Effectiveness of a Standardized Nursing Process Using NANDA International, Nursing Interventions Classification and Nursing Outcome Classification Terminologies: A Systematic Review

1. Introduction

The nursing process (NP) is the scientific method used by nurses to identify, diagnose, intervene in and resolve health issues in the population within the scope of their disciplinary field. Its implementation demands cognitive, psychomotor and affective skills and capacities that underlie the clinical reasoning and care provided by nurses [1]. Each stage of the NP involves carrying out strategies to address the observed phenomenon, from the aspects concerned to the establishment of clinical judgment, including the gathering of information and recognition of health patterns, along with decision-making to determine the main and secondary interventions required for its resolution [2].

The phenomena and activities of nurses can be defined and described using standardized language systems (SLS) through the retrieval of data from electronic records [3]. The use of such nursing terminologies in the scientific literature has been variable, with up to 72% of published studies using NANDA International (NANDA-I) [4] or its combination with Nursing Interventions Classification (NIC) [5] and Nursing Outcome Classification (NOC) [6], thus establishing itself as the most widely-used system by nurses in the international context [7].

Through the review of the scientific literature with regard to the use of NANDA-NIC-NOC (NNN) in clinical practice, it is possible to assess the effectiveness of NP through the successful integration of NNN in electronic health records and validation of concepts.

Two systematic reviews have recently been published that address the use of standardized nursing terminologies [8-9], but have not focused on the exact topic of NNN terminologies. For this assessment, the following review question was posed: Do nursing diagnoses, interventions and people’s health outcomes improve when nurses use standardized NNN terminology?

The main research aim is to synthesize the evidence on the effectiveness of NP using standardized terminology in relation to the benefits of using NNN in care practice, thus improving diagnostic accuracy, nursing interventions, health outcomes and people’s satisfaction.

2. Materials and Methods

A systematic review was carried out following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), 2020 statement [10]. The research protocol was registered in the National Institute for Health Research (NIHR) International Database of Prospective Register of Systematic Reviews (PROSPERO); registration number CRD42020170350.

2.1. Sources of Information

The first step consisted of identifying previous publications on the subject of interest through various searches in PROSPERO and Google Scholar^® that could answer the research question. After this initial check, search strategies were employed in the following databases: Medline and PreMedline (through OvidSP), Embase (through Embase), The Cochrane Library (through Wiley), CINAHL (through EbscoHOST), SCI-EXPANDED, SSCI and Scielo (through WOS) and LILACS (through the Health Virtual Library). To complement these, manual searches were carried out in the Trip Database metasearch engine.

2.2. Search Methods

Searches were conducted on the 12th and 13th of January 2021, establishing methodological limits to publications after 1992, the year in which NNN terminology was officially recognized. Search strategies included the following terms: “nursing interventions classification” OR “nursing outcomes classification” OR “nanda international” OR “nnn terminology” in the title and abstract fields. Similarly, those MeSH most in line with the defined search strategy were selected from the thesaurus of each of the databases. The search strategy was first checked by a documentalist in the Embase database and independently reviewed by two of the authors. Once the definitive strategy was designed, it was adapted to the remaining databases selected.

2.3. Inclusion Criteria

Studies with the following design methodologies were included: Randomized clinical trials (RCT), quasi-experimental (non-randomized clinical trials and pre-post studies) and observational (cohort, case-control, case series), which consider NP assessing the use of NNN in English, Spanish and Portuguese language.

2.4. Exclusion Criteria

Systematic reviews (SR) methodologies and other than those described in the inclusion criteria were excluded. Similarly, studies which did not consider NP assessing the use of NNN were also excluded. Failure to meet any one of these criteria was sufficient for the study to be excluded.

2.5. Quality Appraisal

The records were exported to an Excel^® spreadsheet for the selection process. Following the elimination of duplicates, studies were screened by title and abstract and classified into three groups: “potentially eligible”, “doubtful eligibility” and “excluded”. “Potentially eligible” and “doubtful eligibility” records were retrieved for full-text screening. The process was carried out by two independent reviewers and a third reviewer was consulted in the case of discrepancies. To determine study suitability, Critical Appraisal Skills Programme Español (CASPe) templates appropriate to each type of design were used, so that for cohorts, case-control and RCT (11 items), scores ≤ 5 were considered low quality, scores 6-8 moderate quality and scores ≥ 9 high quality. To verify the suitability of the process, a pilot test was carried out on an initial record sample.

