This approach is based on the premise that if there are not enough nurses with appropriate qualifications, negative outcomes will occur in indicators relevant to nursing practice.43 This outlook has been increasingly and extensively developed throughout the world since the end of the last century and has provided very solid evidence regarding the consequences of nursing understaffing on negative patient outcomes. Thus, multiple studies conducted in 30 countries over two decades with thousands of hospitals, patients, and nurses demonstrate a direct association between the presence of a sufficient number of adequately qualified nurses and a reduction in adverse events such as mortality, infections, haemorrhages, thrombosis, pneumonia, medication errors, etc.1,44,45

After more than two decades of research in this area, the causal relationship between the number of nurses and the occurrence of adverse events and mortality is increasingly solid.46 In OECD countries, it has been found that approximately 15% of hospital expenditure and activity is directly derived from AEs. Thus, in well-funded and technologically advanced hospital settings, approximately one in 10 patients suffers some type of harm, and improved patient outcomes translate into lower costs for the hospital.47,48

These results are in line with data presented in the technical report “Improving Patient Safety in the European Union,” commissioned by the European Commission in 2008, which found that the percentage of AEs in hospitalised patients ranged between 8% and 12%. An estimated 4.1 million patients in the European Union suffer healthcare-related infections each year, and at least 37,000 people die from this cause.49 These infections account for 20% of all AEs in hospitalised patients. In Spain, several studies have assessed the risk of patient care, promoted by the Patient Safety Strategy of the National Health System (NHS),37 such as the National Study of Adverse Events Related to Hospitalisation (ENEAS), the Study of Adverse Events in Primary Care (APEAS), the Study on Adverse Events in Nursing Homes and Social-Health Centers (EARCAS), and the Study of Safety and Risk in Critically Ill Patients (SYREC), and others, such as the Study of Adverse Events in Emergency Departments of Spanish Hospitals (EVADUR), promoted by the Spanish Society of Emergency Medicine. All these studies agree on the high degree of preventability in the production of AEs, ranging from 50% in hospitals to 70% in primary care and emergency departments.

One of the most relevant studies is that of Harrison et al.50 with 231,000 nurses in California, Florida, New Jersey and Pennsylvania, which showed that the probability of surviving to discharge with good cerebral function after CPR increased substantially in those services where nurses had higher levels of training, in those with 70% graduate nurses it increased compared to those with 40% graduate nurses. Furthermore, the probability of surviving to discharge would be lower for patients in hospitals with caseloads of 6 patients per nurse, compared to hospitals with 4 patients per nurse. Derived from this philosophy, instruments such as the Nursing Practice Model Questionnaire (NPMQ),43 Intensive and Critical Care Nursing Competence Scale (ICCN-CS-1),51 Nurse Professional Competence (NPC),52 Healthcare Productivity Survey (HPS), Inventário de Competências Relacionais de Ajuda (ICRA),53 have emerged. Also noteworthy is the RN4CAST study, carried out in 488 hospitals in 12 European countries, including Belgium, England, Finland, Germany, Greece, Ireland, the Netherlands, Norway, Poland, Spain, Sweden, and Switzerland, which has provided valuable insight into the relationship between nursing staffing and AEs. With a sample of nearly 34,000 nurses, the findings showed an increase in the probability of dying for each additional patient assigned to a nurse. Moreover, an increase in the number of graduate nurses decreased this probability. This link implies that patients in hospitals with 60% graduate nurses and an average of 6 patients per nurse, are 30% less likely to die than in hospitals with 30% graduate nurses and a ratio of 8 patients per nurse.54

Although not aimed at nursing human resource management, increasingly more predictive tools for the onset of AEs that are sensitive to nursing practice are being proposed, such as the Predisposition to the Occurrence of Adverse Events Scale (EPEA), VALENF,55 PREDICUIDD,56 the Global Trigger Tool,57 or the extension and adaptation of early warning scales. However, sufficient research is needed to determine the extent to which these tools could act as models of the distribution of nursing resources in different contexts and practice settings.

In the development of this line of research, the concept of care left undone or missed nursing care emerged, referring to care missed or not performed and how the omission of any aspect of care leads to poor patient outcomes.58 The findings of Cho's 2015 study, comparing units with 7 patients per nurse and others with 17 patients per nurse, showed that nurses in units with higher staffing levels performed significantly more care activities, such as repositioning, feeding, bathing, oral care, and the number of assessments per shift. The omission of these activities is associated with the occurrence of pressure ulcers, falls, pneumonia, and supervision and control problems. The economic cost of care rationing has been estimated at approximately €700 per patient and additional day of hospital stay resulting from this phenomenon.59 Therefore, the results suggest that improved nursing staffing can reduce missed nursing care and, consequently, improve patient outcomes.60 This approach, which identifies missed care due to insufficient nursing staffing, has even validated instruments to assess its magnitude, such as the MISSCARE.58

