Impact of Teledermatology on Health Services Outcomes in the Department of Veterans Affairs: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Impact of Teledermatology on Health Services Outcomes in the Department of Veterans Affairs (NCT00488293), a randomized na study. The primary outcome is Skindex-16 Change Scores Between Baseline to Month 9 - Composite Score.

2. Clinical Question

3. Trial Design

The registry identifies this as a randomized, parallel-group study with no masking. It was conducted by the US Department of Veterans Affairs and is titled “Impact of Teledermatology on Health Services Outcomes in the Department of Veterans Affairs.” The listed phase is NA. The record gives enrollment as 392 and marks the study completed. Its recorded start is 2008-11 and primary completion is 2011-02.

The two arms were a store-and-forward teledermatology consult process and a conventional consult process. The primary objective was to compare mean change in quality of life, assessed using Skindex-16 subscale and composite scores. The posted primary outcome is specifically the composite change score from baseline to Month 9. The registry also lists clinical course ratings between baseline and Month 9 and between baseline and the first clinic visit as secondary outcomes. It lists 10 outcome measures and 10 analyses overall.

Random allocation identifies how participants were assigned in the study design, but the registry record does not report the sequence-generation method, allocation concealment, or the process used to implement assignment. Masking is listed as none; further information about which participants, clinicians, or outcome assessors knew the assigned referral process is not reported in the registry record. The record does not report the original sample-size justification, its planning assumptions, or whether enrollment met a prespecified target. Those design details should not be inferred from the post hoc planning illustrations based on the observed effect.

The primary outcome arms list 160 teledermatology referrals and 164 conventional referrals. These are the posted counts for the primary outcome comparison, not a replacement for the registry enrollment figure. The record does not explain the difference between enrollment and those endpoint counts, or specify whether the primary analysis followed an intention-to-treat population, a per-protocol population, or another analysis set.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Baseline to Month 9 Skindex-16 change scores - Composite Score.

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.

Population: One patient assigned to teledermatology was misrandomized and not included in the analysis.

6. Primary Endpoint

The registry lists the primary outcome as Skindex-16 Change Scores Between Baseline to Month 9 - Composite Score, measured over Baseline to Month 9.

The primary endpoint is the change in the Skindex-16 composite score from baseline to Month 9. Skindex-16 is described in the registry as a quality-of-life measure assessing how much skin disease bothers a participant. It includes symptoms, emotions, and functioning, and the composite score combines these domains. Subscale and composite scores range from 0, meaning never bothered, to 100, meaning always bothered.

The endpoint uses change rather than a single score so that the reported outcome reflects movement between baseline and the Month 9 assessment. By the registry’s definition, a negative change score represents better quality of life. The record states that a change score of 10 points is considered clinically significant. That benchmark is useful context for interpreting score changes, but it does not by itself establish whether the difference between study arms is statistically or clinically important.

A mean change is a group-level summary of a continuous outcome. It describes the average direction and magnitude of change in each arm, while the standard deviation describes variability among individual scores. The primary comparison asks whether the average changes differ between the referral processes. The posted mean changes are -12.0 for teledermatology referrals and -13.2 for conventional referrals. These are changes within each arm; the between-arm estimate and its uncertainty are the relevant statistics for the comparative question.

The endpoint is relevant to the stated objective because it measures patient-reported quality of life rather than only service use or clinical course. The registry does not report further detail about endpoint completion, scoring procedures, or how missing questionnaire items were handled.

7. Results Posted in the Registry

• Conventional Referrals: mean -13.20 versus -12.00 on control; difference -1.20 (95% CI -6.23 to 3.83), standardized effect d = -0.05; z-test p = 0.640.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean change summarize?

A mean is the arithmetic average of observed values in a group. For this endpoint, the values are each participant’s change in Skindex-16 composite score from baseline to Month 9. The posted arm means are -12.0 for teledermatology referrals and -13.2 for conventional referrals. Because a negative change represents better quality of life under the registry’s definition, a more negative arm mean indicates a larger average decrease in reported bother. An arm mean describes the group average, not the experience of every participant. It does not show whether most individuals changed by a similar amount or whether a smaller number of large changes influenced the average.

