Eliminating Risk of Preventable Adverse Drug Events at the Hospital-community Interface of Care: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Eliminating Risk of Preventable Adverse Drug Events at the Hospital-community Interface of Care (NCT01164137), a randomized na study. The primary outcome is Health Services Utilization 3 Months Following Hospital Discharge.

2. Clinical Question

The registry record describes a randomized comparison, so the assigned intervention groups are the basis for the posted comparison. It reports mean utilization of 3.33 in the intervention group and 3.4 in the control group. The estimated mean difference, intervention minus control, is -0.07. These are group-level results for a composite of listed service-use categories; the record does not report separate estimates for each category.

The question is about health services utilization, not a direct estimate of medication discrepancies or adverse drug events. Although the initiative’s broader aim mentions medication errors and health system utilization, the listed primary endpoint is utilization during the three months following discharge. The registry record does not report an outcome definition that breaks the overall measure into its component events, or details about how utilization was ascertained.

3. Trial Design

This was a randomized, parallel-group trial with no masking. The registry record lists an enrollment of 156 and a completed status. The study began in 2008-11, and primary completion is listed as 2013-02. The phase is recorded as NA, and the sponsor is Westview Physician Collaborative.

Participants were assigned to either a medication reconciliation intervention or a non-intervention control. The intervention was a pharmacist-led home visit within 72 hours of hospital discharge, intended to identify and correct medication discrepancies. The control description specifies that participants did not receive this home intervention. The registry record does not describe the other services or usual care available to either group, nor does it provide further details about the randomization procedure, allocation concealment, or how the intervention was delivered across participants.

The primary analysis is a comparison of mean health services utilization three months after discharge. The posted group summaries include 77 participants in the intervention arm and 76 in the control arm. These are the numbers associated with the posted primary outcome summaries; the registry record does not explain how these analysis counts relate to the enrollment figure or report participant flow, exclusions, or the number with missing endpoint data.

The study was unmasked, which means masking was not listed for participants, care providers, or outcome assessors. The record does not specify which roles, if any, were masked in practice. It also does not report the original sample-size justification, the planned analysis population, a detailed analysis plan, interim analyses, or handling of missing observations. These details are relevant when interpreting a randomized trial, but their absence from the registry record does not establish what procedures were or were not used.

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: Three participants were excluded from the trail before assignment to groups because they obtained a score of 19 or less on the Mini Mental State Examination (MMSE).

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 Health Services Utilization 3 Months Following Hospital Discharge, measured over 3 Months.

The primary endpoint is mean health services utilization during the three months following hospital discharge. The registry record names four kinds of utilization: hospital readmission, admission to home care, admission to long-term care, and visits to emergency departments. The reported unit is “Health Services Utilizations,” and the endpoint is described as a mean of a continuous measure.

A utilization measure can capture how often participants use a range of services after leaving hospital. Combining service categories gives a broad summary of subsequent health system use, but it also means the overall mean does not reveal which category contributed most to the result. The registry record does not report category-specific means or comparisons, nor does it explain whether the categories were counted in the same way or how repeat events were handled.

For the three-month endpoint, the intervention group mean is 3.33 and the control group mean is 3.4. The registry record reports standard errors of 0.35 and 0.29, respectively. The comparison therefore describes average utilization across the listed measure; it is not itself a direct measure of medication discrepancy resolution or adverse drug events. The record does not report a separate primary result for those outcomes.

Using utilization as an endpoint can address a service-use question that is distinct from whether a medication reconciliation process changes medication-related outcomes. Interpretation depends on the endpoint’s operational definition and completeness of ascertainment. Those measurement details are not reported in the registry record.

7. Results Posted in the Registry

• Experimental: Medication Reconciliation Intervention: mean 3.33 versus 3.40 on control; difference -0.07 (95% CI -0.96 to 0.82), standardized effect d = -0.02; z-test p = 0.878.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous and is summarized by a mean for each group. For the intervention group, the reported mean is 3.33 with a standard error of 0.35, based on 77 participants. For the control group, the reported mean is 3.4 with a standard error of 0.29, based on 76 participants. A mean describes the average value of the endpoint within a group; it does not show how values are distributed among individual participants.

