Subscribe to our Newsletter

Subscribe to our Newsletter


Predictive tool for early detection of delirium in hospitalized patients

Hospital General Universitario Gregorio Marañon

The Hospital General Universitario Gregorio Marañón (HGUGM) is a public hospital in Madrid, Spain. It is the largest hospital in the Community of Madrid and serves more than 2 million people. It has 8,000 employees and more than 1,200 beds. It is a teaching hospital, attached to the Complutense University of Madrid.

In the year 2021, the data on healthcare activity were: Total discharges 42.624, the average length of stay 7,61, total admissions 42.531, emergency admissions 26.691 total emergencies 239.076 and percentage of emergency admissions 11,34%. The incidence of delirium in patients over 65 years old ranges between 11,7% and 18,5%, depending on the type of hospitalization unit [1–3].

Since 2018, the Hospital has been equipped with an Electronic Health Record (EHR) that encompasses relevant information concerning the patient’s acute clinical conditions, baseline medical history, functional and mental state, as well as socio-demographic and clinical variables.

The available Electronic Health Record (EHR) electronically stores a patient’s medical information and care records. It enables healthcare professionals to access, review, and update patient information quickly and efficiently. Currently, with the EHR, it is possible to conduct a comprehensive review of a patient’s medical data, including previous diagnoses, treatments, test results, prescribed medications, healthcare provider notes, and other relevant data for patient care. Additionally, it allows for the registration of comorbidity using the Charlson index[4] and assessment of pain with the relevant scale depending on the cognitive level. Delirium is assessed upon admission and whenever there is a substantial change in the patient’s condition through the Confusion Assessment Method (CAM)[5]

The confusion assessment method (CAM) is a simple tool that can be used by physicians and nurses to integrate their observations and identify when delirium is the most likely diagnosis (Figure 1). In medical and surgical settings, CAM has a sensitivity of 94% to 100% and a specificity of 90% to 95%[5,6].

Delirium is a potentially preventable complication, and several different interventions have been developed over the last decade to prevent and control it. Some of these interventions involve nursing staff, while others focus on treatment, and many attempt to prevent delirium after surgery through pharmacological interventions.

We believe that efforts should be made to improve the identification of patients at risk during admission in order to establish preventive interventions and avoid complications such as falls, increased average hospital stay, and unintentional removal of devices. The CAM has become a standard screening device in clinical studies of delirium conducted in multiple settings, including emergency departments and long-term care [7]. It takes five minutes to administer and can be particularly useful when incorporated into routine bedside assessment, however current workloads and nursing shortages make it difficult to assess on a shift basis.

Since the Hospital does not have an early detection system, we consider that incorporating a delirium algorithm into a third-party analytical solution owned by the hospital with a systematic approach would be an optimal solution for this problem.


Delirium is an acute state of confusion characterized by an altered level of consciousness and impaired attention, resulting in cognitive and perceptual disturbances that cannot be explained by preexisting dementia. Its onset is rapid, typically occurring within a short timeframe of hours to days, and it tends to fluctuate throughout the day. Although some consider delirium to be a specific type of confusional state marked by heightened vigilance, increased psychomotor and autonomic activity, and symptoms such as agitation, tremors, and hallucinations. for the purposes of this project the terms “delirium” and “acute confusional state” are used interchangeably and encompass states characterized by decreased arousal, referred to as “hypoactive delirium.”[8]

The management of delirium is primarily based on expert consensus and observational studies, as conducting controlled clinical trials with cognitively impaired patients poses significant challenges. The strongest evidence supports nonpharmacologic, multicomponent approaches for primary prevention of delirium in high-risk patients [9–11].

Detailed Explanation of Delirium Risk Evaluation in our Hospital

  1. Delirium Prevention:
    • Identifying patients at risk: Nursing professionals assess the risk of delirium within the first 24 hours (72 hours if admission occurs on a weekend). Patients at increased risk include the elderly, individuals with cognitive impairment, those with a history of delirium, or those undergoing specific medical procedures.
    • Managing medications: Reviewing and adjusting medications that may contribute to delirium, especially sedatives, anticholinergics, and medications affecting the central nervous system.
    • Ensuring adequate hydration and nutrition
    • Promoting good sleep hygiene
    • Encouraging early mobilization
  2. Diagnosis of Delirium
    • Nurses, physicians, and other healthcare providers regularly communicate with and observe patients for signs of delirium to ensure early detection. However, high workloads, staff shortages and turnover, and night shifts make it unfeasible to perform a shift-by-shift assessment. Currently, there is no early detection system available.
    • The Confusion Assessment Method (CAM) is used as a widely used screening tool to identify delirium. It involves a series of questions and observations related to attention, disorganized thinking, and altered level of consciousness. The CAM is a test based on the presence of four criteria: acute onset and fluctuating course, inattention, disorganized thinking, and altered level of consciousness.
  3. Management of Delirium:
    • Identifying and treating underlying causes: Addressing the root causes of delirium, such as infections, electrolyte imbalances, pain, or adverse drug reactions.
    • Providing supportive care
    • Implementing non-pharmacological interventions
    • Medication management: In some cases, medications may be prescribed to control severe agitation or distress, but they are used with caution due to the risk of exacerbating delirium.

The importance of having a predictive tool in the hospital is crucial especially in the context of nursing shortages and night shifts. The use of an early detection system for delirium integrated with the EHR can be a valuable tool for the medical team, allowing for early detection and a more effective response to delirium, which enhances the quality of care and patient safety. The solver should provide the technological solution.


Applications closed on the 24th of January 2024 at 17:00 CET.