Study design
This study is a single-center, retrospective, observational, cohort design. The study was approved by the Clinical Research Ethics Committee of the Japanese Red Cross Society Kyoto Daini Hospital (Approval ID Sp2020–7). The Ethics Committee waived the requirement for informed consent because of the anonymous nature of the data.
Setting
The study was performed at the Japanese Red Cross Society Kyoto Daini Hospital in Kyoto City, Japan, which is an urban area with a population of approximately 1.5 million. The total number of ambulance calls is approximately 90,000 cases annually for the entire city [17]. The authors’ 672-bed hospital is one of the 4 critical care medical centers in Kyoto City, and is located at the center of the city. Generally, critical care medical centers are certified by the Japanese Ministry of Health, Labour and Welfare, and can accept emergency and severely ill patients transported by ambulance, including cardiac arrest, trauma, stroke, and sepsis patients, and can provide the specialized treatment in an intensive care unit, stroke care unit, and high care unit. In 2019, the total number of emergency department cases was 7610 patients who arrived by ambulance and 20,769 patients who were “walk-in” status, arriving by other means.
Study population
This study included all adult (age ≥18 years) patients hospitalized for any causes via our emergency department between January 1, 2019, and June 30, 2020.
Visitation restriction
In response to the COVID-19 pandemic in Japan, the study hospital changed the visitation policy to restrict visits beginning on March 28, 2020. The visit restrictions meant that the patient’s family and other close contracts were not allowed to visit the hospital ward or to have contact with inpatients in principle even for short periods of time. The study intervention periods were defined as before visitation restriction, from January 1, 2019, through March 31, 2020, and after visitation restriction, April 1, 2020. through June 30, 2020. We have been accepting inpatients because of COVID-19 since April, 2020, but this has not significantly affected our normal practice and has not restricted our emergency services. Even for inpatients, there were no changes in the hospital’s routine, except for visitation restrictions.
Data collection
Clinical data were obtained by an electronic chart review and the Japanese Diagnosis Procedure Combination (DPC) database of the Japanese Red Cross Society Kyoto Daini Hospital. These clinical data were collected through the electronic chart review: date of admission and discharge, patient age, patient sex, ward type on admission, and psychiatrists' medical record about delirium. The DPC database [18] includes administrative claims and discharge abstract data for all inpatients. The following information for each patient was recorded in the DPC using a uniform data submission form: age; sex; state of consciousness on admission/discharge; activities of daily living on admission/discharge; primary diagnosis; comorbidities, including dementia, on admission; and post-admission complications coded using the International Classification of Diseases, 10th Revision; medical procedures, including ventilator management and general anesthesia surgery; and discharge status. Patients were categorized by age as 18–64 years, 65–74 years, and ≥75 years, and by ward type on admission as emergency ward and general ward; the emergency ward was defined as the intensive care unit, stroke care unit, and high care unit. Patients were categorized as having ventilator management if they required >5 hours of ventilator management.
Outcome
The primary outcome was the incidence of delirium. In this hospital, the psychiatry department provides consultation as needed for the inpatients who are hospitalized in other departments, including their psychiatric diagnosis and intervention. Delirium in this study was defined as receiving a diagnosis of delirium by the psychiatrists and requiring their intervention for delirium during hospitalization. Although delirium is challenging to diagnose and is a diagnosis that tends to be missed [2], this study focused on diagnostic accuracy and clinically problematic delirium by limiting the incidence of delirium to that which required intervention by psychiatrists. The psychiatry department diagnosed delirium according to the 5th edition of American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [19], which is the current reference standard diagnostic criteria [20]. The policy in psychiatric department to diagnose and manage the delirium did not change over the study intervention periods of before and after visitation restriction.
Variables selection
Based on previous studies [1, 20-26], as potential confounding factors to assess the association between incidence of delirium and visitation restriction, the study used these seven variables: patient age, patient sex, ward type on admission, primary diagnosis, ventilator management, general anesthesia surgery, and dementia.
Statistical analysis
Data for patient characteristics were described as a median with an interquartile range (IQR) for continuous variables and as a number with percent for categorical variables. The crude and AORs of delirium incidence with 95% CIs were identified using the multivariable logistic model including all confounders. Furthermore, subgroup analysis was conducted to evaluate the interaction in the association between visitation restriction and delirium incidence. Crude odds ratios with 95% CIs were using univariable logistic regression models and P values for interaction [27]; all statistical results were considered significant at two-sided p<0.05. Missing data were not replaced or estimated. Statistical analyses were performed using JMP Pro 14 software (SAS Institute, Cary, NC, USA).