The standard error describes the uncertainty in an estimated group mean due to sampling variability. It is not the same as the standard deviation, which describes spread among individual observations. The registry record labels the reported dispersion as standard error. The unadjusted recalculation also includes a pooled standard deviation of 2.815, but that is a recalculation from the posted summaries rather than the trial’s stated analysis result. The registry record does not provide the underlying participant-level distribution.

The unadjusted group comparison is expressed as a mean difference: intervention mean minus control mean. The estimate is -0.07. A mean difference retains the endpoint’s original units, making it a direct comparison on the health services utilization scale. A negative estimate indicates a lower mean in the intervention group than in the control group; it does not on its own establish that the difference is reliable or important.

What does the two-sided t-test assess?

The registry reports a two-sided t-test comparing final values between the groups. In general, a t-test evaluates whether the observed difference in means is compatible with a null hypothesis of no difference, while accounting for estimated sampling variability. A two-sided test allows evidence in either direction: the intervention mean could be higher or lower than the control mean. The registry reports a mean difference of -.07, a 95% confidence interval from -.98 to .84, and p = .87.

The recalculation is an unadjusted comparison of the posted summary statistics. It reports p = 0.878. It does not incorporate participant-level covariates, a baseline measure, or any adjustment not present in the posted summaries. The record does not report whether the sponsor’s analysis used additional adjustments beyond the description of a t-test on final values.

Why do confidence intervals accompany a test?

A confidence interval expresses uncertainty about the magnitude and direction of an estimated effect. For a mean difference, the interval gives a range of effect values compatible with the data and the method’s assumptions at the stated confidence level. The sponsor’s 95% interval extends from -.98 to .84 and includes zero. Its width indicates that the estimate is not highly precise. It should not be read as a range containing a fixed percentage of individual participant outcomes, or as a probability statement about the true difference after observing the data.

A p-value and a confidence interval answer related but different questions. The p-value summarizes evidence against a null hypothesis under a statistical model. The interval displays a range of plausible effect sizes at a stated confidence level. Neither statistic alone establishes clinical importance. The registry record does not define a smallest important difference for the utilization measure.

What assumptions accompany an unadjusted comparison?

General statistical interpretation of a t-test depends on features such as the independence of observations and the behavior of the outcome distribution, particularly when sample sizes are limited or the outcome is markedly skewed. Utilization measures can, in general, have uneven distributions because many participants may have few events while some have more. This is a general consideration, not evidence that the distribution in this trial had any particular shape. The registry record does not report distributional diagnostics, robust methods, or an alternative model for the endpoint.

Over-dispersion is a general issue for count-like outcomes when observed variability exceeds what a simpler model assumes. If a utilization measure is modeled as a count, over-dispersion can make standard errors from an overly restrictive model too small and confidence intervals too narrow. The posted primary analysis is described as a t-test of means, not a count-regression model. The record does not report a separate assessment of over-dispersion, and the endpoint’s detailed construction is not reported.

Random assignment is intended to make groups comparable on average, but the registry record does not provide baseline characteristics or details about allocation concealment. The posted means are not adjusted for any listed baseline variables in the recalculation. Without an analysis plan, it is not possible to determine whether other prespecified analyses were intended or how any departures from a plan were handled.

Finally, the inferential result depends on outcome completeness and the population analyzed. The registry record does not report missing-data procedures, participant flow, or whether the posted comparison follows an intention-to-treat principle. These are not minor technical details: they determine which observations contribute to the mean difference and how uncertainty is calculated. The results therefore support interpretation of the posted group summaries and test, while leaving those aspects of the trial’s analysis process unspecified.

