How to interpret conflicting scale results in psychiatric medication lists

First decide what kind of conflict is present

Two different numbers can look like a simple disagreement. In a psychiatric medication list, however, the apparent conflict may arise from different ingredients, formulations, source versions, or coverage states. The first task is to determine whether the records are actually comparable. Only then can the numerical difference be described accurately.

This article offers a comparison framework and a verified source-anomaly example from UnifiedACB. It does not calculate a new consensus rating or declare one scale the correct answer. The point is to make disagreement informative rather than make it disappear.

Match the identity before the value

Confirm the exact medication concept and literal label used by each source. A generic ingredient, a salt form, a combination product, and a route-qualified entry may require different handling. Similar words do not establish that the observations describe the same product.

Aliases help a search find a candidate; they do not authorize every possible merge. If identity is unresolved, the comparison should say so. Otherwise, an apparent disagreement may simply be a comparison of different things, and any calculated agreement statistic would begin with a false match.

Separate numeric disagreement from missingness

If both sources publish values for comparable records, retain those values and their versions. If one source has no entry, describe a coverage difference. A published zero is a third, distinct state and should not be confused with either absence or an unverified search result.

This matters clinically because a blank cell can be misread as reassurance. It also matters analytically because replacing blanks with zero changes the comparison denominator and can produce artificial agreement or disagreement. Missing observations cannot supply evidence about a category that was never published.

A concrete example: chlorpromazine in the KABS supplement

The current public dataset preserves a specific anomaly for chlorpromazine: it is omitted from the KABS 2020 open medication-list supplement despite the category count reported there. UnifiedACB does not infer a supplement score from that count. This is an absence in that source representation, not a published zero and not a new judgment about chlorpromazine. UnifiedACB public evidence dataset, v3.0 RC12 final

The example shows why a familiar medicine can still have a missing source cell. Knowing that a drug is commonly discussed in anticholinergic contexts does not supply a missing row in a particular publication. The source version and document location determine what can be represented as source-native evidence.

Examine the methods without choosing a winner by convenience

Scales can differ in their development, included medicines, evidence basis, and intended application. Those differences should be examined in the original publication. They do not authorize a blanket assumption that the newest, largest, or highest-valued scale is always preferable for every question.

A clinician or researcher may choose a source for a defined purpose, but should state that choice and its rationale. Selecting a different source after seeing a desired result is a different process and should not be described as a prespecified comparison. A transparent choice is more useful than an undocumented search for a reassuring or concerning number.

Keep psychiatric treatment context outside the scoring column

The source comparison does not measure treatment benefit, relapse risk, indication, or every adverse effect. These factors belong in the interpretation, but they should not change a source-native value. A medicine’s importance to the current care plan cannot justify lowering its printed classification; a clinical concern cannot justify raising it.

This separation allows a balanced review. A source may classify a medication more strongly while the treatment has a documented role. Both facts can be stated. The clinical team then considers the evidence alongside the person’s history rather than asking the table to make the decision.

Why an average loses information

Ordinal categories describe ordered classes within a method. Averaging categories from different scales creates a new quantity with no guaranteed clinical meaning. It also hides the range of published observations and makes missingness difficult to see. A “middle” number can look authoritative even though no source reported it.

The more useful display is a set of source-specific records with an explanation of coverage and qualifications. If a comparison uses an analytic mapping for research, that mapping must be explicit and justified; it should not overwrite the original values or be presented as a UnifiedACB patient score.

Write a comparison that can be checked

A clear summary identifies the matched medicine, each source version, the original values or evidence states, and any unresolved identity or publication issue. It then states the clinical question that remains. This allows another reviewer to trace the conclusion back to the observations.

A comparison record should make its unresolved cells reproducible. Record the medication label searched, the exact source version, and whether the outcome was a published value, no listed row, a source anomaly, or an unresolved product match. “The scales disagree” is insufficient when the apparent disagreement came from comparing a value with an omission. In the chlorpromazine example, preserving the supplement’s absence is itself the faithful result; filling it from another publication would change the question being answered.

When a later source or correction becomes available, revisit that cell with the same discipline. It may add evidence, clarify a label, or explain an omission. The previous record should still show why the earlier comparison was limited. A later clarification does not justify silently rewriting the historical source as though it had always reported the new information.

Limits and takeaway

Source agreement does not validate an individual clinical prediction, and disagreement does not make all sources useless. The first tells us how classifications align under defined conditions; the second can reveal methodological or coverage differences. Preserving those distinctions is more scientifically useful than forcing psychiatric medication lists into a single synthetic number.

References

1. Vennard O, Stewart C, Tolia M, Soiza RL, Myint PK. Anticholinergic Medication Burden Scales: A Systematic Review. J Am Geriatr Soc. 2026;74(6):1771-1784. Source
2. Georgiou R, et al. Anticholinergic Burden and Cognitive Performance in Patients With Schizophrenia: A Systematic Literature Review. Front Psychiatry. 2021;12:779607. Source
3. Mancini V, et al. Anticholinergic Burden and Cognitive Function in Psychosis: A Systematic Review and Meta-Analysis. Am J Psychiatry. 2025;182:349-359. Source
4. UnifiedACB. Public evidence dataset. v3.0 RC12 final. Read-only source observations, source registry, and preserved anomaly records. Source

Disclosure

Funding: None. Conflict of interest: None declared for the UnifiedACB educational project.

Educational evidence summary. This content does not provide patient-specific medical advice and does not replace clinical judgment, medication reconciliation, or prescribing guidance.