The use of the Nursing Activities Score in clinical settings: an integrative review

Objective: analyze how studies have approached the results obtained from the application of the Nursing Activities Score (NAS) based on Donabedian’s model of healthcare organization and delivery. Method: CINAHL and PubMed databases were searched for papers published between 2003 and March 2015. Results: 36 articles that met the inclusion criteria were reviewed and double-coded by three independent coders and analyzed based on the three elements of Donabedian’s health care quality framework: structure, process and outcome. The most frequently addressed, but not always tested, variables were those that fell into the structure category. Conclusion: variables that fell into the process category were used less frequently. Beside NAS, the most frequently used variables in the outcome category were mortality and length of stay. However, no study used a quality framework for healthcare or NAS to evaluate costs, and it is recommended that further research should explore this approach. DeScRiPtORS Workload; Nursing Staff; Intensive Care Units; Quality of Health Care; Review. Correspondence Addressed to: Maria Cecilia Gallani 1050, Avenue de la Médecine Quebec, Canada maria-cecilia.gallani@fsi.ulaval.ca USP


IntroduCtIon
Intensive care units (ICUs) are constantly evolving and in recent decades have had to deal with an increasing number of old patients and comorbidities.This in turn leads to a greater number of complex procedures (1) , which calls for closer monitoring and a greater need for nursing interventions, resulting in increased nurse workload and higher costs.
In fact, nursing resources are the largest single component of direct ICU costs, accounting for nearly 50% of the ICU budget (2) .Thus, the use of more objective strategies to evaluate nursing activities is essential to optimize the use of ICU resources and ensure appropriateness, efficacy and cost-effectiveness (3) .Besides this, in the current global context, marked by resource constraints and nursing shortages, it is imperative to portray a more accurate and comprehensive picture of nursing and the value and benefits of the services they provide in line with preestablished objectives and standards (4) .The use of tools for nursing workload measurement serves as a starting point for the analysis, comparison, and processing of data related to nursing activities and can provide important inputs to promote the optimization of the use of resources and establish a more adequate nurse-patient ratio (3) .
Many attempts have been made to characterize nursing activities in ICUs (3) .Beginning with the Therapeutic Intervention Scoring System (TISS) (5) , a number of different tools have been developed, such as the Italian Time-oriented Score System (TOSS), the Canadian Project Research in Nursing (PRN), the TISS-28 (6) and the Nine Equivalents of Nursing Manpower use score (NEMS) (7) ; the latter two of which evaluate severity of illness and complexity of therapeutic procedures.However, it is known that the relationship between severity of illness and the time spent on nursing interventions is not linear, since severity accounts for only part of the variability of nursing activities.Therefore, in 2003, the Nursing Activities Score (NAS) was developed to provide an overall description of all nursing activities in ICUs, not only those correlated to the severity of illness (2) .It has been shown that NAS represents about 80% of the work activities of nurses in ICUs, which is substantially more representative that the 43% obtained using TISS-28 (2) .NAS has been used in over a dozen countries (8) and a number of papers have been published regarding its clinical application.Some studies have demonstrated a relationship between NAS and outcomes such as mortality and adverse events; however, this association is not constant throughout all studies and this inconstancy is yet to be explained.
Nursing care performance involves "the analysis of multiple interacting elements that relate to the diverse aspects of nursing services, their antecedents and their results", reflecting "nursing care as a complex, aggregate entity, comprised of multiple interrelated and interdependent subsystems and components that are logically coordinated and oriented toward the achievement of common goals" (4) .In this sense, Donabedian's model of healthcare organiza-tion and delivery guides understanding on how inputs are acquired from the (nursing) care environment and fed into the service production cycle, where transformation of resources results in changes in patient conditions (9) .
In light of the above, this integrative review assesses how studies have approached the results obtained using this tool based on Donabedian's model of healthcare organization and delivery and aims to contextualize the use of NAS across the healthcare continuum to gain further insight into the interpretation of its results.