The standard deviations are 24.5 and 21.6, respectively. A standard deviation describes how dispersed participant change scores are around their arm’s mean. It is not the uncertainty of the mean itself. The observed dispersion is substantial in relation to the arm means, which is one reason a group-average contrast needs to be read alongside its confidence interval rather than as a complete account of individual outcomes.

What does the mean difference represent?

The posted comparison is conventional referrals relative to teledermatology referrals, with an estimated mean difference of -1.2 units. In plain language, the estimated average change in the conventional-referral arm is more negative by that amount. The difference is a contrast between group means, not a change score for an individual. The effect-size summary d = -0.052 expresses the contrast on a standardized scale using the pooled standard deviation of 23.08. Standardization can help describe the size of a difference relative to outcome variability, but it does not change the clinical meaning of the original Skindex-16 scale.

Why is there a confidence interval?

A confidence interval expresses the uncertainty around an estimated contrast under the statistical model used to calculate it. The reported interval runs from -6.234 to 3.834. Values in the interval are compatible with the observed data under the method’s assumptions; the interval is not a range containing a known fixed proportion of individual changes. Its inclusion of zero means the posted summary does not rule out no average difference. Its extension in both directions also means that the direction of the difference is not determined precisely. The interval should be considered with the endpoint’s scale, the registry’s clinical-significance description, and limitations in the analysis reporting.

What is the role of a t-test?

The registry reports a two-sided t-test for the comparison and gives p = 0.66. A two-sided test considers departures from no difference in either direction, rather than testing only whether one named arm is better. The recalculated unadjusted p-value from the posted summary statistics is 0.64. A t-test evaluates the observed difference relative to the estimated variability and sample information under its assumptions. Its p-value does not indicate the size or practical importance of a difference, and it does not establish that the groups are equivalent when the result is not statistically persuasive.

The recalculation is an unadjusted comparison using the posted means, standard deviations, and outcome counts. It is not a reconstruction of every possible feature of the sponsor’s analysis. The registry does not report whether the sponsor’s analysis used covariate adjustment, how it handled missing outcomes, or whether any other analysis choices affected the reported result. The close p-values are consistent with the same broad reading, while the sponsor’s stated result remains the registry’s reported analysis.

Why can dispersion matter?

General statistical explanation: when individual outcomes vary widely, the group mean is estimated less precisely than it would be if scores clustered closely around the mean, all else being equal. This is why the standard deviation and the confidence interval provide different information. The standard deviation describes variation among participants; the confidence interval describes uncertainty in the estimated contrast. Over-dispersion, in a setting where a model assumes less variability than the data contain, can make standard errors too small and intervals too narrow. The posted endpoint is continuous and its dispersion is described with standard deviations; no over-dispersion assessment or alternative variance model is reported in the registry record.

What does a planning illustration add?

The sample-size planning figures use the observed effect as a starting point and consider fractions of that effect. They describe hypothetical planning under stated assumptions, not the original trial design. The assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The planned effect to detect is the stated fraction of the observed effect. These assumptions support an illustration of how sample size depends on the target effect and desired power, but they should not be read as evidence that the original study was designed using those choices.

The listed planning values are participants per arm for 80% power and participants per arm for 90% power. For 100% of the observed effect, the figures are 5,806 and 7,773; for 75% of the observed effect, they are 10,322 and 13,818; and for 50% of the observed effect, they are 23,223 and 31,089. The figures illustrate why a smaller target effect generally requires more participants to detect it with the same power under the same assumptions. These planning calculations are separate from the observed trial results and are not an observed-power calculation.

10. Confidence Intervals and Precision

The estimated difference in mean change is -1.2 units, and the 95% confidence interval extends from -6.234 to 3.834. This interval conveys uncertainty about the estimated average contrast. It crosses zero, so the data are compatible with no average difference as well as differences favoring either direction. The estimate alone should not be treated as a precise determination of which referral process had the greater change.

