Nurse staffing levels within acute care: results of a national day of care survey

Acute Medical Units (AMUs) are a core component of the acute care pathway in the UK [1]. The AMU model has been replicated in other international health care systems [2]. AMUs provide a single location within each hospital dedicated to receiving acutely unwell medical patients deemed to require inpatient care following initial assessment in the Emergency Department or in the community. Care is typically provided for a short period, before either discharge home or transfer to another inpatient ward for admission. The presence of adequate numbers of appropriately trained nursing staff on the AMU is a tenet of safe care. Hospitals with higher nurse-to-patient ratios tend to report better patient outcomes [3‐6]. In the UK, national policy makers and regulatory bodies have recognised the link between adequate staffing levels, appropriate levels of training and patient safety [7, 8]. Despite this, concerns regarding nurse staffing levels in acute care persist. Staff shortages are a pervasive problem throughout the health-care system but are particularly prevalent in acute care, where many AMU nursing positions are vacant [9]. Unmanageable workloads and a perceived lack of career progression are cited as causes of low morale and poor retention of senior staff, which perpetuate staffing shortages [10].

Nursing on the AMU provides a unique set of challenges. High rates of patient turnover, in combination with high levels of patient acuity, create a demanding environment enriched for risk. Year-on-year increases in the number of emergency medical admissions and seasonal surges in demand compound this problem [11, 12]. Low staffing levels have been shown to affect the ability of individual nurses to respond to abnormal physiology [13]. Failure to identify the deteriorating patient and escalate care appropriately is a well recognised cause of harm [14]. The Royal College of Nursing (RCN) have raised the concerns of members that inadequate staffing levels across the UK are compromising patient safety and have called for staffing decisions to be based on effective models of care, with coherent workforce planning as opposed to cost and availability [15].

Acknowledgement of the importance of staffing levels has not translated into a clear understanding of what constitutes safe and adequate staffing. The Society for Acute Medicine (SAM) has published relevant guidelines, recommending nurse-to-bed ratios of 1:4 for patients on trolleys and 1:6 for patients in beds [16]. While these recommendations provide a useful standard to facilitate measurement and benchmarking, the use of fixed staffing ratios in isolation is problematic. Nurse to bed ratios do not take into account the pace of work driven by patient flow, the skill-mix of the team or relevant patient factors, all of which are likely to influence the number of nurses required. The National Institute for Health and Care Excellence (NICE) have provided specific guidance to aid decision making in relation to nurse staffing levels, where the time required to complete common patient orientated tasks is estimated, to allow a more accurate prediction of the number of staff required [17]. However, these guidelines are based on expert opinion without the benefit of robust empirical evidence.

The Society for Acute Medicine Benchmarking Audit (SAMBA) has run annually since 2012, and by convention takes place on the penultimate Thursday in June [18]. SAMBA records patient level data from individual clinical records to assess performance across a range of pre-specified criteria related to acute medical care. The audit combines organisational and patient level data to provide a detailed picture of how acute medical care is delivered in the UK. We used data from the 2019 SAMBA (SAMBA19) to better understand the provision of nurse staffing resources within the AMU and the skill-mix of the nursing workforce. Our aim was to describe the national nursing staffing levels currently present on AMUs in the UK, including evaluation of skill-mix. We also looked to and compare staffing levels to estimates of the total time required to provide nursing care to patients admitted to the AMU during the study period. This is the first time data on nurse staffing levels has been reported using SAMBA.

Recruitment to SAMBA19 was open to all hospitals in the UK receiving acutely unwell (non-elective, adult) medical patients. Non-acute and community hospitals were excluded. SAMBA19 was undertaken on Thursday 27th June 2019, between 00:00 and 23:59. SAMBA19 was the first SAMBA survey to collect data regarding nurse staffing levels and skill-mix.

SAMBA is a cross-sectional audit with an observational design. Patients aged 16 years or above who were seen for either admission or assessment as part of the internal medicine admission process (‘take’) or ambulatory emergency care (AEC) were eligible for inclusion. Patient level data collection was performed voluntarily by local clinicians, overseen by a named responsible consultant physician. The SAMBA protocol is publicly available [19].

