Patient Outcomes and Hospital Nurses’ Workload: A Cross-Sectional Observational Study in Slovenian Hospitals Using the RN4CAST Survey

Research has consistently shown that higher nursing workloads increase the odds of patient deaths. In hospital settings where RNs provide care to a greater number of patients per shift, where their work environment is unsupportive and managers do not respond to RNs’ concerns, patients have poor health outcomes and RNs report greater burnout, job dissatisfaction, and intent to leave (1 - 4). When nurses care for fewer patients at a time, they are able to spend more time at each patient’s bedside, and as a result patients are less likely to experience an adverse outcome, the length of stay is shorter, and there are cost savings for the hospital (5). Griffiths (6) states that staff shortages can lead to a proportion of nurses abandoning the provision of services, which is a direct indicator of inadequate staffing levels in nursing, as has been shown in international studies (7).

Longitudinal findings revealed an association between increased proportions of bachelor of science RNs (BSN) and improvements in patient outcomes; higher proportions of BSNs were associated with lower odds of 30-day inpatient surgical mortality (1, 6, 8). The BSN educational pathway (four-year study programme) is recommended for at least 80% of the BSN workforce (8 - 11). Higher hospital proportions of BSN nurses, regardless of educational pathway, are associated with lower odds of 30-day inpatient surgical mortality (5, 9).

Schlak et al. (3) found that the work environment had a significant effect on patient outcomes, as a change in the work environment from poor to mixed or mixed to good was associated with a 14% drop in the odds of 30-day in-hospital mortality. Addressing staffing without improving the practice environment would be ineffective (12), because better staffing and practice environments are associated with improved patient outcomes (1, 3, 8, 13 - 15). The cumulative evidence suggests that interventions to improve hospital work environments and patient-to-nurse staffing ratios may be key to addressing quality, safety, and nurse workforce concerns (4), and staffing is a key variable for patient mortality within 30 days of admission, readmission, and hospitalisation time (15).

Slovenia faces a major shortage of RNs, with only 30% of RNs in the nursing workforce group, a far cry from recommendations noted above – 80% of nurses with a bachelor of science in nursing (BSN) (9, 11). The title of nurse (RN) describes anyone meeting the European Union (EU) definition of a trained and licensed nurse according to Directive 2013/55/EU (16). In Slovenia, this corresponds to having an undergraduate three-year professional bachelor’s degree in nursing, level six of the European Qualifications Framework (EQF). The education level of healthcare assistants (HCAs) is four years of secondary school for healthcare technicians, level four of the EQF. For years, health policymakers on Slovenia have been neglecting the need to increase the number of nurses with a minimum of a professional bachelor’s degree in clinical settings and educate them to a four-year bachelor of science in nursing (BSN). Most European Union countries have switched from a three-year professional nursing degree to a four-year science degree because students’ workloads were too high and the three-year professional degree is too short to deliver the competences set out in the 2013 update of the Directive (16).

The first aim of our study is to explore the relation between patient outcomes and nurses’ working conditions in hospitals. The second is to describe nurses’ workloads and place them in an international context.

For this observational study with a cross-sectional design, administrative discharge data on surgical patients for the year 2019 was collected in all acute care hospitals in Slovenia. Nurses in all general hospitals (n=10) and university clinical centres (n=2) were invited to participate in a survey to estimate nurse staffing. Eight general hospitals and two university clinical centres participated in the survey (n=10). Nurses in participating hospitals providing direct patient care on surgical and medical inpatient units (n=2813) were invited to participate in the survey between February and March 2020.

The overall response rate for the nurse survey was 35.90% and varied across hospitals, ranging from 21.97% to 85.71%. Respondents included 848 (83.96%) women and 160 (15.84%) men, while the average age was 37.02 (SD=10.65) years. The average length of employment in nursing was 15.34 years (SD=11.12). Almost all the respondents (96.2%) were employed full time. On average, 35.65% (n=403) of nursing care providers had completed an undergraduate professional bachelor’s programme, which varied from 23.68% to 48.76%, depending on the hospital (Table 2); on average 59.9% (605) were HCAs.

