Impact of a delirium prevention project among older hospitalized patients who underwent orthopedic surgery: a retrospective cohort study

This study, which used a before-after approach to evaluate the effectiveness of a project designed to prevent POD, was conducted at Seoul National University Bundang Hospital, a 1300-bed teaching hospital. We defined two study phases, considering annual variation in the characteristics of orthopedic surgical patients, as follows: 1) an evidence-based delirium prevention project phase conducted between December 2018 and February 2019 (intervention group) and 2) a phase involving patients admitted to the same wards during the same time period 1 year earlier, from December 2017 to February 2018 (control group). All inpatients aged ≥65 years who underwent orthopedic surgery and had been admitted to either of the two orthopedic wards (ward 61 and 62) from both the intervention and control study phases were included. Patients who were delirious at the time of hospital admission were excluded from the study. To identify the effect of the delirium prevention protocol application in the general ward, we analyzed the data of all patients in both the intervention and control groups, although some patients missed screening or intervention for reasons such as an examination or scheduled surgery in the intervention group.

The study protocol was reviewed and approved by the Seoul National University Bundang Hospital Institutional Review Board, and the requirement for informed consent was waived (IRB No. B-1904/534–104).

Through an expert team meeting, we developed a screening tool to identify patients who had an increased risk of POD based on reviews of previous studies investigating risk factors for POD [1, 14]. In the screening tool, because the most universal delirium prevention activities are intended to be performed at the age of 70 years and older and because the risk of delirium is known to increase dramatically after 80 years of age, patients aged < 70 years were excluded from screening and were classified as low risk, while patients aged ≥80 years were classified as high risk for POD [1, 5]. Patients aged 70–79 years were classified as high risk for POD if they had a history of dementia or delirium, took dementia medication, or had a Korean version of the AD8 (K-AD8) score ≥ 2, which was identified using a Clinical Dementia Rating over 0.5 (cognitive decline) from 0 (normal) [15] (Fig. 1).

Based on a focus group interview, preliminary meetings were held among team members, including geriatricians, a geriatric nurse-specialist, a pharmacist, the senior nurse in charge of the two orthopedic wards, an orthopedic surgeon, and a neurologist. The POD prevention project, comprising four components: education, screening, prevention, and scheduled assessment, was developed (Fig. 2). The first component of the prevention project aimed to enhance education for nurses, orthopedic surgical patients, and their caregivers. We revised an 8-page mini-handbook to include information on delirium (definition, risk factors, symptoms, prevention, prognosis, and Questions and Answers) for patients and caregivers. Nurses working in wards, orthopedic specialists, and residents received a 30-min training session on the pathophysiology, risk factors, prevention, screening, and treatment of POD. The training session was delivered by a geriatric nurse-specialist or by geriatricians. During the pre- and post-training sessions, we evaluated the degree of knowledge regarding POD, previously developed and validated in Korea [16]. Additionally, to maintain ongoing education, training material concerning delirium was added to the education program for new nurses on the team.

The second component of the project aimed to establish a routine screening process to identify patients at risk of POD. We adopted our newly introduced screening tool to be used at the time of admission by a nurse. POD prevention was undertaken in patients identified to be at high risk of POD.

The third component of the project involved the use of an interdisciplinary preventive approach. Nurses provided a mini handbook on POD to patients and their caregivers who had been identified via screening to be at risk of POD. A pharmacist and a group of geriatricians defined a list of 15 medications (chlorpheniramine, diphenhydramine, hydroxyzine, amitriptyline, scopolamine, alprazolam, clonazepam, diazepam, etizolam, lorazepam, triazolam, zolpidem, pethidine, famotidine, and ranitidine) that were reported to potentially cause delirium based on a review of previous studies and the frequency of use of these medications [17, 18]. To make a list of inappropriate medications for the elderly, we used the medication list in our hospital (2012 SNUBH Inappropriate Medication for Elderly Criteria), which is based on the 2008 Screening Tool of Older People’s Prescriptions criteria and 2012 Beers criteria [19]. The medication list for high-risk patients was reviewed by a pharmacist and a geriatrician at admission, and information was provided to orthopedic units by progress notes regarding duplicate prescriptions or medications that could cause delirium or were inappropriate for older adults. Based on this information, alternative medications were suggested by pharmacists and geriatricians. A checklist was provided to high-risk patients detailing the implementation of the evidence-based interdisciplinary approach to prevent POD. The items on the checklist included blood pressure and blood glucose control, pain control, defecation and urination, volume status pre- and post-operatively, electrolyte status, sleep hygiene, and sensory deprivation (the use of glasses and hearing aids), and relevant preventive interventions were performed from the day of admission to 3 days after surgery. Among the check-list items, pain assessment, pain control, and sleep hygiene education were conducted primarily by nurses. Blood pressure, blood glucose, and fluid and electrolyte balance were evaluated and modified by geriatricians. The entire checklist was screened and modified by a geriatrician on a daily basis and was included in the progress notes in the electronic medical record system for ward nurses and orthopedic staff to consider.