There are other elements besides nursing staffing that determine patient safety outcomes, such as the practice environment. Its influence, which is conceived as “the set of organisational attributes in the workplace that facilitate or hinder professional nursing practice,”61 has been shown to be associated with the emergence of AEs. Despite the complexity of its conceptualisation, the Practice Environment Scale of the Nursing Work Index (PES-NWI) has allowed its measurement with consistent results,62 but research has made available more than thirty instruments aimed at measuring this construct.63 It typically includes dimensions such as: 1) the nurse's autonomy in decision-making regarding patient care, 2) the nurse's participation in the decision-making bodies of their center, 3) the styles of collaborative relationships between medical and nursing staff, and 4) the adequacy of staffing or concern for the quality of care. In our country, it has been proven that in primary care, a practice environment perceived by nurses improves the control of patients with high blood pressure.64

When these environments are inadequate, they can hinder the implementation of evidence-based safety interventions, exacerbating the gap between knowledge and practice, or leaving nurses without sufficient autonomy to develop their skills to their full potential. Furthermore, nurses often work under austere conditions, with increased care intensity and prioritising budgetary constraints over sustainable nursing service provision.66 This can lead to emotional stress, burnout, and higher turnover or intention to leave the job. There is also a significant correlation between nurses' age, burnout, and intention to leave the job, especially among those under 30 years of age.65,67,68

Furthermore, factors such as work shifts, overtime, shift length, the number of night shifts, and lack of adequate rest between shifts are associated with increased fatigue in nurses. For example, fatigue levels increase when nurses work more than 10 h in a single shift, which can compromise patient safety.69 Another dimension that affects the quality of care provided by nurses is job satisfaction, which has been widely studied. Job satisfaction refers to the comparison between the professional's current and ideal work situation. It is an integral part of the quality of care provided and has a direct impact on the outcomes of healthcare activities.70 In hospitals with a high patient-to-nurse ratio, nurses who perceive a poorer work environment show a greater intention to leave their jobs, which is closely related to job satisfaction.71 Conversely, high job satisfaction leads to a lower risk of stress-related illnesses, depression, anxiety, and burnout.72 The reasons why nurses may consider leaving their positions may also be linked to the interventions of their organisations' formal leaders and, therefore, potentially preventable. As nursing managers adopt transformational leadership behaviours, nurses are more likely to choose to remain in their positions.73

The concept of complexity, expressed in terms such as: complex patient, complex case, complex care, complex practice, and complex needs, refers to both clinical challenges, such as therapeutic or procedural difficulty and the chronicity of the processes, and the impact of the disease and its treatment on the patient's daily life, as well as their personal and social circumstances. Therefore, the degree of complexity is determined by the instability, variability, and uncertainty inherent in the healthcare process.77,78

There is currently a profusion of proposals, instruments, and criteria for identifying patients with complex chronic conditions, making it very difficult to establish a gold standard to guide the development of health policies, the organisation of health services, and the management and evaluation of care based on these criteria. Many proposals focus on clinical components, which are not always the most explanatory from the perspective of the frequency and demand for health services, or on functional criteria, or with a moderate use of social and behavioural criteria. Thus, a review by Davis et al.79 identified 90 definitions of complex patients, which may also have different purposes: to stratify, segment, or identify target populations for interventions. Furthermore, a high use of subjective criteria was observed in the definitions and very different purposes, with a predominance of systems that seek to identify “hyper-utilising” patients, based on cost criteria (with extreme variability in the definition of utilisation), compared to others with a more predictive approach or others that seek to capture the presence of certain clinical or functional situations. Similarly, another review that attempted to bring together the elements comprising the definitions of complexity proposed to date80 identified 83 articles that defined patient complexity. Of these, it was defined according to clinical aspects in 36.1%, a classification tool was used in 43.4%, and a conceptual model was used in only 8.4% of cases.

Schaink et al.81 proposed the complexity framework more than a decade ago to synthesize the multiple perspectives that existed until then, although it did not cover the dimensions of frailty and disability. Zullig et al., proposed the complexity cycle model82 with six domains: 1) burden and effort of the person in carrying out ADLs, 2) acute clinical events, 3) accessibility and utilisation of health services, 4) patient preferences and expectations, and 5) interpersonal, organisational, and community factors. This proposal attempts to address conceptual gaps in previously proposed models, such as those by Shippe et al.,83 Safford et al.,84 Giovannetti et al.,85 Grembowski et al.,86 etc. Furthermore, there is considerable conceptual ambiguity in the use of other factors frequently associated with the need for care, such as frailty or vulnerability.87,88 For example, the causal relationship between frailty and multimorbidity continues to present many gaps.89 Ultimately, proposals for identifying complexity may have purposes other than the need for nursing care, and may be aimed solely at the likelihood of health service utilisation, or the need for activation or referral to certain levels of care, etc.