The interval’s width matters as much as its midpoint. While the point estimate is close to zero, the interval includes values meaningfully below and above zero on the score scale. The registry states that a 10-point change score is considered clinically significant. That statement is about change scores, while the interval is for a difference between group mean changes. These quantities address related but distinct questions, so the threshold should not be used as if it were a prespecified equivalence margin for the between-arm comparison.

The arm standard deviations, 24.5 for teledermatology referrals and 21.6 for conventional referrals, indicate notable variation in change scores. The uncertainty of a mean comparison reflects both variability and the amount of endpoint information, not simply the distance between the two arm means. The primary outcome counts are 160 and 164, and the registry enrollment is 392; the reason these counts differ and the handling of incomplete outcome data are not reported in the registry record.

A confidence interval is not a guarantee that the true effect lies within its limits, and it does not assign probabilities to specific effects after the data have been observed. It is best read as a range of estimates reasonably compatible with the data under the analysis method. Here, that range leaves substantial uncertainty about both the direction and size of the between-arm difference. The p-values, 0.66 in the registry analysis and 0.64 in the unadjusted recalculation, are consistent with that cautious interpretation, but the interval gives more direct information about precision and possible effect sizes.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrative calculations based on the observed effect, not the trial’s original design assumptions; the original sample-size justification is not reported in the registry record. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect.

The 80% power column gives the hypothetical number of participants per arm needed for an 80% chance of detecting the specified effect under those assumptions, if that effect were present and the model assumptions held. The 90% power column gives the corresponding planning number for a 90% chance. These are prospective design concepts, not probabilities that the observed result is true or that a treatment works. They are not observed power.

The entries are 5,806 participants per arm for 100% of the observed effect at 80% power and 7,773 at 90% power. For 75% of the observed effect, the entries are 10,322 and 13,818 participants per arm. For 50% of the observed effect, they are 23,223 and 31,089 participants per arm. The pattern illustrates a general principle: smaller effects are harder to distinguish from random variation and therefore require more participants to detect at a given power and significance level.

These figures should not be interpreted as a recommended design or as proof that the original study was underpowered. Their inputs derive from the observed estimate, which itself is uncertain, and the registry does not report the trial’s own target effect, power goal, or planned sample-size method. They serve as conditional planning illustrations rather than a retrospective verdict on the study.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.200; pooled standard deviation σ = 23.078.
2. n = 2 × (1.96 + 0.842)² × 23.078² / -1.200² = 5,805.7, rounded up to 5,806 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.2                      # observed difference in means
sigma <- 23.0775784546211                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(488293)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 5,806 (80% power), 7,773 (90% power)
 75% of observed effect: n per arm = 10,322 (80% power), 13,818 (90% power)
 50% of observed effect: n per arm = 23,223 (80% power), 31,089 (90% power)
power.t.test: 5806.7 
Simulated power at n = 5,806 per arm: 0.803 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a two-sided t-test with p = 0.66. The unadjusted recalculation from the posted arm means, standard deviations, and endpoint counts gives p = 0.64. The recalculated comparison estimates a mean difference of -1.2, with a 95% confidence interval from -6.234 to 3.834. The registry record marks no discrepancy between the posted result and the recalculation.

The recalculation is a simple check based on the summary statistics available in the registry. It is not a substitute for the sponsor’s analysis or for an individual-participant-data analysis. The record does not report enough detail to determine whether the sponsor used additional adjustment, particular missing-data procedures, or other analysis choices beyond the stated t-test. The sponsor’s reported p-value should therefore be identified as the registry analysis, while the recalculated estimate and interval should be understood as unadjusted summary-statistic results.

Both p-values lead to the same broad statistical reading: the posted comparison does not provide clear evidence of a difference in mean Month 9 change. Neither result establishes that the referral processes are equivalent, and neither indicates the probability that a particular effect is true. The confidence interval is useful for seeing the uncertainty in the estimated contrast; the p-value alone does not show which effect sizes remain compatible with the data.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures and 10 analyses, including the primary Skindex-16 composite change score and secondary clinical course ratings between baseline and Month 9 and between baseline and the first clinic visit. When several outcomes or comparisons are assessed, the chance of obtaining at least one apparently unusual result can increase if each is tested separately without accounting for the number of tests. Whether and how multiplicity was handled is not reported in the registry record. The primary result should therefore be distinguished from exploratory or secondary findings, and no adjustment strategy should be assumed.