The RN4CAST survey instrument comprises a core battery of extensively validated instruments and questions (17). In this article, we will present data on nurse staffing levels and education, and a description of recent changes in hospitals (duration of shift, aspects of responsibility for patients, roles in nursing care, number of RNs and HCAs on the ward, performance of various tasks, and tasks they should have completed and did not). The possible answers to each question are described in the results. The last part of the instrument includes demographic data. Using the same approach as in RN4CAST (18), forward and backward translations of the survey instrument were conducted independently by two qualified translators.

Patient data was provided by the National Institute of Public Health in Slovenia to analyse the links between patient outcomes and nurses’ working conditions. The primary outcome of interest is in-hospital mortality within 30 days of admission. Primary and secondary diagnoses were coded using the tenth version of the International Classification of Diseases (19). When modelling mortality, only patients with a primary diagnosis in ICD-10 blocks S (Injury), M (Disease of musculoskeletal system), K (Disease of digestive system), I (Disease of circulatory system), and C (Malignant neoplasm) were included. The analysis included surgical patients aged 18 years or older with a hospital stay of at least two days or more, and for whom complete data was available for comorbidities, discharge status, and other variables used for risk adjustment. Risk adjustment variables at the individual patient level include patient age, gender, admission type (emergency or elective), and the Charlson Comorbidity index (20, 21).

Permission to conduct the study was obtained from the National Medical Ethics Committee (No. 0120-488/2019/6, January 7, 2020), and all actions were carried out in accordance with relevant guidelines and regulations (22). The participants gave their written informed consent to participate in the study and permission to use the data collected at the national level for professional and scientific purposes. The informed consent to publish in scientific journals (online, open-access) was also obtained from the study participants.

Nurse staffing was calculated by averaging the patient-to-nurse ratios – as reported by each nurse – for the hospital level. The patient-to-nurse ratio at the individual nurse level was calculated by dividing the number patients by the number of nurses with an undergraduate professional bachelor’s degree present on the ward on their most recent shift. A low ratio reflects a more favourable staffing situation. The association between nurse staffing, nurse education and 30-day in-hospital mortality is estimated before and after adjusting for hospital characteristics and risk-adjustment for individual patient characteristics. Using a generalised linear model, individual patient outcomes were modelled using a combination of both hospital and patient characteristics. This approach allowed us to take the hierarchical structure of the data (i.e. patients nested within hospitals) into account. Descriptive and bivariate statistics were used to analyse the data obtained from the nurses using the instrument. Analyses were conducted using SAS version 9.4 for Windows and SPSS version 20.0.

The mean age of patients sampled was 59.62 years (SD=17.25), all other patient characteristics are provided in Table 1. Of the 36,037 patients included, 18,768 were men (52.08 %) and 17,269 (47.92 %) were women, and 29.64% had a primary diagnosis in ICD10-block M, while only 6.53% had a primary diagnosis in ICD10-block C. The most common comorbidities were diabetes without complication, congestive heart failure, and malignancy (Table 1).

Mortality data was obtained for 36,037 patients; the number of discharges varied across hospitals (Table 2). The 30-day in-hospital mortality for surgical patients was 1.00% in the hospitals sampled and ranged from 0.27% to 1.62%.

Table 3 shows the effect of nurse staffing and nurse education on 30 day in-hospital mortality, both unadjusted and adjusted for potential confounders. In the unadjusted model, only nurse education is significantly associated with mortality. After considering the severity of patients’ illness, patient characteristics, and hospital characteristics (bedside), only nurse staffing was significantly associated with mortality in the adjusted model (Table 3). The odds ratio for staffing suggests that each increase of one patient per RN is associated with a 6% increase in the likelihood of a patient dying within 30-days of admission.

The mean number of patients per inpatient ward during the last work shift of the respondents was 25.17 (SD=14,390). The mean patient-to-RN ratio was 15.56 (SD=2.50) and varied from 10.29 to 19.39 (Table 2). One HCA was responsible for an average of 8.62 (SD=5.675) patients. The difference between RNs and HCAs is significant, with the latter significantly less burdened than the former (p<0.001).

The mean number of RNs per ward during the most recent work shift was 2.06 (SD=1.412), compared to 3.60 HCAs (SD=2.800) (p<0.001).