Method
The review included the following steps: problem identification, literature search, data extraction, and data analysis and reporting (10) .The PICO strategy was used to formulate the question, whereby the population was defined as "Intensive Care Units" and intervention as "application of the Nursing Activities Score -NAS".Comparison criteria were not applied and outcome was analyzed from an exploratory perspective, in accordance with Donabedian's model (9) .
Problem identification: the central question of this integrative review was: "what are the main purposes for using NAS in ICU settings?The following aspects were analyzed: study objective, the context in which NAS was used, sample size (for patient and NAS application), the strategy used for applying the tool, and the purpose of using NAS, according to the SPO elements of Donabedian's model: structure/process/outcome.
Search strategy: a search was conducted of the CINAHL and PubMed databases of articles produced between 2003, which was when the first article on NAS was produced, and March 2015, using the following concepts: 1) "Nursing Activities Score" or "Nursing Activity Score" and "scale"; 2) "intensive care units" or "critical care"; 3) "workload measurement" or "workload".This search strategy was adapted to the terminology of the databases.In addition, the bibliographies of potentially eligible studies were analyzed manually to identify studies not brought up by the search strategy.
Eligibility: the following inclusion criteria were used -1) fully peer-reviewed papers published in English, French, Italian, Portuguese or Spanish; 2) papers should demonstrate the use of NAS in a clinical setting; and 3) papers should indicate the context in which the tool was used (for example type of ICU -adult, pediatric, general, clinical); 4) papers should specify the method of NAS application; and 5) specify the purpose for using the tool.Both exploratory and correlational studies were included since they met the mentioned criteria.Studies addressing cross-cultural adaptation, conceptual aspects of the tool, and the measurement properties of NAS were excluded; as were published abstracts or papers without abstracts found to be lacking information on the context in which the NAS was used, methods of application, or the purpose for using the tool during data extraction.Articles reporting the use of NAS in settings other than ICUs were excluded, considering that the original focus of the tool were ICUs (2) .
The studies were sorted by title and abstract by JL in order to select the articles that met the inclusion criteria.Duplicated articles were discarded.Thereafter, the full text of the articles that met the inclusion criteria were inde-pendently screened by JL, FD and MCG and compared.A total of 36 papers were selected.Details of the methodological steps are shown in a flowchart developed in accordance with the PRISMA Statement (11) (Figure 1).Records after duplicates removed n=137 Record identified through databases searching n=187 Studies included in qualitative synthesis n=36 Full-text assessed for eligibility n=67 Data extraction: first coding was carried out with papers to check agreement of data extraction and to confirm the quality of the coding sheet.Where there was disagreement between the two reviewers, the final decision was taken after discussion with reviewer MCG for language reasons.
Data analysis: NAS serves to assess the percentage of time spent on nursing interventions delivered to an ICU patient.Thus, the results obtained with NAS could serve as one of the means to evaluate the quality of health services provided by nurses in ICU.A multitude of models or frameworks were developed for health services quality improvements such as Donabedian (9) and more recently (4) .Since Donabedian framework is already largely used in nursing studies to evaluate quality of care (12)(13) , this model was chosen to analyze the retained studies.
Donabedian's framework evaluates three elements of the quality of health service delivery: structure, process and outcome, whereby each component is interdependent and influenced by the other components (9) .Structure refers to all technical, human and organizational resources related to the service or institution, and stable patient characteristics, such as age and gender.In this study, other patient characteristics at the time of admission that affect process and cannot be influenced by the nurse were included in the structure category, such as clinical severity, patient origin, and reason for admission to ICU.Process includes all activities carried out by professionals involved in direct patient care, care coordination, and the interaction between and response to these activities.Outcome includes changes in the health status of a patient after direct contact with the health system such as healing, survival, mortality and adverse events.It should be noted that this analysis of the SPO does not imply a separation between means and ends, but rather an unbroken care process chain continuum (9) .Thus, NAS can be seen as a connecting element between the structure and process components, since it enables the quantification of the nursing activities that characterize the process.

