Understanding Disparities in Healthcare and Primary Care Provider Quality: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Understanding Disparities in Healthcare and Primary Care Provider Quality (NCT02626910), a registered phase not stated study. The primary outcome is Satisfaction With Medical Care Survey.

2. Clinical Question

The underlying rationale is that healthcare disparities associated with race or ethnicity, economic circumstances, and education have been documented, while less is known about disparities related to disability. Accordingly, the research question is descriptive and comparative: what do participants report about medical care, and how do reported experiences vary across the populations and settings identified in the record? The available registry information establishes the topic and outcome, but it does not establish that the listed groups are directly comparable. Recruitment approaches and group composition differ, and no randomized allocation is stated. The registry also does not report enough information to determine whether the participants represent the broader population of adult Medicare beneficiaries with disabilities or those without disabilities.

3. Trial Design

The registry describes a completed project sponsored by the University of Pennsylvania. Its enrollment is listed as 49. The record gives a start date of 2013-08-01 and a primary completion date of 2017-07-01. These dates and the completed status describe the registry record; they do not establish when individual participants entered or completed study activities.

The listed groups are Second Life: People with disabilities, Second Life: People without disabilities, Second Life: Clinicians, and Urban. The first three groups are described as recruited from the virtual world Second Life. The urban group is described as including people with and without disabilities recruited from an urban setting. The registry does not report how people were assigned to groups, and it explicitly characterizes the study as not being a randomized comparison. These groups therefore should not be treated as randomized arms or as a like-for-like control comparison.

The project’s stated purpose is to investigate disparities in healthcare and primary care provider quality from the perspectives of adult Medicare beneficiaries with disabilities compared with those without disabilities. The record lists one primary outcome measure, Satisfaction With Medical Care Survey, with a one-year timeframe. It describes a scale-based outcome and says multinomial logistic regression models will be used to determine the degree of satisfaction reported by people with and without disabilities.

The registry does not report a detailed protocol, participant flow, recruitment dates, group-specific eligibility criteria, or the number analyzed for the primary outcome. It also does not report a complete statistical analysis plan or results from the planned model. As a result, the design can be described in broad terms, but the record does not support a detailed account of how recruitment, follow-up, or analysis was conducted.

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: Satisfaction with healthcare

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 Satisfaction With Medical Care Survey, measured over One year.

The primary outcome is the Satisfaction With Medical Care Survey, described as measuring satisfaction with healthcare. Its scale ranges from 1 (not satisfied) to 4 (very satisfied), with higher scores representing more satisfaction with medical care. The unit is units on a scale, and the registry identifies the outcome parameter as a mean with standard deviation as the dispersion measure. The timeframe is one year.

Satisfaction is relevant to this research question because the project aims to understand healthcare experiences from the perspectives of people with and without disabilities. A measure of reported satisfaction can describe how participants evaluate care and can be used to examine whether reported experiences vary across the groups of interest. It is not, by itself, a direct measure of clinical outcomes, provider technical performance, access to care, or the causes of any differences in experience.

The outcome description also states that multinomial logistic regression models will be used. The registry does not explain how a scale identified as a mean-based continuous measure will be represented in that model, nor does it report the model’s specification, covariates, or estimates. The record therefore establishes the intended construct and the scale’s interpretation, but not the details needed to reproduce the analysis. No observed mean, standard deviation, model estimate, or confidence interval is reported, so the amount and precision of any difference in satisfaction cannot be assessed.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The record identifies a scale-based primary outcome, gives a mean and standard deviation as the intended summary, and describes multinomial logistic regression as the planned modeling approach. These statements identify parts of the intended analysis, but they do not provide a complete statistical specification. No analyses or fitted estimates are listed. The following general explanations clarify the methods without implying that particular results or procedures were reported.

What does a mean describe?