The self-assessed number of patients for whom respondents were directly responsible during the last shift was 17.42 (SD=11.84); RNs’ assessment was significantly higher at 19.09 (SD=14.42) compared to HCAs’ assessment at 16.38 (SD=9.64) (p<0.001). A total of 55.4% respondents reported their direct responsibility for patients did not differ from the usual level, while 31.1% felt they were less burdened than usual.

The mean duration of the most recent shift was 9.42 hours (SD=5.518), with no significant differences between RNs and HCAs.

An average of 8.85 (SD=7.555) patients for whom the respondents were responsible during the last shift needed assistance with all activities of daily living and 5.40 (SD=5.923) needed monitoring for an hour or less or healthcare interventions. RNs were significantly more burdened in patients’ daily living activities (M=9.55, SD=8,83, p=0.023) compared to HCAs (M=8.41, SD=6.57), with no differences in monitoring the patients between the two nursing groups (p=919).

A total of 54.9% (n=523) respondents estimated they provided most care to the patients themselves, 31.3% (n=298) supervised other providers in addition to providing direct nursing care, 11.1% (n=106) responded that they were involved in the direct work (dressing wounds, administering medication) in a minor way, with most of the direct nursing care being provided by others. The difference is significant in the category of supervision of the nursing care provided, as RNs (46.6%, n=179) (p<0.001) act as both providers and as those responsible for the nursing care provided in the nursing team, in contrast to HCAs (21%, n=119).

Table 4 shows the frequency of the different tasks during the last work shift. RNs are typically more likely to answer the telephone, perform routine blood draws, handle patient discharge tasks, and obtain supplies or equipment. HCAs typically answer the telephone, take care of food service, clean patient rooms and equipment, perform non-nursing care, transport non-threatened patients, obtain supplies or equipment, and perform other services (Table 4).

The respondents were asked which tasks they should have performed during their last shift but did not due to lack of time (Table 5). The tasks that stood out the most were ensuring comfort and engaging in conversation with patients, educating patients and family, appropriate supervision of patients, care planning, etc. On average, four of the planned tasks were not performed during the most recent shift. More detailed results are shown in Table 5.

The findings call for the unification of staffing norms in Slovenian hospitals and the consideration of the fact that the staffing policies of developed countries aim for an 80% share of BSN-educated nurses in the nursing workforce (11). Our findings support promoting multiple BSN educational pathways to reach the Institute of Medicine recommendation (8, 9). It has been shown that reducing the nursing skill mix by adding nursing associates and other categories of assistive nursing personnel without professional nursing qualifications may contribute to preventable deaths, erode the quality and safety of hospital care, and contribute to shortages of hospital nurses (13).

It is important to adopt staffing norms that will increase the number of BSN staff by 50% over the next decade. There is also an urgent need to reorganise the work and tasks at the level of the hospital ward and hospital so that non-nursing work and tasks, particularly those performed by RNs, are delegated to non-healthcare staff involved in helping the hospital ward run and performing administrative tasks.

There is also a need to reform the nursing curricula and expand the nursing curriculum by one year to reach the BSN level. Employers should take the approach of sending HCAs to obtain higher nursing education in order to improve the ratio of HCAs to RNs as soon as possible, thereby also improving the input of nursing skills and the patient-to-RN ratio.

The results of this study have important implications for the management of Slovenian hospitals and for health policy. The long-known fact that Slovenia has the lowest proportion of adequately educated RNs according to the European Directive has been confirmed by our study. The consequences of the policy of overloading existing RNs are also related to patient mortality within 30 days of hospital admission in our study, which should be a clear message to health policymakers and healthcare managers that the long-standing limitation on the recruitment of RNs has negative consequences. Slovenia is the worst performing country for RNs in the EU. RNs are the most overworked and are also forced to omit the care that patients should receive – and here, too, the result is the worst in the EU. Economic analyses have shown that work-related complications have negative effects beyond making it difficult to recruit an adequate number of nurses and enabling them to develop their careers. Griffiths et al. (28) clearly show that staffing plans with higher baseline staff levels are highly cost effective. It is therefore high time to regard the recruitment, training and performance of nurses as an investment and not as a cost. Mandatory training in working with staff should be introduced for the appointment of health managers at macro and meso levels. Only educated managers and leaders can properly address the problems of providing an appropriate and supportive working environment in nursing, an issue which has been described by other authors in our cultural context (29-31).