ALIGN: Aligning Medications With What Matters Most (Demonstration): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for ALIGN: Aligning Medications With What Matters Most (Demonstration) (NCT06335953), a randomized na study. The primary outcome is Patients Who Stop One or More Medications - Per Protocol.

2. Clinical Question

3. Trial Design

ALIGN is a completed, randomized, parallel-group clinical trial sponsored by Johns Hopkins University. The registry lists no masking. Enrollment is 474, and the record gives a start date of 2024-05-07 and a primary completion date of 2025-09-16. The record describes the study as a demonstration of a pharmacist-led, primary care-based deprescribing intervention for people living with dementia and their care partners.

The intervention includes three strategies: a deprescribing educational brochure intended to activate the patient and care partner; one telehealth discussion with an embedded clinical pharmacist about medication benefits and harms in the context of goals and preferences; and pharmacist-to-primary-care-clinician communication with tailored deprescribing recommendations. The comparator is described as a delayed intervention or waitlist control. Further details about the control period and the timing of the delayed intervention are not reported in the registry record.

Although the design fields describe randomization and parallel groups, the registry's analysis note identifies the trial as cluster-randomized. The cluster unit and randomization procedures are not reported in the registry record. This matters for statistical interpretation: participants assigned within the same cluster may have correlated outcomes, so treating all participants as independently randomized can misstate uncertainty. The analysis note therefore says that a simple individual-level re-analysis would be misleading and that only the sponsor's own analysis is shown.

The record identifies a per-protocol primary analysis at 3 months. The primary endpoint is the number of participants who stopped at least one medication, with discontinuation determined from medication lists in the electronic health record. Protocol details, the intended sample-size justification, and further information on analysis populations 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.

Before assignment: People living with dementia (PLWD) were enrolled independently or together with a care partner. A total of 360 PLWD patients were enrolled and 114 care partners, with available data obtained from the EHR only for the patients. Pharmacists and primary care physicians (PCPs) delivered the intervention and were not considered enrolled study participants.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Per protocol analysis

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Baseline characteristics were collected from the electronic health record (EHR) for PLWD patients only. Baseline data were not collected from care partners.

6. Primary Endpoint

The registry lists the primary outcome as Patients Who Stop One or More Medications - Per Protocol, measured over 3 months.

The primary endpoint counts participants who stopped one or more medications by the 3-month assessment. A medication is considered discontinued when it appears on a medication list at one time point but not on the subsequent list. The registry says the medication information comes from the electronic health record, and that the endpoint is analyzed per protocol.

This participant-level outcome asks whether a person had at least one discontinuation, rather than how many medications were stopped in total. As a general statistical point, a binary endpoint can be summarized by the proportion of participants meeting the definition in each group and compared using a relative or absolute measure. The reported analysis uses a risk ratio. The registry does not provide the underlying group counts or proportions for this primary endpoint, so those cannot be reconstructed from the reported risk ratio alone.

Medication-list changes are the operational definition of discontinuation in this record. The registry does not report additional rules for distinguishing a planned stop from a medication omitted for another reason, nor does it describe how conflicting or incomplete medication lists were handled. Such details can affect how consistently the endpoint reflects medication discontinuation. The endpoint is used to assess a change in medication status, not to establish why a medicine was stopped or whether the change improved a health outcome.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is reported as a risk ratio from a mixed models analysis. The endpoint is binary at the participant level: each person either did or did not stop at least one medication by 3 months under the registry definition. The analysis population is per protocol. The registry identifies the study as cluster-randomized in its analysis note, which is an important part of understanding the model and the uncertainty around the estimate.

What does the risk ratio describe?

As a general statistical explanation, a risk ratio is the risk of an outcome in one group divided by the risk in another group. A risk ratio of 1 indicates equal risks; a value above 1 indicates a higher risk in the numerator group, while a value below 1 indicates a lower risk in that group. The registry reports 1.45, but its result entry uses group identifiers rather than naming the numerator and denominator. It is therefore not possible to attribute the direction to a named trial arm based only on that entry. The risk ratio is relative: without the group-specific proportions, it does not show how many additional or fewer participants experienced the outcome in absolute terms.

