The list is only useful when differences are understood.

NICE's medicines optimisation guidance includes reconciliation and communication when people move between settings (NICE, n.d.). A local process review should follow how information is obtained, checked and passed on. It should distinguish a missing source from a difference between sources, and a documented explanation from an assumption. The same apparent discrepancy may require different action depending on context. Our proposed organisational evaluation would leave clinical interpretation with the responsible professionals and focus on whether the preparation supports that judgement.

An unresolved item needs an accountable route: who can clarify it, what information is needed and how the response returns to the reviewer. A generic status such as pending is not enough for an operational plan. The proposed brief would make those dependencies explicit and distinguish confirmed facts from open questions. It should not infer that a medicine was intentionally changed because two documents disagree. Showing uncertainty clearly is a design strength when the alternative is a confident but unsupported explanation.

For an organisational discrepancy log, distinguish information not supplied, conflicting sources, an explanation requiring confirmation and an action awaiting an accepted owner. These are coordination categories, not judgements about clinical severity. Track the source and time of each observation. The distinction prevents a team treating every unresolved item as the same kind of problem and helps it direct clarification requests to the appropriate role instead of repeating the entire review.

A first study could use synthetic transition cases and approved process documents. Participants would prepare an organisational discrepancy brief using the current approach and a PHARMORIS-supported approach. An independent clinical panel would define the important missing information, misleading statements and required escalations. The task would exclude prescribing, treatment changes and live record updates. This provides an early test of whether the proposed support helps reviewers organise the work, before any decision about access to real records or operational deployment.

A medicines-transition lead could review recurring preparation gaps at an agreed service meeting and after a material pathway change. A prospective sponsor would supply approved process documents, representative synthetic cases and qualified reviewers who can define important omissions. Acceptance would require the brief to preserve contradictions, assign only agreed organisational routes and expose every unsupported inference, with checking effort counted. Further commissioning would depend on persistent preparation or coordination work that the existing method leaves unresolved, not on replacing professional reconciliation.

Measure important issues correctly identified, unsupported inferences, traceability to the supplied material and the effort needed to check and correct the brief. Speed is a secondary question unless the output meets the agreed quality threshold. Include examples with conflicting and incomplete information, rather than testing only clean records. A later study could examine the coordination burden in practice, but an improvement on synthetic examples would remain an early development finding rather than evidence of fewer medication errors or better patient outcomes.

A useful first conversation involves the medicines team responsible for the transition and the people providing the information it needs. Bring a process description and a representative example stripped of confidential details. We would define the preparation task, the comparator and what would justify further work. The value proposition is a more inspectable and potentially less burdensome organisational process. Whether PHARMORIS can deliver that value is the subject of the proposed evaluation, not a result assumed in advance.

NICE (n.d.) Medicines optimisation, NG5. Available at: https://www.nice.org.uk/guidance/ng5/chapter/Recommendations (Accessed: 14 October 2026).