10. Confidence Intervals and Precision

The sponsor-reported 95% confidence interval for the mean difference runs from -.98 to .84, with the estimate at -.07. The interval includes zero and extends to values on both sides of it. Thus the estimate’s direction is uncertain: the data are compatible with a lower mean in the intervention group, little or no difference, or a higher mean in that group, within the interval and under the analysis assumptions.

The interval is broad relative to the point estimate. That means the observed difference should not be treated as a precise estimate of the size of any group difference. The p-value of .87 likewise does not establish that the true difference is zero. It indicates that this test did not provide strong evidence against its null hypothesis, but the interval shows why that should not be converted into a claim of equivalence: a range of nonzero values remains compatible with the reported analysis.

Confidence intervals concern uncertainty in an estimated group contrast, not the range of participant-level utilization values. They also do not account automatically for every possible source of uncertainty, such as incomplete outcome measurement or missing observations. The registry record does not report details that would allow those issues to be assessed.

The unadjusted recalculation and the sponsor’s reported estimate are both -0.07, while their reported p-values are 0.878 and .87, respectively. The registry’s interval is the sponsor-reported interval. The comparison should be read in light of the analysis method and the information reported in the record, not as evidence that all plausible effects are equally likely or that a clinically meaningful threshold has been excluded.

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 values are participants per arm, not total events. They are illustrations based on the observed effect, rather than a report of the trial’s original sample-size calculation. The registry record does not report the trial’s own sample-size justification or original design assumptions.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set as a stated fraction of the observed effect. For planning at 100% of the observed effect, the listed values are 25,380 participants per arm for 80% power and 33,976 participants per arm for 90% power. At 75% of the observed effect, they are 45,120 per arm for 80% power and 60,402 per arm for 90% power. At 50% of the observed effect, they are 101,518 per arm for 80% power and 135,904 per arm for 90% power.

In plain terms, the 80% power column gives the illustrative number of participants per arm under which the planning calculation would have an 80% chance of detecting the specified effect, if the assumptions and effect were correct. The 90% power column describes a higher target chance of detection under those same assumptions and therefore requires more participants per arm. As the assumed effect becomes smaller, distinguishing it from random variation becomes harder, so the planning calculation requires more participants.

These values are not observed power. They are prospective planning illustrations calculated from the observed effect, not a description of the study’s actual design rationale. The effect estimate is close to zero, so these figures should be understood as consequences of planning around that estimate and the stated assumptions, rather than as evidence about what sample size the investigators planned before conducting the trial.

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 = -0.070; pooled standard deviation σ = 2.815.
2. n = 2 × (1.96 + 0.842)² × 2.815² / -0.070² = 25,379.5, rounded up to 25,380 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 <- -0.0699999999999998                      # observed difference in means
sigma <- 2.81462594411727                     # 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(1164137)
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 = 25,380 (80% power), 33,976 (90% power)
 75% of observed effect: n per arm = 45,120 (80% power), 60,402 (90% power)
 50% of observed effect: n per arm = 101,518 (80% power), 135,904 (90% power)
power.t.test: 25380.5 
Simulated power at n = 25,380 per arm: 0.801 (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’s reported analysis is a two-sided t-test of the mean difference in final values. It reports a mean difference of -.07, a 95% confidence interval from -.98 to .84, and p = .87. The unadjusted recalculation from the posted summary statistics gives a mean difference of -0.07 and p = 0.878.

The estimates agree at the precision shown for the recalculated mean difference, and the p-values are close. The registry record does not flag a discrepancy. The sponsor’s reported analysis remains the primary inferential result because it is the analysis posted for the trial; the recalculation is a check based on summary values and should not be treated as a replacement for the trial’s analysis.