Why is a confidence interval useful?

In general, a confidence interval presents a range of estimates compatible with the observed data and the statistical model at a stated confidence level. The reported interval is 1.09 to 1.92 at the 95% level. It conveys uncertainty more fully than the point estimate alone. It is not a range that contains a fixed percentage of individual participant outcomes, and it does not indicate that every value inside the interval is equally likely. Its interpretation depends on the correctness of the model and design assumptions, including appropriate handling of clustering.

What does the p-value add?

The reported result is p < 0.05. As a general explanation, a p-value summarizes how unusual data at least as incompatible with a null hypothesis would be if that null model and the analysis assumptions held. It is not the probability that the treatment effect is absent, nor does it measure the effect's importance. Because the registry gives only an inequality, the exact p-value is not reported in the registry record. The p-value and confidence interval describe related aspects of statistical evidence, but neither supplies the omitted group-level counts.

Why does the cluster design matter?

In general, cluster randomization assigns groups of people rather than individuals, and outcomes within a cluster may be more alike than outcomes across clusters. This within-cluster correlation reduces the amount of independent information in a given number of participants. A model that treats every participant as independently randomized may therefore give uncertainty estimates that are too small. The registry's analysis note explicitly warns that a simple individual-level re-analysis would be misleading and indicates that only the sponsor's own analysis is shown. The cluster unit, number of clusters, and details of how the mixed model accounted for clustering are not reported in the registry record.

Mixed models can, in general, represent correlated observations through model structure such as random effects, depending on the particular model specified. The name “Mixed Models Analysis” alone does not identify the link function, covariates, correlation structure, or estimation procedure. Those specifications are not reported in the registry record, so the analysis cannot be independently reproduced from the registry result.

How should the analysis population be understood?

Per-protocol analyses generally include participants who meet criteria for following the assigned intervention sufficiently to be analyzed under the protocol definition. As a general statistical consideration, this can answer a different question from an intention-to-treat analysis, which retains participants according to randomized assignment. Per-protocol analyses can also be affected by differences between those who do and do not adhere or remain eligible. The registry identifies the primary population as per protocol but does not report its operational definition, the number included, exclusions, or an intention-to-treat result for the primary endpoint.

The record lists 10 outcome measures and 10 analyses, while the reported primary result is one comparison. In general, testing multiple outcomes or hypotheses can increase the chance that at least one result appears statistically notable by chance. Whether and how multiplicity was addressed is not reported in the registry record. The reported p-value should therefore be read as the result of the stated analysis, not as evidence that all possible multiplicity concerns were addressed.

The record also includes a planning-illustration description based on an observed effect, with assumptions of a two-sided significance level of 0.05, equal allocation, and a normal approximation. No calculated planning values appear in the trial details for discussion. In general, such sample-size illustrations are hypothetical planning exercises and do not represent observed power or a new estimate of trial effectiveness. They cannot substitute for the trial's design rationale, which is not reported in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Intervention", "Delayed Intervention (Waitlist Control)")
n      <- c(146, 146)
events <- c(103, 73)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Intervention  146  103  70.5  62.4  77.8  62.7  77.3  62.7  77.3
ARM  Delayed Intervention (Waitlist Control)  146  73  50.0  41.6  58.4  42.0  58.0  42.0  58.0

11. Analyses Reported by the Sponsor

The primary result comes from the sponsor's mixed models analysis: a risk ratio of 1.45, a 95% confidence interval from 1.09 to 1.92, and p < 0.05. The registry's analysis note says the trial is cluster-randomized and warns that a simple individual-level re-analysis would be misleading. For that reason, interpretation should rely on the sponsor's own analysis rather than on a recalculation that treats participants as independently randomized.

No simple recalculated estimate is reported for the primary outcome. The registry does not provide the arm-specific outcome counts needed to independently reconstruct the risk ratio from summary counts, and it does not provide all details of the mixed model. The group identifiers in the result are not mapped to the named intervention and waitlist groups in that entry, so the direction of the relative comparison cannot be established from the result alone.