Therefore, although it is a very suggestive and logical approach, it is not easy to resolve the current conceptual labyrinth to identify, stratify, and segment the complexity of care required by a person.

An instrument that has been validated and disseminated in Spain for the purpose of identifying complexity is INTERMED.90 Originally designed to identify care needs upon admission of patients to acute hospitalisation, it classifies patient characteristics into four dimensions (biological, psychological, social, and health services) that generate a range of scores between 0 and 60. A cutoff point of 20–21 has been proposed to differentiate complexity in hospitalised patients.91 Its use has spread to many countries and settings, including primary care,92 although validation methods have often been disperse and have had significant limitations, particularly in sample size and the determination of construct and predictive validity.

Juvé et al., proposed the Care Complexity Index (CCFI),93 which differentiates hospitalised patients according to five domains: developmental, mental-cognitive, psychoemotional, sociocultural, and comorbidity/complications. It has been used to assess the onset of AEs,94 or the use of acute services.95

The Safer Nursing Care Tool (SNCT), recommended by NICE in the British NHS,96 is an instrument for managing nursing staff based on dependency and exacerbation levels, which has been linked to the occurrence of some AEs that are sensitive to nursing practice. The SNCT classifies patients into five levels and assigns a weighted multiplier to the number of nurses required according to each level. This is based on the observation of more than 40,000 episodes of care in the British NHS and has recently been tested in Canadian hospitals,97 and in South Korea.98 However, the instrument combines both clinical characteristics of patients and interventions performed (presence of oxygen therapy; chest drains; arterial blood gases; central catheterisations; tracheostomy; mechanical ventilation; neurological monitoring; vasoactive drugs, and need for vital signs monitoring) for stratification. This means that some of its indicators are subject to the already described potential bias of interinstitutional variability when establishing certain actions (e.g., the decision to initiate arterial blood gases or the frequency of vital signs taking in the same patient could be different depending on the hospital they are admitted to). An evaluation in the British NHS has shown a 3% reduction in mortality for each additional hour of nursing staffing99 (HR: 0.97; 95% CI: 0.94–1.00), with mortality increasing when the rate of nurses on temporary contracts increased, although it was not able to identify a recommended cut-off threshold for the instrument.100

In another methodological approach, the Nursing Outcomes Classification (NOC) has been proposed as a framework for developing research on health outcomes. It allows for the assessment of states, behaviours, or perceptions susceptible to nursing intervention, regardless of the institution in which the patient is admitted, as it only uses indicators of the patient's condition.101

Additionally, it has another advantage: its focus on outcomes sensitive to nursing practice, that is, those attributable to the presence or absence of an intervention on a previous health condition.102 Since 2006, research at INICIARE in Spain has been developing the design and validation of a tool based on these characteristics.103 After more than a decade, supported by several multicenter projects in 20 of the 28 hospitals of the Andalusian Public Health System, with different levels of complexity and management, and which have covered nearly 10,000 patients and more than 2000 professionals, the Level of Care Inventory through Nursing Outcomes Classification Indicators (INICIARE) instrument provides a structured evaluation format, developed from standardised nursing language and compatible with the digital clinical history, using a coding system included in the different health information systems1. INICIARE is based on a solid conceptual substrate, unlike many other proposals, with the capacity to describe a large number of phenomena and care situations.104 It has shown excellent reliability and internal consistency,104 with good construct validity in its short 26-item versión.105 It is an independent predictor of hospital mortality,76 has shown good acceptance for use by nurses in qualitative studies,106 and has been adapted to other settings such as Brazil.107 It is free of items related to nursing activities and is based on a conceptual approach that sought to avoid at all costs the inclusion of items derived from the actions of professionals, due to the institutional bias generated by the variability of practice styles and decision-making styles in identical clinical situations. Although this possibility was originally raised, during the statistical analysis, the difficulties encountered in validating items pertaining to nursing interventions forced them to be discarded in the internal consistency analysis phase.80 Except for the SNCT, neither INTERMED nor INICIARE have yet established indices for weighting nursing resources based on the identified level of care. However, in the case of INICIARE, it is at a maximum level of development. It has recently been evaluated in 6 countries in the Mediterranean basin (Italy, Greece, Lebanon, Egypt, Spain and Tunisia) on 20,000 users of hospital, primary and nursing home care108 and has also provided study on the incorporation of the latest advances in Machine Learning and intelligent models, which will enable the comparison of the distribution of nurses who care for patients with similar levels of dependency or care needs in different hospital.109

The origin of INICIARE stems from a practice setting conceived by care units not by diseases or medical specialities, the special unit or hospitalisation unit in the Alto Guadalquivir Hospital.80 As a result the research has been based on offering health services a tool that aids modernisation and adaptation of the convention hospitalisation units, in a reorganisation focused on care needs and care complexity.89