The certainty of the evidence was assessed (random sequence, allocation concealment) blinding bias of participants and researchers (concealment of allocation to study arm, intention to blind, method of blinding, blinding effectiveness), blinding bias to outcome assessors (reported, requiring researcher judgment, not requiring researcher judgment), attrition bias (incomplete data, omitted from analysis) and reporting bias (selective outcome reporting), identifying each as: low risk, high risk, uncertain risk or not applicable. A pilot test of bias risk assessment was conducted on a sample of studies. Bias risk was considered in determining the degree of certainty of the evidence using grading of recommendations assessment, development and evaluation (GRADE) methodology.

2.6. Data Abstraction

The research outcomes analysed correspond to information on improvements in diagnostic accuracy, efficacy of interventions, health outcomes and people’s satisfaction. Separately, general study data were extracted. Data extraction was performed independently by two researchers and resolved through consensus with a third researcher in the case of discrepancies. The Mendeley^® bibliographic reference manager was used for data extraction and recorded in detail in the data extraction document. A pilot test of the extraction process was carried out on a sample of studies.

2.7. Synthesis

To organize the presentation of results, firstly, criteria established by Joanna Briggs Institute (JBI) was followed to determine the levels of evidence (LE) for the effectiveness of each of the studies. The results were then organized according to the research outcomes below.

3. Results

The number of records identified was n = 4455; following elimination of n = 1545 duplicates, the number was n = 2910. During the title and abstract screening process n = 2820 were excluded, limiting the number of retrievable full-text records to n = 90. Of these, n = 4 could not be retrieved, so that the number of studies assessed for eligibility was n = 86, of which n = 69 did not satisfy inclusion criteria. Thus, the final number of included studies was n = 17, as can be seen in the flow chart below in Figure 1.

Following the screening process, those studies meeting eligibility criteria were distributed among the authors for critical reading in pairs (CARS-CEMA; PRBB-MNHDL; DAFG-HGDLT) and subsequent measurement of interobserver agreement through determination of Cohen’s weighted kappa coefficient are shown in Table S1: Interobserver agreement on included studies. When the coefficient did not reach statistical significance, a third reviewer was consulted (CARS and MNHDL) to resolve agreement discrepancies.

All the studies showed high or moderate quality following critical reading with CASPe. The studies that showed high quality were the RCT (score 9/11) by Corcoles et al. [11], Guerra et al. [12], Gencbas et al. [13] and Sampaio et al. [14]. The remaining studies showed moderate quality in Table S2: Critical reading scores for the included studies.

With regard to the design methodology, the studies included 9 experimental designs (5 RCT, 1 pseudo RCT and 3 quasi-experimental) and 8 observational (1 case control and 7 cohort), which are shown together with sociodemographic characteristics in Table 1.

Following GRADE methodology criteria, the overall quality of the certainty of scientific evidence was determined for each of the outcomes assessed. GRADE stipulates that studies with experimental designs show greater initial certainty, while observational studies do so with lesser initial certainty, although following application of compensation criteria for lowering or raising the quality of this initial certainty corresponding to each of the GRADE domains, their estimation is corrected. Final certainty was shown to be high in the studies outcomes by Corcoles et al. [11], Silva et al. [15], Pascoal et al. [16], Silva et al. [17], Pascoal et al. [18], Reis & Jesus [19] and Pascoal et al. [20]. JBI criteria were simultaneously applied to assign the level of evidence to each one, as shown in Table S3: JBI level of evidence and degree of certainty using GRADE methodology.

Regarding research outcomes, the included studies assessed improvements in diagnostic accuracy (n = 6) and in people’s health outcomes (n = 11). No studies were identified that assessed outcomes in the efficacy of interventions or improvements in population satisfaction.