The sponsor's model-based result and a basic calculation from participant counts would not necessarily answer the same statistical question in a cluster-randomized trial. As a general point, accounting for within-cluster correlation can change the estimated uncertainty, even where the underlying outcome counts are unchanged. The registry does not report enough model detail to reproduce the sponsor's estimate or assess its specific assumptions. Thus, the result can be described as posted, but not independently verified from the registry's summary information.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures and 10 analyses. The primary endpoint is patients who stop one or more medications, analyzed per protocol at 3 months. The secondary outcomes include starting new medications, stopping or starting potentially inappropriate medications, and total number of medications; several are listed under both per-protocol and intention-to-treat populations. The registry does not report results for those secondary outcomes in the analysis information summarized for the primary result.

Whether the trial specified a hierarchy among outcomes, selected a single primary hypothesis for formal testing, or adjusted for multiple comparisons is not reported in the registry record. In general, when several outcomes or comparisons are tested, the chance of at least one apparently notable result can increase. Adjustment methods and prespecified testing strategies can address that issue, but no such approach should be assumed. The p-value for the primary result is therefore interpreted as the reported result of that analysis, without claiming that multiplicity was or was not controlled.

The record identifies the primary analysis as per protocol, but the definition of that population, the number included, and the reasons participants may have been excluded are not reported in the registry record. The record names intention-to-treat versions of some secondary outcomes, but does not provide a primary intention-to-treat result. Missing outcome data, imputation procedures, and sensitivity analyses are also not reported in the registry record.

These gaps matter because analysis-population choices and missing-data assumptions can affect an estimate and its precision. A per-protocol result does not automatically describe the effect among everyone randomized, and the enrollment total alone does not show how many contributed to the primary analysis. The record does not report enough information to evaluate attrition, compare included and excluded participants, or determine whether conclusions are robust to alternative missing-data assumptions.

13. 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.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: ed visit (3.7%), hospitalization (0.5%).

For serious adverse events, the registry reports 9 affected participants among 214 at risk in the intervention group and 0 among 146 at risk in the delayed-intervention waitlist control group. Deaths are reported for 2 participants in the intervention group and 0 in the waitlist group. The “other adverse events” counts are 0 among 214 and 0 among 146, respectively. These are descriptive counts and do not, by themselves, establish that the groups differ in safety.

The registry also lists emergency department visit at 3.7% and hospitalization at 0.5% among the top serious adverse event terms. The record does not provide the event counts or clarify the denominator for those percentages in the information available for this summary. Those terms should not be added together or treated as distinct participants without further detail: one participant may have more than one event, and the record does not explain how terms were counted.

As a general statistical point, a zero count in a group does not prove that the underlying event risk is zero. Interpretation depends on the number followed, event definitions, observation time, and completeness of reporting. The registry does not report enough detail about event timing, relatedness, or comparative safety methods. These adverse-event figures should be read as reported observations, not as a causal comparison or a treatment recommendation.

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

The posted primary result describes a relative difference in the probability of stopping one or more medications at 3 months under a per-protocol analysis. The risk ratio is 1.45, with a 95% confidence interval from 1.09 to 1.92 and p < 0.05. The result is from the sponsor's mixed models analysis, and the registry notes that the trial is cluster-randomized. The interval communicates uncertainty around the relative estimate; it does not specify an absolute change in the proportion of participants.

The result does not establish the direction of the comparison as displayed, because the reported analysis uses group identifiers without stating which is the numerator or denominator. It also does not provide the number of participants in each named group who stopped one or more medications. The registry's primary outcome definition is based on medication lists, so it establishes whether a medication is absent at a subsequent time point, not the reason for the change or whether it was consistent with the person's preferences.

The reported finding is specific to the per-protocol analysis and cannot automatically be extended to all randomized participants. A primary intention-to-treat result, outcome counts by group, the per-protocol sample size, and missing-data methods are not reported in the registry record. The mixed model's detailed specifications and the cluster unit are also not reported, limiting independent reproduction.

Nor does this single endpoint establish effects on other listed outcomes, such as starting medications or total medication count. The registry lists those outcomes, but their results are not part of the reported primary analysis. The serious adverse event counts are descriptive and do not establish a causal safety difference.