The record describes the sponsor’s result as a comparison of final values, but does not provide participant-level data or a fuller statistical plan. It is therefore not possible to independently reproduce the sponsor’s test or determine whether it incorporated analysis details not represented in the posted summaries. The reported confidence interval and test should be interpreted as the sponsor’s stated analysis, while recognizing that important details about missing data and analysis populations are not reported in the registry record.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists five outcome measures and five analyses, but it reports no secondary outcomes in the posted results. The primary comparison is the mean health services utilization measure at three months. The record does not report how the other listed measures or analyses were defined, whether they were prespecified, or what results they produced.

Multiplicity arises when several outcomes, time points, subgroups, or statistical tests are considered. In general, making more comparisons can increase the chance that at least one result appears statistically notable by chance. Methods such as adjusting significance thresholds or defining a testing hierarchy can address that issue, but the registry record does not report a multiplicity strategy. It is not possible to infer whether any adjustment was planned or used.

The record does not report missing-data methods, the number of participants with observed primary endpoint data, or participant flow by reason for non-completion. The outcome summary counts are 77 for the intervention group and 76 for control; the record does not explain missingness or how those counts were determined. It also does not state whether missing values were imputed, whether complete cases were analyzed, or whether sensitivity analyses were performed. A missing-data approach can affect both the estimated difference and its uncertainty, so this omission limits assessment of the robustness of the result.

The analysis population is not reported in the registry record. In particular, the record does not state whether the primary analysis included all randomized participants according to assigned group, only participants who completed follow-up, or another defined population. Nor does it report details of adherence to assignment or protocol deviations. Randomization establishes an assignment framework, but conclusions about the analysis depend on how participants and outcome observations were handled after assignment.

The trial’s original statistical analysis plan, sample-size justification, and interim-analysis procedures are not reported in the registry record. No assumptions about those procedures should be inferred from the posted test or the planning illustrations. These gaps do not change the arithmetic of the reported comparison, but they make it harder to assess whether analyses were selected in advance and how complete the primary analysis was.

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.

The serious adverse event listing records 0 serious events in the intervention group and 0 in the control group. It lists 80 participants at risk in the intervention group and 76 in the control group for this safety measure. The listing also records 0 other adverse events in each group, with 80 and 76 participants at risk, respectively.

These figures describe what is recorded in the registry’s adverse-event fields. They do not, by themselves, establish that no adverse events occurred outside the categories or reporting procedures captured in the record. The registry record does not provide serious adverse event terms, event narratives, severity details, or a fuller description of collection and attribution methods. It also does not report deaths in these fields.

The safety listing’s intervention group count at risk is 80, while the primary outcome summary includes 77 participants in that arm. The registry record does not explain the difference between these counts. Safety and efficacy denominators can differ for various reasons, but no explanation is reported and none can be inferred. The adverse-event counts should therefore be read with their stated at-risk denominators and the limitations of the registry record in mind.

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

The posted results estimate a mean health services utilization difference of -0.07 between the intervention and control groups over the three-month period after hospital discharge. The intervention group mean is 3.33 and the control group mean is 3.4. The sponsor’s 95% confidence interval, from -.98 to .84, includes zero and allows for differences in either direction. The reported p-value is .87.

These results do not provide strong statistical evidence of a difference in the primary endpoint under the reported test. At the same time, they do not show that the groups are equivalent, nor do they rule out all nonzero differences. The registry record does not define a margin for equivalence or identify a threshold at which a difference in utilization would be considered important. The interval’s width also means that the estimate is not especially precise.

The endpoint is an overall health services utilization measure that includes hospital readmission, home care admission, long-term care admission, and emergency department visits. Because separate category results are not reported, the mean difference does not establish what happened for any one category. It cannot be interpreted as a direct measure of medication discrepancies, medication errors, or adverse drug events.

The intervention was a pharmacist-led home medication reconciliation visit within 72 hours of discharge, compared with no such home visit. The result pertains to that randomized comparison and the reported endpoint. The registry record does not report enough detail about other care, endpoint ascertainment, participant flow, missing data, or analysis populations to assess all potential sources of bias or uncertainty.

