pen-and-paper and 0.02 for the computerized talking touchscreen. The registry defines higher values as higher DIF. It does not report uncertainty estimates or an inferential comparison, and it lists zero analyses. Serious adverse events, other adverse events, and deaths are recorded as zero in both listed groups. Phone administration is listed as an intervention, but no corresponding primary outcome result is reported.