The posted primary comparison gives outcome counts of 160 for teledermatology referrals and 164 for conventional referrals. Registry enrollment is 392. The record does not explain the difference between enrollment and the counts contributing to the primary endpoint, identify reasons for unavailable scores, or state how missing outcome data were handled. It does not report whether participants with incomplete scores were excluded, imputed, or addressed by another method. The effect of missing data on the comparison cannot be assessed from the posted summary alone.

The analysis population is also not specified. The registry does not say whether the primary result followed the randomized assignment regardless of consultation received, included only participants who completed the assigned process, or used another population definition. Without that information, the relationship between randomization and the analyzed endpoint sample cannot be fully evaluated.

Baseline score distributions, adjusted estimates, subgroup analyses, and sensitivity analyses are not reported in the registry record. These details could help assess comparability and robustness, but their absence does not justify assuming a particular bias or analysis choice. The most defensible interpretation is limited to the posted group means, standard deviations, counts, confidence interval, and registry-reported test, with the unreported handling of multiplicity and incomplete outcomes kept in view.

15. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: death unrelated to study participation (1.0%).

The registry lists serious adverse events for 196 participants at risk in each arm. It records 2 participants affected in the teledermatology referrals group and 1 in the conventional referrals group. For other adverse events, the record lists 1 participant affected among 196 at risk in the teledermatology group and 2 among 196 at risk in the conventional group. These are the recorded counts; they should not be converted into additional rates or interpreted as a complete account of safety without further definitions.

The most frequent serious adverse event term listed is “Death Unrelated to Study Participation,” at 1.0%. The deaths fields in the arm-level safety summary are null, so the registry does not provide arm-specific death counts in that field. The listed event term explicitly describes the death as unrelated to study participation; it should not be reclassified as an event caused by a referral process.

The serious and other adverse-event figures are descriptive counts. They do not establish that the groups had different underlying safety risks, particularly without details about event definitions, collection procedures, follow-up completeness, or statistical comparisons. Those details are not reported in the registry record. The adverse-event denominators are 196 per arm, which differ from the primary outcome counts of 160 and 164; the registry does not explain this difference. The safety figures should therefore be read as the registry’s recorded counts and not combined with the primary endpoint results.

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

The primary result describes average change in patient-reported quality of life between baseline and Month 9. The mean change was -12.0 in the teledermatology referrals group and -13.2 in the conventional referrals group. Since negative change represents better quality of life, both arm means are in the direction of reduced reported bother. The comparative estimate is -1.2 for conventional referrals relative to teledermatology referrals, with a 95% confidence interval from -6.234 to 3.834.

The estimate is compatible with a small average difference, but the uncertainty interval also allows differences in either direction. The result does not establish that one referral modality produced a greater improvement. Nor does the absence of a clear statistical difference prove equivalence or show that the modalities have identical effects. The registry does not report an equivalence margin or a noninferiority analysis.

The result also does not describe every participant’s experience. Mean change and standard deviation summarize groups, and individual scores can vary around those summaries. The registry’s 10-point clinical-significance statement applies to a change score; it is not itself a test of the difference between group averages. Interpreting a mean contrast against that benchmark requires care because the contrast and the individual change measure are not the same quantity.

The finding is limited to the posted primary composite outcome and its recorded analysis. The record does not report missing-data methods, analysis-population details, adjusted results, or multiplicity handling. The endpoint counts are smaller than the stated enrollment, with no explanation of that accounting in the registry record. As a result, the estimates should be read with uncertainty about how the analyzed endpoint sample was formed.

Overall, the posted results show similar average changes in the two groups, while leaving meaningful uncertainty about the between-group effect. They do not by themselves establish comparative effectiveness across other outcomes, settings, or populations. They also do not answer questions about implementation, service use, or clinical course beyond the outcomes specifically listed in the registry.

