ICMM cross-case summary

This is a publication-oriented summary of the canonical mechanism findings. It must not become a second thematic register.

The current series contains 71 fully extracted reports from 87 retained candidates; 16 candidates were excluded against the protocol. It combines complete 2018 publication-cohort enumeration/extraction with 36 retained cohort decisions, targeted 2019–2026 discovery and a limited check for later-migrated 2018 reports. The other 343 cohort rows lack retained contextual dispositions. One included PFD covers three deaths, so report and death counts are not interchangeable. These are corpus counts, not prevalence estimates.

Headline synthesis

• Information can be stored without becoming perceptible, delivered or actionable in the workflow used for care.
• Viewing, filing or acknowledging an item is not the same as assigning ownership, escalating delay and recording clinical closure.
• Interoperability includes person identity, source completeness, meaning, workflow receipt and visible missingness—not transport alone.
• Paper, live, exported and investigation views can silently diverge unless authority and provenance are explicit.
• Decision support is a sociotechnical control whose value depends on severity, actionability, workload, ownership and governance.
• Degraded operation and migration must preserve observable safety barriers, not rely on informal recovery.

Each statement is unpacked with named report examples, variation and limits in the Findings section. The case–mechanism map shows the underlying non-exclusive coding, and the Defence requirements register contains the controlled analytical translation.

Publication boundary

Before submission, rerun the search to the closing date, independently recheck coding and source lineage, and refresh derivative outputs from the canonical records. Do not use “systematic”, “exhaustive”, “prevalence”, “first study” or a digital-PFD trend claim without the additional methods and evidence recorded in the validation queue.