Critical Need to Address Accuracy of Nurses’ Diagnoses

  • Margaret Lunney, RN, PhD
    Margaret Lunney, RN, PhD

    Margaret Lunney is a Professor and Graduate Nursing Programs Coordinator at the College of Staten Island, the City University of the New York (CUNY), and Doctoral Faculty at CUNY’s Graduate Center. For 25 years, her research and professional activities have focused on community health nursing, critical thinking, and the concept of accuracy of nurses’ diagnoses, measurement of accuracy, and use of standardized nursing languages. Among the honors she received are the 2001 Distinguished Nurse Researcher Award from the New York State Nurses’ Foundation and a 2007 Fulbright Award as research consultant and lecturer in Japan. She received her M.S. in nursing from Hunter College, CUNY, and her PhD in Nursing Science from New York University.

Abstract

Studies published from 1966 to 2006 describe how nurses’ interpretations of clinical data vary widely, thus significant percentages of nurses’ diagnoses may be of low accuracy. This is important because data interpretations, or diagnoses, serve as the basis for selection of interventions and the subsequent achievement of patient outcomes. Accuracy of nurses’ diagnoses is defined as a rater’s judgment of the match between a diagnostic statement and patient data. Low accuracy can lead to wasted time and energy, harm to patients, absence of positive outcomes, and patient and family dissatisfaction. The purpose of this article is to appeal to nurses in both practice and education to address the accuracy of nurses’ diagnoses. This appeal is based on three factors: (a) research evidence indicates the need for greater consistency among nurses in making nurses diagnoses, (b) accuracy of nurses’ diagnoses will always be an issue of concern because diagnosis in nursing is complex, and (c) with implementation of electronic health records, the degree of accuracy of nurses’ diagnoses will have broad-based implications. In this article, the need for nurses to be accountable for addressing diagnostic accuracy is explained and strategies to improve accuracy related to the diagnostician, the diagnostic task, and the situational context are recommended. Some of these strategies include a greater focus on educational methods and content for development of nurses as diagnosticians, adoption of partnership models of nurse-patient relationships, an increase in opportunities for critical thinking and clinical decision making, selection of software with appropriate structures and content libraries, and a change in health care policies.

Key words: administration, clinical judgement, education, evidence-based practice, nursing diagnosis

The research evidence is strong that it is time to address the accuracy of nurses’ diagnoses... Accuracy of nurses’ data interpretations (diagnoses) should be a serious concern of nurses in both practice and education because interpretations of patient data serve as the basis for selecting the nursing interventions that will achieve positive patient outcomes. Accuracy of nurses’ diagnoses is defined as a rater’s judgment of the match between a diagnostic statement and patient data (Lunney, 1990, 2001).

The research evidence is strong that it is time to address the accuracy of nurses’ diagnoses and consider strategies to improve accuracy. In an analysis of 20 studies published from 1966 to 2000, Lunney (2001) reported that in all clinical simulation studies, and also in a study involving clinical cases, nurses’ interpretations of the same data varied widely. Since 2000, investigators of nurses’ clinical reasoning and critical thinking abilities have also indicated that interpretations from the same data vary from nurse to nurse (Brannon & Carson, 2003; Ebright, Patterson, Chalko, & Render, 2003; Ferrario, 2003; Hicks, Merritt, & Elstein, 2003; Junnola, Eriksson, Salantera, & Lauri, 2002; Puntillo, Neighbor, O’Neill, & Nixon, 2003; Redden & Wooten, 2001; Reischman & Yarandi, 2002). When interpretations vary, some of the interpretations represent low accuracy. This is serious because low diagnostic accuracy contributes to harm to patients through: wasted time and energy, implementing ineffective interventions, absence of positive outcomes, and patient and family dissatisfaction.

...diagnosis in nursing is complex [and]...low diagnostic accuracy contributes to harm to patients...A concern about the accuracy of nurses’ diagnoses relates to all nursing care whether or not standardized nursing diagnoses, such as NANDA International (NANDA-I, 2007), are used. Nursing interventions are based on data interpretations, whether or not nursing diagnoses are stated. For example, even the act of a nurse deciding to help a patient move from a bed to a chair is based on the nurse’s interpretation of that patient’s data. Data interpretations are considered as diagnoses when they are complex enough to vary in accuracy and when they serve as the basis for interventions. For example, interpreting a patient’s skin color as pale is not complex enough to serve as the basis for interventions. Instead, it is a data element that may contribute to making a nursing diagnosis, such as Acute Pain, Fear, or Fluid Volume Deficit.

The purpose of this paper is to appeal for nurses in all settings to address the accuracy of nurses’ diagnoses of human responses. This appeal is based on three factors: (a) research evidence indicates the need for greater consistency among nurses in making nursing diagnoses, (b) accuracy of nurses’ diagnoses will always be an issue of concern because diagnosis in nursing is complex, and (c) with implementation of electronic health records, the degree of accuracy of nurses’ diagnoses will have broad-based implications. Nurses’ accountability for accuracy is described and strategies to achieve accuracy of nurses’ diagnoses are presented.

Appeal for Nurses in Practice and Education to Address the Accuracy of Nurses’ Diagnoses

The need to address the accuracy of nurses’ diagnoses derives from three major factors. First, there is sufficient research-based evidence in relation to the diagnostician, the nature of the diagnostic task, and situational contexts to establish that diagnostic accuracy varies widely. Second, accuracy of nurses’ diagnoses will always be an issue of concern because the diagnosis of human responses is complex. Third, implementation of electronic health records (EHRs) will result in the degree of accuracy having more significant effects on nursing care outcomes and the quality of nursing care than paper records have had.