results
A large majority of the articles reviewed by this study were published in nursing journals (86.1%), while 9.3% were published in multidisciplinary journals  (Chart 1). With rgard to the nursing journals, 24% were journals specifically focusing on critical care or cardiac nursing, while the rest were general nursing journals.Year of publication varied from 2005 to 2014, with a major concentration (81.4%) in the period 2010 to 2014, indicating that the use of NAS for clinical purposes is a more recent phenomenon.Most articles were produced in Brazil (69.4%), followed by Europe (27.8%) and one publication (2.8%) from China.Patient sample size varied from 34 to 5,856; only one paper failed to mention the number of patients who participated in the study.Sixteen articles did not clearly state the number of NAS applications.Those articles that stated the number of applications reported between 34 and 28,390 applications including patient follow-up.The large amount of variability between the samples made it impossible to compare study data.All studies indicated the context in which NAS was used, the application method, and the purpose for using the tool.(20) , 2006/Brazil Acta Paul Enferm.
To compare nursing care needs of elderly and non-elderly patients.
To verify the association between nursing workload and pressure ulcer risk and illness severity.
To compare illness severity and nursing workload and to identify predictors of nursing workload for the treatment of elderly patients.
To evaluate the demand for nursing care.
To identify the factors affecting the length of stay in hospital.Gerasimou-Angelidi et al. (30) , 2014/Greece J Nurs Manag.
To study the association between family satisfaction and nursing workload.NAS was used in one or more ICUs and some studies compared different ICUs.The studies targeted different types of ICUs, and therefore patients: clinical, clinical-surgical, and specialized ICUs (cardiac, neurological, and burns).Five studies reported the use of NAS with pediatric or neonatal patients.

The use of NAs ANd The NursiNg cAre sysTem
The use of NAS across the unbroken care process chain continuum was analyzed (9) .The variables described in the studies were categorized into one of the three elements of the framework: structure, process and outcome.
The variables that fell into the structure category were far more frequent and were related to the stable patient characteristics (age and gender), clinical profile at admission (reason for admission, origin, associated clinical conditions, organ dysfunction, illness severity and mortality risk, ulcer pressure risk) and environmental conditions (type of unit, nurse-patient ratio, bed-occupancy rate, professional background of the nursing team, absenteeism).However, it is important to highlight that these variables were not always tested against NAS: for example, age and gender were mentioned by 33 and 29 studies, respectively, but only nine studies tested the association between NAS and age and only four tested the association between NAS and gender.Severity of illness and risk of mortality were tested against NAS in 14 and four studies respectively, and only mentioned in six and two studies, respectively.Other variables tested against NAS included: clinical condition (one study), origin (five studies), reason for admission (five studies), different types of unit or distribution (four studies), risk of pressure ulcer, and bed occupancy rate (one study).
With regard to process, besides NAS, few variables were described, and even fewer variables were tested  (47) 2012/Brazil Acta Paul Enferm.
To explore the correlation between patient illness severity and nursing workload.against NAS: TISS-28 in two studies, type of treatment (described in six studies and tested in one study), and the interval between emergency department-ICU.
With respect to outcome category, the most frequently used variables were mortality and length of stay (LOS).Mortality was mentioned in 20 papers and tested against NAS in 14 studies.LOS was measured in 28 papers, and effectively tested in 10 studies.Additionally, a range of different variables that fall into this category were used and tested against NAS by separate studies: clinical evolution, nosocomial infection, readmission, HAI, adverse events, noninjurious incidents and family satisfaction.
Finally, no study reported using a quality framework for healthcare or NAS to evaluate costs.

