# Validation of the confusion assessment method for intensive care

**Authors:** Carla Cristina Nunes Teixeira Veiga, Paulo Alexandre Oliveira Marques, José Manuel Monteiro Dias, Amâncio António de Sousa Carvalho

PMC · DOI: 10.1515/med-2026-1394 · Open Medicine · 2026-03-02

## TL;DR

This study validates a tool for detecting delirium in ICU patients using European Portuguese, showing it is reliable and accurate.

## Contribution

The study validates the CAM-ICU flowsheet in European Portuguese for nonverbal ICU patients.

## Key findings

- Delirium was detected in 57.1% of patients using the validated tool.
- The tool showed high interobserver agreement (κ=0.952) and excellent sensitivity (100%).
- The inattentiveness feature was critical for identifying delirium.

## Abstract

Delirium is a common but frequently underdiagnosed complication in adult intensive care units (ICU), associated with increased morbidity, mortality, and healthcare costs. Early detection using validated tools is essential. This study aim to translate, adapt, and validate the Confusion Assessment Method for Intensive Care Units flowsheet (CAM-ICU flowsheet) into European Portuguese.

A methodological, cross-sectional, descriptive-correlational study was conducted in a northern Portuguese ICU. The CAM-ICU flowsheet was translated and back-translated, followed by expert validation. Forty-two adult ICU patients on invasive mechanical ventilation were assessed using the adapted tool by two trained observers. Diagnostic accuracy was evaluated using DSM-5 criteria as the gold standard. Statistical analysis included sensitivity, specificity, kappa coefficient, and ROC curves.

Delirium was detected in 57.1 % of patients. The CAM-ICU flowsheet in European Portuguese demonstrated high interobserver agreement (κ=0.952), excellent sensitivity (100 %), and specificity (81.8–86.3 %). ROC AUC ranged from 0.909 to 0.932. The inattentiveness feature was the core for the presence or absence of delirium.

The CAM-ICU flowsheet in European Portuguese was validated to assess delirium in nonverbal ICU patients on mechanical ventilation through this study and demonstrated to be a reliable and valid tool.

## Linked entities

- **Diseases:** delirium (MONDO:0045057)

## Full-text entities

- **Genes:** CALM3 (calmodulin 3) [NCBI Gene 808] {aka CALM, CAM1, CAM2, CAMB, CPVT6, CaM}
- **Diseases:** Pain (MESH:D010146), Altered (MESH:D004408), visual impairment (MESH:D014786), inattention (MESH:D001308), critical illness (MESH:D016638), trauma (MESH:D014947), anxiety and depression (MESH:D001007), Sleep Disruption (MESH:D019958), diabetes (MESH:D003920), muscle weakness (MESH:D018908), Mental Disorders (MESH:D001523), Confusion (MESH:D003221), chronic obstructive pulmonary disease (MESH:D029424), Delirium (MESH:D003693), Organ Failure (MESH:D009102), brain death (MESH:D001926), falls (MESH:C537863), respiratory failure (MESH:D012131), disorganized thinking (MESH:D012562), frailty (MESH:D000073496), Acute (MESH:D000208), Acute Respiratory Distress Syndrome (MESH:D012128), fever (MESH:D005334), hyperactive (MESH:D006948), post-traumatic stress disorder (MESH:D013313), hypertension (MESH:D006973), intellectual disability (MESH:D008607), brain dysfunction (MESH:D001927), Altered level of consciousness (MESH:D003244), died (MESH:D003643), mental health disorders (OMIM:603663), lethargy (MESH:D053609), COVID-19 (MESH:D000086382), Agitation (MESH:D011595), neurocognitive disorder (MESH:D019965), hypoactive (MESH:D020018), arterial (MESH:D012078), cardiovascular disease (MESH:D002318), aggressive behavior (MESH:D010554), dementia (MESH:D003704), sepsis (MESH:D018805), coma (MESH:D003128), cognitive decline (MESH:D003072), movements (MESH:D009069), TISS-28 class III (MESH:D008313), septic shock (MESH:D012772)
- **Chemicals:** fentanyl (MESH:D005283), propofol (MESH:D015742), benzodiazepine (MESH:D001569), IMV (-), dexmedetomidine (MESH:D020927), tramadol (MESH:D014147), morphine (MESH:D009020)
- **Species:** Homo sapiens (human, species) [taxon 9606]

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12952290/full.md

## References

48 references — full list in the complete paper: https://tomesphere.com/paper/PMC12952290/full.md

---
Source: https://tomesphere.com/paper/PMC12952290