General: a mean summarizes the central value of observations on a quantitative measure. For a satisfaction scale, it can provide a compact description of the average reported score. A mean alone does not show how responses are distributed, whether responses cluster at particular scale values, or whether differences between groups are driven by a subset of observations. The standard deviation describes how spread out observations are around the mean. It is not a measure of uncertainty in the estimated mean; a confidence interval or standard error would address estimation precision.

The scale is bounded, with a stated minimum of 1 and maximum of 4. General: when a bounded rating scale is summarized with a mean, that summary treats differences between adjacent scale values as meaningful on a quantitative scale. Whether that treatment is appropriate depends on the measurement properties and analytic objective. The registry does not report a rationale for the scale’s statistical treatment, a distribution of responses, or diagnostics that would help assess it.

What does multinomial logistic regression do?

General: multinomial logistic regression models the probability of outcomes in categories when the response has more than two possible categories without a required ordering. It can compare category probabilities across groups and, depending on the model specification, account for other measured variables. Its estimates are model-based and depend on how the response and predictors are defined, which categories are used as references, and which covariates are included.

The registry says this type of model will be used to determine satisfaction among people with and without disabilities. However, it also characterizes the outcome as a mean of a continuous measure with standard deviation. The record does not explain whether the scale responses were to be modeled as categories, how categories would be defined, or whether the stated mean and standard deviation were descriptive summaries alongside a categorical model. It also does not report reference categories, covariates, interaction terms, model diagnostics, or a method for handling the clinician and urban groups. Those details are not reported in the registry record, so the analysis cannot be reconstructed from the description alone.

Why does study design matter to the model?

General: regression can adjust comparisons for variables that are measured and included in a model, but adjustment does not make a nonrandomized comparison equivalent to a randomized one. Unmeasured or poorly measured differences can remain, and different recruitment settings can affect who participates and how they respond. The registry describes recruitment from Second Life and an urban setting, but does not provide enough information about recruitment procedures or participant characteristics to assess the comparability of groups.

General: model estimates also rely on assumptions about the relation between predictors and outcome, the independence or dependence of observations, and the appropriateness of the selected outcome distribution. For a multinomial model, adequate information across outcome categories is important for stable estimates. The record does not report category counts, diagnostics, or methods for addressing sparse data. It is therefore not possible to assess model fit or stability.

What can planning calculations tell readers?

General: sample-size calculations are planning tools that describe how much information a proposed study might need under stated assumptions. The registry’s planning illustration is described as using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. Those are planning assumptions, not evidence that allocation was equal or that the study had a particular achieved power.

A planning calculation based on an observed effect is especially dependent on that effect estimate and on the assumptions used to translate it into a future study design. It should not be presented as observed power or as proof that the completed study could detect a meaningful difference. The registry does not report a sample-size justification for the actual project, a target enrollment rationale, or a result from the planned satisfaction analysis. The listed enrollment alone cannot fill those gaps.

General: interpretation should consider the estimate, its uncertainty, the outcome’s meaning, and the design that generated the data. A statistical result, even if available, would not by itself establish that a difference was caused by disability status or by a particular care provider. In this record, the absence of posted estimates means that even the descriptive magnitude and precision of satisfaction scores cannot be evaluated.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Second Life: People With Disabilities", "Second Life: People Without Disabilities", "Second Life: Clinicians", "Urban")
n   <- c(23, 10, 6, 10)
m   <- c(3.4, 3.8, 3.6, 3)
sp  <- c(1.2, 1.7, 1.9, 2.9)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Second Life: People With Disabilities  23  3.400  1.200  2.881  3.919
CARM  Second Life: People Without Disabilities  10  3.800  1.700  2.584  5.016
CARM  Second Life: Clinicians  6  3.600  1.900  1.606  5.594
CARM  Urban  10  3.000  2.900  0.925  5.075

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure, Satisfaction With Medical Care Survey, and no analyses. It does not describe additional primary hypotheses, secondary outcomes, or a multiplicity adjustment. The number of planned comparisons and whether the analysis involved subgroup or interaction testing are not reported in the registry record. Without those details, it is not possible to determine how many inferential tests were planned or whether the reported significance of any result would require adjustment for multiple testing. No significance test for the primary outcome is posted.