dIsCussIon
This literature review summarizes studies assessing the use of NAS in given contexts, application methods, and the purpose for using this tool in settings.
The large majority of studies concerning the clinical use of NAS were published in nursing journals, which is to be expected given that NAS deals specifically with nursing interventions.However, some studies were published in multidisciplinary journals, which may be viewed as a strategic effort to increase the visibility of the unique role of nursing in ICUs.
The majority of studies were conducted in Brazil during the last four years, indicating that this country has contributed most to the clinical application of this tool.Considering that NAS is widely used around the world, particularly in Europe, it is very important to stimulate the application of this tool from a research perspective and the dissemination of results in Europe.
Important insights about nursing activities in ICU and their contribution to the continuum of healthcare could be gained by comparing the results of the application of NAS in different cultures.International collaborations involving multicenter studies could be an interesting way of making this possible.
NAS was applied in various types of ICUs, including neonatal and pediatric units.However, it is important to mention that NAS was developed for the adult ICU context (2) and, to date, validation studies of the NAS in these settings do not exist.
Regarding the SPO elements of Donabedian's framework, it was observed that the most frequently described, but not always tested, variables were those related to stable patient characteristics (age and gender) and severity of illness, measured using different tools (APACHE, SAPS).A wide range of variables that fall into the structure category were also mentioned; however, the lack of a minimum set of standardized variables makes it impossible to compare the different contexts in which NAS was applied.Considering the significant impact of structure on process (and thus on nursing activities) and, consequently, on outcomes, the determination of a minimum set of variables could make a particularly important contribution to improving the analysis and interpretation of NAS results.Moreover, it is important to explore how these variables may influence NAS and outcomes.
Beside NAS, the studies mentioned few variables that fall into the process category.Type of treatment was mentioned in seven studies and tested against NAS in only one.It is important to determine which other variables in the process category influence NAS results and their effects on outcome.For example, coordination of care is a variable that potentially falls into the process category, since it is influenced by the setting and other structure variables and has a presumed causal effect on patient outcome (50) .
The most frequently measured variables in the outcome category were LOS and mortality.These are key-variables in the evaluation of ICU outcomes, since these units have higher mortality rates than other hospital units.In the United States, there are approximately four million ICU admissions per year, with average mortality rates of between 8 and 19%, which is equivalent to around 500,000 deaths per year (51) .Furthermore, ICU is a costly component of the health care budget and this cost is largely accounted for by LOS.Despite the importance of these outcome variables, they should not be used in isolation from other variables to indicate the impact of nursing care since they reflect the contribution (or failure) of several systems of care and healthcare disciplines.LOS and mortality are part of a group of outcomes in the nursing care performance framework (4) that reflect the joint contribution of nursing care, health status, readmission, and complications.Other quality indicators that are more sensitive to nursing interventions include: (a) patient safety (for example patient falls, injuries, medication errors, pulmonary, intravenous and urinary tract infections, pressure ulcers); (b) patient comfort and quality of life in the care context (meeting patients' needs in relation to personal hygiene, nutrition, management of symptoms such as pain and dyspnoea, and continence, and avoiding unnecessary interventions during hospitalization, such as physical or chemical restraints, nasogastric tubes or prolonged use of urinary catheters); (c) changes in patients or families' knowledge, skills, and behaviors; (d) patients' functional status (physical, psychosocial and cognitive); and (e) patient and family satisfaction with the care experience.The systematic inclusion of such variables by studies using NAS could make a significant contribution to the evaluation of the impact of nursing activities (4) .This review also highlights the importance of a framework to guide the use of NAS in clinical practice, the elaboration of good hypotheses for studies of NAS and its implications for healthcare quality, and the interpretation of results.Furthermore, despite the fact that one of the underlying aims of NAS is to promote effective cost management and utilization of nursing resources (30) , which can account for up to 50% of the total spending of an ICU (2) , this review shows that there is a lack research exploring the use of NAS to evaluate and manage costs and further research in this area is therefore highly recommended (2) .

DeScRiPtOReS
Carga de Trabajo; Personal de Enfermería; Unidades de Cuidados Intensivos; Calidad de la Atención de Salud; Revisión.reFerenCes meta-analysis was not conducted due to the variability and lack of information in some studies.

ConClusIon
This review shows that NAS has been applied in clinical settings in various types of ICUs.The analysis of the results based on Donabedian's model indicates that NAS was used to test several variables that fall into the structure category (principally age, sex and severity of illness), but few variables related to process.With regard to outcome, the most frequently tested variables were mortality and LOS, which are not nurse-sensitive variables.Further research should consider the use of a conceptual model to guide the use of NAS and the interpretation of results, and address the notion of health economics.

Figure 1 -
Figure 1 -Flowchart showing the methodological steps of the integrative review.

*
Variables tested against NAS; AIS: Abbreviated Injury Scale; APACHE: Acute Physiology and Chronic Health Evaluation; ER: Emergency Room; HAI: Healthcare-Associated Infections; ICU: Intensive Care Unit; ISS: Injury Severity Score; IU: semi-intensive unit; LODS: Logistic Organ Dysfunction System; LOS: Length of Stay; NAS: Nursing Activities Score; NEMS: Nine Equivalents of Nursing Manpower Use Score; NS: Not Specified; PU: Pressure Ulcer; SAPS: Simplified Acute Physiology Score; SOFA: Sepsis Related Organ Failure Assessment;TISS-28: Therapeutic Intervention Scoring System; RASS: Richmond Agitation Sedation Scale.

Chart 1 -
Summary of the studies of the application of NAS in clinical settings analyzed by the present study.
The use of the Nursing Activities Score in clinical settings: an integrative review www.ee.usp.br/reeusp