Delirium during inpatient rehabilitation: detection, risk factors, clinical implications, and risk assessment

One of the main goals of epidemiology is to identify factors that increase the risk of a single individual or a population developing a certain disease to reduce its morbidity or mortality. For this purpose, observational analytical studies, such as cohort studies or case-control studies, are often appropriate, as these kinds of studies are generally cheaper, more generalizable, and can be conducted on a larger scale than experimental studies. Rehabilitation is a fundamental process in modern medicine that aims, through an interdisciplinary approach, to restore and maintain optimal physical, sensory, intellectual, psychological, and social functioning in patients with acute or chronic disabilities. Delirium is known as a brain syndrome that dramatically impairs consciousness, attention, perception, thought, memory, psychomotor behavior, emotions, and the sleep-wake cycle of affected inpatients. In both acute and rehabilitation settings, delirium has been associated with prolonged length of stay and increased mortality. Furthermore, as patients suffering from delirium are unable to fulfill the rehabilitation therapy schedule, they often achieve poor rehabilitation outcomes, such as reduced autonomy after discharge. Delirium consequences cause annually additional costs for rehabilitation facilities and the entire healthcare system. Preventing delirium and thus minimizing its clinical implications is possible, non-pharmacological interventions, such as cognitive stimulation or patient mobilization have been proven to be effective. However, to implement targeted interventions, it is essential to identify patients potentially at risk of developing delirium during inpatient rehabilitation. This thesis aimed to identify a set of risk factors for the development of incident delirium during inpatient rehabilitation, to assess the impact of delirium on functional rehabilitation outcomes and length of rehabilitation, and to develop a clinical prediction model based on parameters available on admission able to quantify the individual’s risk of developing delirium during inpatient rehabilitation. For this purpose, the thesis was structured on three projects based on data from a collective of approximately 10’000 patients who underwent inpatient rehabilitation at ZURZACH Care, Rehaklinik Bad Zurzach, an inpatient rehabilitation clinic in Switzerland. We first developed and validated a chart-based method to detect incident delirium episodes within our study population. The developed method, based on delirium predictive key words identified within medical notes, was able to detect incident delirium episodes with low to moderate accuracy. However, supplemented with experts’ review, this method was able to identify a validated population of 125 incident delirium episodes. Based on the validated delirium episodes and routinely collected clinical data, we conducted a retrospective matched case-control study in the subsequent project. The results identified older age, neurological rehabilitation, reduced Functional Independence Measure (FIM), and high disease (CIRS) or anticholinergic burden (ACB) at admission as factors associated with a considerably increased risk of incident delirium during rehabilitation. Patients with infectious diseases, disorders of fluid, electrolyte and acid-base balance, and Parkinson’s disease at admission, and patients treated with laxatives, heparins, antidepressants, dopaminergic agents, and antipsychotics during rehabilitation, were at an increased risk of developing delirium compared to patients without these conditions or patients not receiving these drug classes. Furthermore, we observed that patients who became delirious during rehabilitation had a longer mean stay and a poorer functional rehabilitation outcome than patients without delirium. In the last project, we developed a clinical prediction model based on parameters available at rehabilitation admission. We observed that age, FIM, disorders of fluid, electrolyte and acid-base balance, use of other analgesic and antipyretics or anti-Parkinson drugs on admission, and an ACB ≥3 were selected as predictor parameters. Upon validation, this model would allow an individual’s risk calculation of developing incident delirium during inpatient rehabilitation