Prior to the collection of patient level data, a lead acute medicine clinician at each participating AMU completed an online questionnaire to provide organisational and unit level data, relating to staffing levels on the AMU. Data on staffing levels could only be submitted once from each AMU. The questions regarding nursing numbers, level of nursing experience and the specific roles performed by staff were developed by a group of nurses affiliated with SAM. Total number of nurses present on the AMU during the 24-h data collection window was measured, with nurse-to-bed ratios during daytime (08:00–19:59) and nighttime (20:00–07:59) hours calculated using the reported number of available beds within each AMU (occupied and unoccupied). The number of non-registered nurses, band 5 nurses, band 6 nurses and band 7 or above nurses were recorded separately [20].

The proportion of units which achieved the SAM standard was calculated. In order to determine whether the use of non-registered nurses was higher in units with fewer registered nurses, the ratios of registered to non-registered nurses were compared between units with nursing levels above and below the recommended SAM standard.

The unit and patient level data collected were used to generate an estimate of the total number of nursing hours available within all contributing AMUs. This estimate was compared with an estimate of the anticipated time required to complete nursing tasks informed by the relevant NICE guidance on safe staffing levels [17]. The total number of available AMU beds taken from the unit level data were combined with the total number of admissions to each AMU recorded at the patient level.

To estimate the amount of nursing time required to complete tasks at the system level. This was achieved by multiplying the anticipated number of occupied beds, or the number of AMU admissions, by the proportion of patients deemed likely to require each nursing intervention and the time the time taken to perform each nursing intervention as stated by the NICE guidelines.

These estimates were made under multiple assumptions. A bed-occupancy rate of 90% was used to inform the estimate, consistent with national averages during the study period [21]. It was anticipated that patients with features of frailty would have increased nursing needs. Patients with features of frailty were anticipated to require complex admission assessment and discharge planning. The prevalence of frailty was approximated at 50% based on previous evidence [18, 22]. The frequency of physiological observations was based on the distribution of NEWS2 values obtained on admission, combined with national guidance on the frequency of observation [31].

All admitted patients were assumed to require administration of medications (including assessment of whether ‘as required’ medications were needed), 80% were assumed to require administration of intravenous medication [23]. Intravenous fluid therapy was felt likely to be required in 50% of patients [24] (rate of 8 hourly in 30%, > 8 hourly in 10% and complex fluid management involving hourly fluid balance in 10%) [25]. All patients were deemed to require some form of direct communication each day (routine in 50% and significant in 50%). Consensus opinion as to whether these assumptions reflected clinical experience was obtained by discussion (HH, TK, CA), and disagreements reviewed by all authors. The assumptions made for individual tasks are provided in Table 1.

The administration of oral and parenteral medications and clinical tasks associated with care coordination were deemed appropriate for registered nursing staff in accordance with NMC guidance [26]. Tasks suitable for non-registered nurses were identified and summated separately. The number of nursing hours required to meet the demands of nursing care on the day of SAMBA was compared with the actual number of nursing hours available recorded by individual units. It was assumed that matrons and ward managers did not undertake direct clinical care, due to other responsibilities.

SAMBA19 collected paired unit and patient data from 130 AMUs. Complete data relating to nurse staffing and workforce planning was provided by 122 AMUs (93.8%), with individual patient level data collected for 5849 patients seen through these units, 8 AMUs did not provide data on nurse staffing levels and were not included in the analysis.

The most common working pattern, employed by 70 units (57.4%) was exclusive 12-h shifts. A combination of 12-h and 8-h shifts were implemented in 51 units (41.8%). Only one unit used solely 8-h shifts. During the daytime, 1159 nurses were on duty; this included 913 staff nurses and 246 senior nurses (ratio 3.7:1). At nighttime, 970 nurses were on duty; this included 801 staff nurses and 169 senior nurses (ratio 4.7:1).

SAMBA19 provides insight into nurse staffing levels in acute medical care at the system-level. The results suggest a short-fall in staffing levels within the 24-h study period. SAM recommend a nurse to bed ratio of 1:6, but this was not achieved in approximately 20% of AMUs during daytime hours. Performance against the standard was considerably poorer at night. Nighttime staffing was also characterised by a lower ratio of band 6 or higher nurses to more junior colleagues.

