The study was designed as an observational study with cohort of patients and corresponding nurses in an Emergency Department (ED) at a University Hospital within the Capital Region of Denmark. This study is reported according to the STROBE guidelines (16).
Setting
The Danish health care system is predominantly tax-paid and access to the EDs use a “telephone-triage before presentation” and as such walk-ins at the EDs are not encouraged (17). The ED at Copenhagen University Hospital, Amager Hvidovre covers a catchment population of 517,000 citizens and serves all acute medical patients referred to the hospital by pre-hospital services (General Practitioner, Out-of-Hours services and Emergency Medical Services) with the exceptions of pediatric, gastroenterological and obstetric patients who are directed to specialized EDs within the hospital. The ED comprises of 29-beds, has roughly 70,000 patient contacts/year (2019 numbers) and coordinates the admissions of 45–60 patients per day. In 2017 admitted patients had a median length of stay of 7.8 hours (interquartile range (IQR), 3.6–17.5). Maximum length of stay within the ED is 48 hours after which the patients must be transferred to a specialized ward (18). Nursing work within the department is organized in three groups, each nurse is the primary nurse for 4–6 patients but help each other if needed.
Data collection
The data collection took place from the 22nd of February to the 27th of March 2021 and was carried out by two trained health care professionals (one registered nurse and one nurse student - both from the department). On days of data collection, the data collector screened the patient board in the staffs’ office and made notes of the patients’ names, room numbers and corresponding nurses. All patients ≥ 18 years, cognitively intact (excluding demented, mentally or physically too weak to participate and/or under the influence of alcohol or drugs) and Danish or English speaking were eligible to participate. Nurses regardless of seniority and gender were eligible for participation.
The data collection consisted of a survey aimed at the patient and the patients’ primary nurse. The patient’s answer to the DOW question: “how worried are you about the condition you are here today on a scale from 1 to 10, where 1 is minimally worried and 10 is maximum worried” was registered alongside gender, age, co-morbidity, triage level, and medical reason for encounter.
The nurses’ survey was centered around the question: “how worried do you think your patient is about the condition he/she is there today on a scale from 1 to 10, where 1 is minimally worried and 10 is maximum worried?” Nurses also supplied data on gender, age, seniority as a registered nurse (RN) and in the ED.
Variables
Patients’ self-reported DOW and the nurses’ estimation of the patients’ DOW was registered on a scale from 1 to 10, however for the primary analysis the variable was categorized into three levels of DOW; DOWlow (degree-of-worry 1–3), DOWmiddle (degree-of-worry 4–6), DOWhigh (degree-of-worry 7–10) based on the association between degree-of-worry and acute hospitalization (9). Age was collected as a continuous variable and categorized into 18–39 years, 40–59 years, 60–79 years, and 80 + years. Likewise, nurses’ seniority was collected as a continuous variable but categorized into 0–1, 2–5, 5 + years. Information on co-morbidity was collected by self-report as the presence/absence of co-morbidity. Triage level was registered as the level registered on the patient board in the staff office. Each patient is provided a triage level on arrival which is estimated based on vital parameters such as saturation, blood pressure, pulse, temperature, and alertness and indicates the urgency of the condition. The five-level Danish triage manual resembles the Manchester triage manual (19,20). Patients could only participate once but if a nurse participated more than once he/she was included as a new nurse each time, as the aim of the study was to investigate the agreement of DOW-rating in the patient-nurse dyad.
Statistical methods
Sample size calculation was based on a two rater, patients and nurses, weighted kappa for agreement on the three-level categorization of DOW. With expected distribution of the DOWlow, DOWmiddle and DOWhigh of 21%, 36% and 43% respectively and an expected Kappa value of 0.7 was assumed and tested against a null hypothesis kappa value of 0.80 with a power of 0.80 and critical value of 0.05, resulting in a sample size of 213. Sample size calculation was done in R 4.1.2 using the Power3Cats function from the kappa Size package (21).
Descriptive statistics are presented as frequencies and percentages for categorical variables and mean and standard deviation or median and inter quartile range (IQR) for continuous variable normal-and non-normal distributed, respectively.
The agreement between the categoric DOW (patient/nurse) was assessed with weighted Cohen’s Kappa and tested against the null-hypothesis kappa value of 0.8, for total population and stratified on patient’s gender, age, co-morbidity and nurses’ gender, age and seniority. Results were reported as the weighted Kappa coefficient and 95% confidence intervals (95%CI).
To access absolute bias of DOW (absolute value patient DOW minus nurse DOW), bias regardless of direction, one-sample Wilcoxon sum-rank test was used, as absolute bias could not be assumed to be normally distributed. Further associations of DOW absolute bias with nurses’ gender, age, seniority as a nurse and ED experience were analyzed by Kruskal-Wallis test. Estimates are presented as median absolute DOW bias with IQR.
Normality assumptions were evaluated using QQ-plots. P-values < 0.05 were considered statistically significant. Data was analyzed using SAS enterprise 8.3.
Ethics and approvals
Patients and nurses were invited to participate and informed of the project and that participation was voluntary. All participants gave written informed consent. The study was approved by the National Data Protection Agency reference number P-2021-127.