In general, statistical evidence about one endpoint should be separated from claims about broader health effects or clinical value. The registry result concerns medication-list discontinuation at a defined follow-up time. It does not establish the clinical consequences of stopping a medication, and it does not on its own show that the intervention should be adopted in other settings.

15. Why This Trial Matters Statistically

This trial raises a statistical question that is especially important when evaluating a multi-component intervention: how should a relative change in a binary medication outcome be interpreted when the study uses cluster randomization? The registry's warning against a simple individual-level re-analysis highlights that the randomization unit affects uncertainty. A participant-level count may look straightforward, but the dependence among participants within clusters can matter to valid inference.

The primary endpoint also illustrates the difference between a measurable process outcome and a broader clinical outcome. Whether at least one medication disappears from a subsequent medication list is operationally defined and can be counted. It does not, by itself, convey why the medicine was stopped, whether the stop reflected the patient's goals, or what followed. This distinction is important for interpreting what the risk ratio represents and what it cannot establish.

The result is reported with both a confidence interval and a p-value, allowing readers to consider uncertainty and statistical evidence rather than relying on the point estimate alone. At the same time, the absence of arm-specific counts, population details, and full model specifications limits the degree to which the posted analysis can be checked or translated into an absolute difference.

Finally, the record lists multiple outcomes and analysis populations, while the primary result is per protocol. Transparent reporting of the analysis population, missing-data approach, and any multiplicity strategy is relevant to understanding how robust a finding is. Those elements are not reported in the registry record, so statistical interpretation must remain focused on the reported primary comparison and its stated limitations.

16. 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 does not report the randomized cluster unit, number of clusters, or the detailed randomization procedures. These details are important for understanding the design and the mixed models analysis. Although the record identifies cluster randomization and warns against individual-level re-analysis, it does not give enough model information to reproduce the estimate or assess how clustering was handled.

The primary result is reported as a risk ratio, but the analysis entry uses group identifiers without mapping them to the named intervention and waitlist arms. Group-specific outcome counts and proportions are not reported. Consequently, the direction of the comparison cannot be assigned from the analysis entry alone, and the absolute outcome difference cannot be determined. The exact p-value is also not reported beyond p < 0.05.

The primary analysis is per protocol, but the registry does not report the operational definition of that population, the number analyzed, exclusions, or reasons for exclusion. It also does not report a primary intention-to-treat result, missing-data methods, or sensitivity analyses. These omissions limit evaluation of attrition and the effect of analysis-population choices.

Multiplicity procedures and a hierarchy for the listed outcomes are not reported in the registry record. Further protocol details, the sample-size justification, and the full specification of the mixed model are not reported either. Although enrollment is 474, that total does not establish how many participants contributed to the per-protocol endpoint analysis.

The intervention arm descriptions in the registry are incomplete, and details about the waitlist period and timing are not reported. For safety, event timing, relatedness, and the denominators for the listed serious adverse event percentages are not described in the information summarized. These reporting limits constrain interpretation; they do not establish that the underlying procedures or analyses were absent.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: ALIGN: Aligning Medications With What Matters Most (Demonstration) — NCT06335953.
• Registry record last updated 2026-08-06; status last verified 2026-07.
• Results first posted 2026-08-06.
• 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.

20. Record Summary

ALIGN: Aligning Medications With What Matters Most (Demonstration), NCT06335953, is a completed randomized parallel-group trial sponsored by Johns Hopkins University. The registry lists enrollment of 474, no masking, and a primary completion date of 2025-09-16. It identifies polypharmacy, Alzheimer's disease and related dementias, and mild cognitive impairment among the conditions.

The study evaluates a pharmacist-led deprescribing intervention involving education, a pharmacist telehealth visit, and communication with the primary care clinician, compared with a delayed-intervention waitlist control. The primary endpoint is participants who stopped one or more medications at 3 months, assessed from electronic health record medication lists and analyzed per protocol.

The reported primary result is a mixed models risk ratio of 1.45, with a 95% confidence interval from 1.09 to 1.92 and p < 0.05. The registry notes cluster randomization and cautions that a simple individual-level re-analysis would be misleading. Group outcome counts, the mapping of result identifiers to named arms, and several analysis details are not reported in the registry record.