Accordingly, the result is best understood as an imprecise estimate of the difference in mean recorded utilization, rather than as a conclusive answer about medication safety or every form of post-discharge health service use. It neither establishes a benefit nor demonstrates that the intervention had no effect.

17. Why This Trial Matters Statistically

This trial illustrates why an estimated effect, its uncertainty, and the endpoint’s meaning need to be considered together. The mean difference of -0.07 is near zero, but the confidence interval from -.98 to .84 is broad enough to include values on both sides of zero. A non-significant test is not the same as a precise estimate of no difference, and it is not a formal equivalence result.

It also highlights the distinction between a study’s motivating problem and its measured primary outcome. The initiative was intended to address medication errors and health system utilization after discharge, while the posted primary endpoint is mean utilization across several listed service categories. That endpoint can inform a service-use question, but it does not alone tell readers whether medication discrepancies or adverse drug events changed. Separate outcome results would be needed to answer those questions, and they are not reported in the registry record.

For biostatisticians and clinical researchers, the record is a reminder to inspect denominators, definitions, and analysis populations, rather than relying only on a point estimate and p-value. The safety fields list 80 participants at risk in the intervention arm, while the primary outcome summary lists 77. The registry record does not explain the difference. Details of missing data, participant flow, multiplicity, and the original statistical plan are also not reported.

The planning illustrations demonstrate how a small observed effect can imply large sample-size requirements when planning around that effect under stated assumptions. They should not be confused with the study’s original design calculation or with evidence that the trial had a particular power. Together, the endpoint and precision considerations show how registry results can support a focused statistical interpretation while leaving important design and measurement questions unresolved.

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 limitations follow from what is not reported in the registry record. The endpoint is described as a mean of health services utilization, but its construction is not detailed. The four included categories are named, yet no category-specific results are reported, and the record does not explain how repeated service use was counted or ascertained. This limits interpretation of the aggregate mean.

The record does not describe participant eligibility in detail, baseline characteristics, or the distribution of the primary endpoint. Without those details, readers cannot assess the enrolled population’s characteristics or whether the mean summarizes a skewed or otherwise uneven distribution. The unmasked design is recorded, but the roles of participants, clinicians, and outcome assessors are not specified. Utilization may be based on records or reports, but the ascertainment method is not reported, so no assumption should be made about it.

The primary summary lists 77 intervention participants and 76 controls, while the safety listing has 80 intervention participants at risk and 76 controls. The registry record does not explain this difference. It also does not report participant flow, missing outcome data, the analysis population, adherence, or protocol deviations. These omissions prevent a full assessment of whether the posted groups and denominators represent all randomized participants and how incomplete follow-up may have influenced the comparison.

The analysis description is limited to a two-sided t-test of final values. The registry record does not report the original sample-size rationale, a full statistical analysis plan, multiplicity handling, interim analyses, or prespecified thresholds for clinical importance. The planning figures are illustrations based on the observed estimate and do not fill those gaps. Finally, secondary posted results are absent, so the primary utilization finding cannot be placed alongside reported estimates for the other outcome measures.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Eliminating Risk of Preventable Adverse Drug Events at the Hospital-community Interface of Care — NCT01164137.
• Registry record last updated 2013-02-15; status last verified 2013-02.
• Results first posted 2013-01-30.
• 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, unmasked trial of a pharmacist-led home medication reconciliation intervention after hospital discharge. Enrollment is listed as 156; the intervention was intended to occur within 72 hours of discharge. The primary endpoint was mean health services utilization during the three months after discharge, including hospital readmission, home care admission, long-term care admission, and emergency department visits.

The posted means are 3.33 for the intervention group and 3.4 for control, with a mean difference of -0.07. The sponsor reports a 95% confidence interval from -.98 to .84 and p = .87. The interval includes zero and is broad, so the estimate is uncertain and does not establish equivalence. Secondary results are not posted. Details about missing data, analysis populations, endpoint ascertainment, and the trial’s original statistical plan are not reported in the registry record.