The fourth component of the project comprised a scheduled assessment at every nursing staff shift, three times a day, for signs or symptoms of POD. Orientation to time, place, and person was assessed, and if a patient was disorientated, a nurse provided correct orientation (re-orientation). In addition to the ward nurse’s assessment, a geriatric nurse specialist and geriatricians examined the patients daily and evaluated symptoms throughout the day to diagnose POD. If POD was suspected, physicians were encouraged to prescribe atypical antipsychotics when needed or to co-consult with a psychiatrist concerning the patient.

The primary outcomes were the sensitivity and specificity of the delirium screening tool in predicting POD during the intervention period. The secondary outcomes were the effectiveness of the delirium prevention project in reducing the risk and incidence rate of POD, as diagnosed by a psychiatric consultant or using the Diagnostic and Statistical Manual of Mental Disorders (4th edition) criteria in both cohorts through a retrospective chart review from admission date to discharge date. These diagnostic criteria for delirium were also adapted in the analysis to identify the sensitivity and specificity of the delirium screening tool. To assess the effectiveness of the delirium prevention project, we compared outcomes between the project study period (December 2018 to February 2019) and the same time period 1 year earlier (December 2017 to February 2018). We compared these two phases to assess the effectiveness of the risk reduction of POD and of length of hospital stay. A validation of the delirium screening tool and a comparison of the degree of knowledge of delirium before and after education were performed in the intervention group.

Continuous variables were represented as means (standard deviations [SD]) or medians (interquartile ranges [IQRs]) if the variables were not normally distributed. The paired- or independent t-test and the chi-squared test were used for continuous and categorical variables, respectively. The relationship between the introduction of the POD prevention project, sex, age, or other conventional risk factors and outcomes was determined using logistic regression models. The length of hospital stay was compared using a nonparametric test, the Mann-Whitney U test. Statistical analysis was performed using SPSS version 25.0 (SPSS Inc., Chicago, IL, USA).

A total of 275 older adults who were admitted during the study period (intervention group) and 274 older adults who were admitted to the same wards during the same time period 1 year earlier (control group) were included in this study. Baseline patient characteristics such as age, sex, body mass index, and the American Society of Anesthesiologists (ASA) classification were similar between patients in the intervention and non-intervention groups. In addition, confounding factors that could influence the baseline severity of patients, such as the number of medications, number of diseases, existence of dementia medication, education level, or living situation, were not different between two groups (Table 1). Among the 275 intervention group patients, 78 patients were aged < 70 years and 53 were aged ≥80 years. Among the 144 patients (aged 70–79 years) who required screening for delirium/dementia owing to their medical or medication histories and K-AD8 scores at admission, with the exception of 16 patients who missed screening, 99 and 29 patients were classified as low and high risk, respectively. Finally, among the all 275 patients in the intervention group, 82 were classified as being at risk of delirium (Fig. 3).

During the study period, with the exception of the 16 patients who missed screening, 82 patients (53 patients aged ≥80 years and 29 patients aged between 70 and 79 years who screened positive) were classified as being at risk of POD and were targeted with individualized delirium prevention strategies. A total of 177 patients (78 patients aged < 70 years and 99 patients aged between 70 and 79 years who screened negative) were classified as being at low risk of POD. Among the patients (n = 82) who received the multicomponent interdisciplinary POD prevention intervention and a review of medications in use, 17.1% (n = 1 4) and 36.6% (n = 30) of patients, respectively, were taking medications inappropriate for older adults or medications reported to potentially cause delirium.

During the course of the delirium prevention QI project, 17 patients experienced POD. Of these, POD occurred in 16 patients who were classified as high risk according to the screening tool. The sensitivity and specificity of the delirium screening tool in estimating POD risk were 94.1 and 72.7%, respectively.

The incidence rates of POD in the intervention and control groups were 6.2% (n = 17) and 10.2% (n = 28), respectively. To test the effectiveness of the delirium prevention project in POD prevention, we used multiple logistic regression models to obtain adjusted relative effect measures for the incidence of POD. Variables that could affect the incidence of POD and basic demographic characteristics were adjusted, and the results are shown in Table 2. The POD prevention project showed significant effectiveness with an odds ratio of 0.316 (95% confidence interval [CI]; 0.125–0.800, p = 0.015) after adjustment for possible confounders. In the multivariate analysis, the number of medications, existence of dementia medication, type of surgery, and advanced age were also associated with an increased risk of delirium.

The pre- and post-educational training session test scores involving the definition, classification, incidence, risk factors, signs and symptoms, prevention, screening, diagnosis, and treatment of POD were 40.52 and 43.24 out of 49 total points, respectively. The effect of education throughout the study period was analyzed using a paired t-test and was shown to be statistically significant (p < 0.001) The median length of hospital stay for the intervention and non-intervention groups was 6.0 days (IQR, 4–9) and 7.0 days (IQR, 4–10), respectively. (Table 1) The delirium prevention project did not shorten the length of hospital stay significantly (p = 0.167).