Estimates of the time required to perform direct clinical tasks, based on current NICE guidelines, suggest a significant gap between the time required to deliver care and the total number of nursing hours available. These estimates only calculate the nursing time required for direct clinical care. The provision of direct clinical care does not include time between tasks or documentation, which represent a significant proportion, if not the majority, of time used in delivering nursing care. A previous observational study using a validated tool to measure working patterns and the time distribution of nursing tasks suggested only a fifth of available time was devoted to activities involving direct patient contact [27]. The study findings did not account for the additional nursing time required to provide high dependency level care to those at the severest end of the acuity spectrum with conditions such as septic shock, respiratory failure requiring non-invasive ventilation, or diabetic ketoacidosis. These conditions are relatively common on the AMU and place significant additional demands on nursing staff. Furthermore, these study findings do not account for nursing time devoted to bed coordination; this is a complex and time consuming role often designated to a member of the nursing team. The ability to perform direct clinical activity is likely to be severely limited when undertaking the coordination role. Combining all these factors together, nurse staffing levels in AMUs are under-resourced.

Nurse to bed ratios in isolation are recognised as a poor measure for staffing requirements as they do not adequately account for the multitude of factors that determine whether demand and supply are adequately matched [28, 29]. NICE recommend the use of an in-depth analysis of demand and skill-mix incorporating a time task approach to ensure safe staffing levels [17]. Nursing levels should be matched to the acuity and complexity of the patients and to the pace of the work dictated by patient flow. In this regard, staff planning requires consideration of patient factors, the staffing skill-mix in terms of experience and competencies, and ward factors that relate to the process of care and added responsibilities. Patients managed on AMU, particularly during the first 24-h, are at the peak severity of their illness [30]. This is the point at which the need for medical intervention and intensive monitoring is highest and the probability of deterioration greatest. Many patients are reaching the end-of-life and in these circumstances compassionate nursing care requires time to ensure both patients and relatives receive appropriate support. The changing nature of acute medical admissions, with an increase in multiple co-morbidities and frailty, also creates pressures on the time required to provide quality care. It is not unusual for patients on the AMU to require 1:1 care to ensure their safety. The variable nature of acute care can make workforce planning challenging. The results of this study raise concerns as to whether current staffing levels are adequate for the demands placed on nursing staff to provide quality care. This deficiency in staffing has the potential to materially affect outcomes and the experience of care by acute medical patients.

The deficit in the number of hours to provide care was driven by tasks that could be appropriate for non-registered nurses. Classification of tasks on this basis is a simplification of reality. It would be expected that a proportion of a registered nurses time would involve delivering basic care in addition to more complex tasks which would be inappropriate for non-registered colleagues. Although our estimates suggest a surplus of hours to perform tasks only appropriate for registered nurses, this did not account for the indirect time required to perform these tasks, such as documentation, collecting of equipment and communicating with other members of the medical team. Not accounting for this indirect time is likely to have led to a significant under-estimate of the time required to conduct these tasks. Increasing the number of non-registered nurses may help reduce some of the gap between anticipated workload and available staff but is unlikely to provide a solution to inadequate staffing levels in isolation. High ratios of non-registered to registered nurses have been linked to higher mortality during in-patient care [31]. The development of new roles such as nursing associates, with enhanced spectrum of practice, may help provide additional nursing care while also performing some tasks traditionally only performed by registered nurses [32].

The NHS staff survey highlighted the relationship between overstretched staff with excessive workloads and poor morale [33]. These workforce factors are known to affect patient care [6, 34]. A work environment in which staff feel overstretched and without the necessary resources to provide high quality care leads to stress and staff burnout [35]. In the NHS, nursing positions in acute medical care have the highest vacancy rates [9]. High staff turnover and poor retention inhibits the passage of knowledge from senior staff to those less experienced. The principles laid out in the Interim NHS People Plan to increase the intake of new nurses and ensure appropriate career progression should be implemented urgently [36].

This UK-wide study suggests that despite reasonable compliance with current SAM staffing levels, when based on a nurse per bed ratio, in reality acute medical care faces a staffing crisis among nurses. There is not enough nursing time within the system to complete direct clinical tasks, as predicted by our study findings. Ultimately our findings highlight the need for an increase in the number of nurses working in AMUs. A concerted effort to recruit and retain nursing staff is urgently required. A valid fear would be declining quality in patient care, especially if the current situation deteriorates further.