3.1. Diagnostic Accuracy

Studies assessing diagnostic accuracy of NANDA-I determined the effectiveness of related factors (RF) (n = 3) and defining characteristics (DC) (n = 3).

The NANDA-I nursing diagnoses that addressed the effectiveness of the RF were: Risk of delayed surgical recovery (00246), Dysfunctional ventilatory response to weaning (00034) and Risk of falls (00155). The effect measures of these RF were found to be statistically significant in most of the etiological indicators assessed, as shown in Table 2.

The articles that assessed the effectiveness of the DC (n = 3) concerned the NANDA-I nursing diagnoses: Impaired gas exchange (00030), Ineffective airway clearance (00031) and Ineffective respiratory pattern (00032), as shown in Table 3.

3.2. People’s Health Outcomes

Articles that addressed effectiveness in people’s health outcomes did so from two perspectives.

First, regarding the general aspects of effectiveness (n = 2). On the one hand, with respect to the assessment of care planning using NNN and, on the other, concerning clinical reasoning. The study carried out by Cárdenas-Valladolid et al. [21] evaluated the implementation of care planning in primary care centers using standardized NNN terminology in the intervention group (IG) compared to the usual recording of non-standardized care as a control group (CG) through prospective follow-up of a cohort (n = 23488) over 2 years, demonstrating that both groups experienced a moderate reduction in cardiovascular risk factors observed at 12, 18 and 24 months for systolic blood pressure (SBP), diastolic blood pressure (DBP), glycosylated hemoglobin (HbA1c), LDL cholesterol and body mass index (BMI). The effect measure improved in the IG for all outcomes except LDL cholesterol and DBP. Following adjustment of the reference parameters for age, sex, type of treatment and physical activity, a reducing effect was observed in all outcomes except HbA1c, which was statistically significant for DBP (mean = -0.33 (CI = -0.63-0.04); p = 0.02). In general, the changes in the values for SBP, DBP, HbA1c, LDL cholesterol and BMI were greater in the IG than the CG, despite only reaching statistical significance in favour of the IG in HbA1c (p<0.01), while the CG reached statistical significance in SBP (p<0.01).

With regard to clinical reasoning, Müller-Staub et al. [21] developed a training program for nurses using guided clinical reasoning as an IG, compared with nurses who received training through classic discussion of clinical cases as a CG, showing greater acquisition of critical thinking skills for the application of NNN in clinical practice in the IG due to better internal consistency between diagnoses, interventions and outcomes, as shown in Table 4.

Secondly, studies that assessed the effectiveness of health outcomes in specific situations (n = 9) corresponded to the NANDA-I nursing diagnoses: Functional urinary incontinence (00020), Risk of falls (00155), Ineffective health management (00078), Risk of perioperative postural injury (00087), Ineffective airway clearance (00031), Nutritional imbalance: less than the body needs (00002), Anxiety (00146) and Sleep pattern disorder (00198). These studies assessed the interrelationship of NANDA-I diagnosis respect to NIC and NOC terminologies. On the other hand, Guerra et al. [12] did not use NOC terminology to measure the effect of fall prevention on the reduction on risk of falls, while Bjorklund-Lima et al. [22] assessed the risk of perioperative postural injury using various NOCs but without reporting the NICs performed in the NP.

The statistically significant effect measures for each of the indicators of effectiveness on improving people’s health outcomes are shown in Table 5.

4. Discussion

Brazil is the context with the greatest number of publications included, showing a marked tendency to explore aspects related to the clinical applicability of NNN, while Spain ranked second with a distinct emphasis on the growing interest in the study of nursing terminologies in our environment. The increase in the use and effectiveness of nursing SLS in clinical practice is accompanied by improvements in the diagnostic reasoning capacities of the nurses [24].

Regarding the quality of evidence in these studies, the use of traditional systems such as the proposal by JBI to establish the LE has been refined with the application of GRADE methodology such that it is possible to adjust the focus and quality of the initial evidence rating granted according to the design of these studies’ methodologies, readjusting the factors or domains that confer the final certainty of the evidence to reduce it (assessing the risk of bias, inconsistency, indirectness, inaccuracy and publication bias) or increase it (assessing magnitude of effect, response gradient, and absence of residual confounding) with greater certainty [25].