Missing-data procedures are not reported. The record does not say whether participants could omit survey items, how incomplete surveys were handled, whether missing outcomes were imputed, or whether analyses were restricted to participants with complete responses. It also does not provide outcome-specific denominators or a participant flow description. Consequently, enrollment of 49 should not be treated as the number with complete primary-outcome data.

The analysis populations are not defined. The record does not state whether the intended analysis included all enrolled participants, only those who completed the survey, or a particular subgroup. It does not report rules for classifying disability status, handling participants who might fit more than one group description, or including clinicians and the Urban group in the primary model. Allocation is not stated, and the record says the study is not a randomized comparison, so conventional randomized-trial populations such as intention-to-treat cannot be assumed.

General: missing-data methods matter because results can depend on who has an observed outcome and why other outcomes are absent. A complete-case analysis, for example, describes participants with recorded responses and may not represent everyone enrolled if missingness is related to satisfaction or participant characteristics. Imputation and weighting methods also require assumptions. Since no method or missingness pattern is reported, the potential influence of missing data on the intended comparison cannot be assessed.

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.

The adverse-event figures list 0 serious events affected and 0 deaths for Second Life: People With Disabilities, among 23 at risk. For Second Life:People Without Disabilities, the record lists 0 serious events affected and 0 deaths among 10 at risk. For Second Life: Clinicians, it lists 0 serious events affected and 0 deaths among 6 at risk. For Urban, it lists 0 serious events affected and 0 deaths among 10 at risk. The record also reports 0 affected for other events in each of those groups.

These figures are registry-reported counts. A count of 0 means that no event in the relevant recorded category is reported as affecting participants in that group; it does not establish that the risk of an event is absent. The record does not provide event descriptions for the serious-event category, and no serious-event terms are listed. It also does not state the adverse-event collection period or provide further detail about ascertainment. The listed figures should therefore be read as the safety information recorded, not as a comprehensive account of every possible harm or as evidence about comparative safety.

The study is not described as a randomized comparison, and the groups differ in their recruitment descriptions. The event counts consequently do not support a causal comparison between groups. They should be interpreted alongside the limited group-specific enrollment information and the lack of detail about event collection and follow-up.

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

The record establishes a research focus: disparities in healthcare and primary care provider quality, considered from the perspectives of people with and without disabilities. It identifies satisfaction with medical care as the primary outcome and describes a scale on which higher scores indicate greater satisfaction. It also documents recruitment descriptions for participants from Second Life and an urban setting, alongside a clinician group.

It does not establish the direction or magnitude of any difference in satisfaction. No observed score, group contrast, model estimate, confidence interval, or p-value is reported. Although the registry describes multinomial logistic regression as a planned method, no fitted analysis is listed. Thus, the record cannot support a claim that one group reported greater or lower satisfaction, that any difference was statistically distinguishable from chance, or that an estimate was precise.

Even if comparative estimates were available, this record describes a nonrandomized comparison with allocation not stated. General: group differences in nonrandomized research may reflect differences in recruitment, participant characteristics, setting, or other factors, in addition to the characteristic being compared. Statistical adjustment can address measured variables included in a model, but cannot ensure that all relevant differences have been accounted for. The registry does not give the details needed to evaluate these issues.

The satisfaction measure addresses reported experience, not every dimension of healthcare quality. A satisfaction score alone does not establish clinical benefit, access, quality of technical care, or the reason for a participant’s response. Similarly, the registry’s completed status is not evidence that a particular hypothesis was supported. The most defensible interpretation is that the record describes a study question and an intended measure, while leaving the study’s comparative findings and their uncertainty unreported.