17. Why This Trial Matters Statistically

This trial raises a useful statistical question about how to compare care pathways when the primary outcome is patient-reported and continuous. The mean change offers a direct summary of average movement in Skindex-16 composite score, while the standard deviations show that individual changes vary considerably. Reporting both the group summaries and a between-arm interval helps prevent the point estimate from being mistaken for a complete description of the data.

The result also illustrates why a nonsignificant comparison should not be treated as proof that two approaches are equivalent. The estimated difference is -1.2, but the 95% confidence interval ranges from -6.234 to 3.834. It includes zero and values on either side, so the direction and size of a possible difference remain uncertain. Equivalence would require a prespecified margin and a design and analysis aimed at that question; the registry does not report these.

Another statistical issue is the distinction between outcome counts and enrollment. The registry gives enrollment as 392, while the primary endpoint arms list 160 and 164. Safety denominators are listed separately as 196 per arm. Without an explanation of these populations or the missing outcome process, readers cannot fully assess the path from randomized enrollment to analyzed data. Clear reporting of analysis populations and missing-data handling is important for understanding what an estimate represents.

Finally, the planning illustrations demonstrate how required sample size depends on the effect a study aims to detect. Smaller effects require more participants under the stated assumptions. Because the calculations use the observed effect and do not report the trial’s original design assumptions, they are best viewed as conditional planning illustrations rather than a retrospective assessment of design quality. Together, these points make the study useful for discussing estimation, precision, analysis populations, and the limits of interpreting a single p-value.

18. 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 reporting limitations constrain interpretation. Although the registry states random allocation, it does not report the sequence-generation method or allocation concealment. Masking is listed as none, but the record does not clarify which groups of people knew the assigned referral process or whether outcome assessment could be masked.

The registry reports enrollment of 392, while the primary outcome comparison lists 160 teledermatology referrals and 164 conventional referrals. The safety summary lists 196 at risk in each arm. The record does not explain these differing counts, give a participant flow account, or report the analysis population for the primary outcome. It also does not describe the amount, reasons, or handling of missing Skindex-16 data. These omissions limit assessment of whether the outcome comparison reflects the randomized groups and how sensitive it may be to incomplete follow-up.

The endpoint is a patient-reported quality-of-life measure, but further details on questionnaire completion, scoring rules for incomplete responses, and timing windows are not reported in the registry record. The record does not provide baseline outcome summaries, adjusted estimates, or sensitivity analyses. It lists multiple outcomes and analyses but does not report whether any multiplicity procedure was used.

The reported primary comparison is an unadjusted summary-statistic recalculation alongside the registry’s two-sided t-test. The registry does not provide the full statistical analysis plan or enough detail to assess all modeling decisions. It also does not report the original sample-size justification or target effect. These are reporting limitations; they do not by themselves establish that the results are biased, but they restrict how completely the analysis can be evaluated.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Impact of Teledermatology on Health Services Outcomes in the Department of Veterans Affairs — NCT00488293.
• Registry record last updated 2015-06-08; status last verified 2015-05.
• Results first posted 2014-12-15.
• 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.

22. Record Summary

The registry record describes a completed randomized, parallel-group study with no masking, comparing store-and-forward teledermatology referrals with conventional referrals. Enrollment is listed as 392. The primary outcome is Skindex-16 composite change from baseline to Month 9, where negative change represents better quality of life.

Mean change was -12.0 among teledermatology referrals and -13.2 among conventional referrals. The estimated mean difference for conventional referrals relative to teledermatology referrals was -1.2, with a 95% confidence interval from -6.234 to 3.834. The registry reports a two-sided t-test with p = 0.66; the unadjusted recalculation gives p = 0.64.

The interval includes zero and values in either direction, leaving uncertainty about the size and direction of any average difference. The record does not report key details about the original sample-size justification, missing outcomes, analysis populations, or multiplicity handling. The posted result is best read as an imprecise comparative estimate rather than evidence of equivalence.