As background to this research, a study conducted by Müller-Staub et al. [26] assessed, among other aspects, the accuracy of the Standardized Nursing Terminology, in addition to the coherence between diagnoses, interventions and people’s health results. The authors identified deficits in the diagnostic process as well as in the notification of signs, symptoms and aetiologies, arguing for the need to implement training measures that ensure accuracy in nurses’ diagnostic reasoning [27]. To complement these criteria, the present study adds the importance of linking nurses’ critical thinking to the use of clinical indicators based on the best scientific evidence available from the results of rigorous research.

With respect to diagnostic accuracy, the diagnoses Risk of delayed surgical recovery (00246), includes people aged over 80 years in the NANDA-I classification, although the SR only reported results that indicated absence of statistical significance in this population with extreme ages. In contrast, the remaining aetiologies presented showed semantic variations.

Concerning the diagnoses Dysfunctional ventilatory response to weaning (00034), most of the statistically significant RF reported by Silva et al. [15] were not included.

Regarding the analysis of diagnostic accuracy through the study of DC, the diagnoses Impaired gas exchange (00030) showed that abnormal skin colour and hypoxemia indicate the presence of this health issue with greater statistical accuracy. These DC that could be considered higher are found to be included in the 2021-2013 NANDA-I edition [4] along with a substantial number of DC with lower predictive accuracy to reach clinical judgment. As such, it would be beneficial to add diagnostic accuracy criteria that distinguish between major and minor DC to NANDA-I. The diagnoses Ineffective airway clearance (00031), showed that the only DC not included in the 2021-2023 NANDA-I edition (4) correlates statistically significant with Open eyes, albeit with an excessively wide CI. On the other hand, the diagnoses Ineffective respiratory pattern (00032) showed effectiveness for diagnostic accuracy in all DC, including others that were not observed in the study, suggesting that it would be valuable in future research to assess the rest of the DC included in NANDA-I.

In the effectiveness analysis for the resolution of specific health issues, certain modifications or the elimination of some diagnoses in the latest published edition of NANDA-I were notable [4]. Thus, Functional urinary incontinence (00020) was replaced by another diagnoses called Disability associated urinary incontinence (00297). Similarly, in the 2021-2023 NANDA-I edition, the diagnoses Risk of falls (00155) was removed from the classification and replaced by new diagnosis which distinguish between the population of adults, with the diagnoses Risk of falls in adults (00303), and Risk of falls in children (00306). Likewise, for the diagnoses Dysfunctional ventilatory response to weaning (00034), the 2021-2023 NANDA-I edition included diagnoses called Dysfunctional ventilatory response to adult weaning, which differs from the previous definition by specifying that it refers to individuals over 18 who required mechanical ventilation for at least 24 hours.

In recent years, there has been growing interest among nurses in studying the clinical application of NNN with more rigorous methodological designs, including cohort studies with adequate follow-up and randomized interventions with control groups that estimate the risk of bias. However, it is still essential to diversify international contexts and sample sizes in the populations studied with the aim of increasing effect measures in the population. Separately, it is vital that the results of these studies are transferred more quickly to the subsequent published NNN editions in order to improve nurses’ clinical impact.

The limitations of the current research are due to the heterogeneity of the studies included in the SR, addressing distinct clinical situations corresponding to various health issues and NNN labels independently, which prevents comparison of results and the accumulated meta-analysis of their effect measures. Taking this into account, future research should examine larger sample sizes and the effect of longer follow-up periods in the populations studied.

5. Conclusions

It must be concluded that, at present, scientific literature using NNN is very extensive but that there is still a deficit regarding the amount and quality of evidence, and the degree of certainty concerning the effectiveness of the NP using these terminologies. It is essential to increase the number of studies with rigorous methodologies that approach diagnostic accuracy and the health results in people using NNN terminologies from the clinical perspective. Similarly, it is important to implement the findings of new studies that assess the use of these terminologies with respect to improvements in the efficacy of nursing interventions and the satisfaction of the population with NP.