14. Why This Trial Matters Statistically

The statistical importance of this study lies in the question it seeks to make measurable: whether healthcare experiences and satisfaction differ for people with and without disabilities. The registry’s background notes that disparities related to race or ethnicity, economic circumstances, and education have been documented, while disability-related disparities are less understood. Measuring reported satisfaction can help frame that question in participants’ own evaluations of care, rather than treating provider quality only as an administrative or clinical construct.

The design information also illustrates why comparisons require careful interpretation. Participants were recruited through different settings, including Second Life and an urban setting, and a clinician group is listed. The groups are not described as randomized, and allocation is not stated. These features make it important to distinguish a descriptive association from a causal effect and to know how participant characteristics and recruitment context were handled. The registry does not report enough analytic detail to make that assessment.

For biostatistical interpretation, an outcome definition and a planned model are not substitutes for results. A usable comparison requires estimates, uncertainty measures, outcome denominators, and enough information about the model and analysis population to judge how the estimates were obtained. Here, the primary outcome is described, but analyses are not listed and no results are reported. That gap prevents readers from assessing whether the project found a difference or how informative the data were.

The record is therefore relevant as a documented research question and study framework, not as evidence for a particular comparative conclusion. Its statistical contribution cannot be judged from the outcome description alone. Full interpretation would require posted results and clearer information about recruitment, group composition, missing data, and the implemented analysis.

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.

The most direct limitation of the registry record is the absence of posted analysis results. It lists the primary outcome and a planned modeling approach but reports no estimates, uncertainty intervals, test results, or descriptive outcome summaries. Consequently, neither the direction nor precision of any satisfaction difference can be assessed.

The design description is also limited. The project is explicitly not a randomized comparison, and allocation is not stated. Recruitment settings vary: people with and without disabilities and clinicians are described as recruited from Second Life, while the Urban group includes people with and without disabilities recruited from an urban setting. The record does not describe how these settings contributed to the intended comparisons or how group membership was determined. It also does not report detailed eligibility criteria or participant characteristics. The extent to which findings would represent adult Medicare beneficiaries is therefore not established.

The relationship between the outcome summary and the planned model is unclear. The measure is characterized as a mean of a continuous scale with standard deviation, while the analysis description names multinomial logistic regression. The registry does not explain how the scale was encoded for that model or provide the model specification. Covariate selection, diagnostic procedures, and approaches to sparse categories are not reported.

Outcome completeness and analysis populations are not described. Although enrollment is listed as 49, the number with observed survey responses and the number analyzed are not reported. Procedures for missing data, multiplicity, and subgroup comparisons are also absent. The adverse-event information gives zero recorded counts for the listed groups but does not describe ascertainment or the reporting period. These limitations concern what can be concluded from the registry record; they do not establish what procedures may have been used outside the record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Understanding Disparities in Healthcare and Primary Care Provider Quality — NCT02626910.
• Registry record last updated 2017-12-15; status last verified 2017-10.
• Results first posted 2017-12-15.
• 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

The University of Pennsylvania-sponsored study, titled Understanding Disparities in Healthcare and Primary Care Provider Quality, is marked completed. The registry lists enrollment of 49, a start date of 2013-08-01, and a primary completion date of 2017-07-01. The project concerns healthcare disparities and primary care provider quality from the perspectives of adult Medicare beneficiaries with and without disabilities.

The primary outcome is Satisfaction With Medical Care Survey, measured on a scale from 1 (not satisfied) to 4 (very satisfied), with a one-year timeframe. The outcome is described using a mean and standard deviation, and the record says multinomial logistic regression models will be used. Listed groups include people with disabilities, people without disabilities, and clinicians recruited from Second Life, plus an urban group. The study is not described as randomized, and allocation is not stated.

No analyses or primary-outcome results are reported in the registry record. The record therefore does not provide estimates of satisfaction, comparative effects, or their precision.