Research-Based Evidence that Accuracy Varies

Research evidence supports that nurses’ interpretations of patient data vary widely. Hence the matter of diagnostic accuracy needs to be addressed. Carnevali and Thomas (1993) and Gordon (1994), theorists describing the diagnostic process in nursing, have identified the many factors influencing variations in accuracy. These factors have been classified as the: (a) diagnostician, (b) nature of the diagnostic task, and (c) situational context. Research reports related to diagnostic accuracy in each of these areas will be discussed below.

Diagnostician. Because nurses interpret patient data to make human response diagnoses, such as pain, they are considered diagnosticians. The characteristics of nurse diagnosticians, such as experience, education, and abilities in intellectual, interpersonal, and technical domains are important influencing factors in clinical decision making, including accuracy of diagnosing human responses. In a clinical study, for example, only eight (12.9%) of 62 staff nurses in three hospitals achieved the highest accuracy score on a seven point scale of diagnostic accuracy when rated for accuracy by two clinical nurse experts who assessed the same patients (Lunney, Karlik, Kiss & Murphy, 1997).

The research support is strong...in indicating that teaching nursing students about nursing diagnoses and the use of the diagnostic process is associated with higher accuracy. Regarding the element of experience, it is not years of experience in nursing that influences accuracy of diagnosis, but rather the experience of working with the same types of patients (Carnevali & Thomas, 1993; Gordon, 1994; Lunney, 2001). This is consistent with Benner’s (1984) findings related to the value of nursing experience, and with the findings of cognitive scientists that the critical thinking needed for clinical judgments is not a generic skill but one that is knowledge-specific (Willingham, 2007). In  addition, nursing experience can generate both good and bad habits of mind, habits known as heuristics. Good habits facilitate positive effects on accuracy whereas bad habits may decrease accuracy (Brannon & Carson, 2003; Thompson, 2003). One heuristic that would have a negative effect on accuracy, for example, is overconfidence. Overconfidence occurs with a belief that experience automatically provides excellent ability to interpret clinical data; it represents flawed self-assessment (Gambrill, 2005).

The research support is weak that higher educational levels are related to higher diagnostic accuracy (Lunney, 2001). The research support is strong, however, in indicating that teaching nursing students about nursing diagnoses and the use of the diagnostic process is associated with higher accuracy. Eight studies were reported to support this latter relationship (Lunney, 2001).

In relation to nurses’ abilities in the intellectual domain, a large number of studies have shown that critical thinking and clinical reasoning abilities vary widely resulting in variance in diagnostic ability (e.g., Brannon & Carson, 2003; Ebright, et al., 2003; Ferrario, 2003; Hicks, et al., 2003; Junnola, et al., 2002; Puntillo, et al., 2003; Redden & Wooten, 2001; Reischman & Yarandi, 2002). In relation to specific types of critical thinking and clinical reasoning, it is not possible to draw conclusions because, in both older and recent studies, a tremendous variety of theoretical frameworks were used, making it difficult to combine study results for knowledge development. One conclusion that can be drawn from the above-cited research, however, is that a variety of thinking abilities are needed for diagnostic accuracy based on the complexity and variety of clinical cases.

Abilities in the interpersonal domain are also important for accuracy because the formation of trusting relationships with patients and families enables them to share important personal information (Carnevali & Thomas, 1993; Gordon, 1994; Lunney, 2001). In a review of 200 studies related to critical thinking, Tanner (2006) concluded that one of the factors that influence nurses’ critical thinking was the nurse-patient relationship. In a qualitative study of eight patients’ experiences of how nurses communicate, McCabe (2004) found that patients experienced differences associated with patient-centered communications versus task-oriented communications. Regarding patient-centered communications, patients noted that the nurses spoke to them using their names and asking about their experiences. In contrast, patients described task-oriented communications as those that focused on getting a job done, such as administration of medications, without considering the patients’ perspectives. A lack of patient-centered communications was noted in this study. Patient-centered communications are needed for accurate assessment of human responses.

Patient-centered communications are needed for accurate assessment of human responses. Abilities in the technical domain are needed to conduct diagnosis-focused assessments, such as pain or fluid volume shifts, and physical examinations, as well as ability to understand data from equipment, such as mechanical ventilators, all of which affect the quality of data interpretation and thus nursing interventions (McCaffrey & Ferrell, 1997; Puntillo et al., 2003; Redden & Wotton, 2001; Wilkinson, 2001). The diagnosis of pain is the only nursing diagnosis that has been adequately studied from the perspective of diagnosis-focused assessment, and these studies have shown variations in accuracy (e.g., McCaffrey & Ferrell, 1997; Puntillo et al., 2003).

Nature of the Diagnostic Task. The nature of the diagnostic task refers to clinical situations that influence the interpretation of human responses to health problems and life processes. Factors within the nature of the diagnostic task that have been studied are relevance of data, amounts of data, and complexity of the diagnostic task (Lunney, 2001). 

The relevance of data to diagnoses is generally considered as high, moderate, and low relevance (Carnevali & Thomas, 1993; Gordon, 1994). Having adequate numbers and types of high relevance data helps to ensure diagnostic accuracy. Studies from the early 1980’s showed that high amounts of low relevance data are associated with low accuracy (Lunney, 2001). An example of this is when nurses collect admission data without a concern for ‘what is the diagnosis.’ The results are data that have no specific implications for helping patients. In comparing expert and novice cue utilization when making critical care diagnoses, Reischman and Yarandi (2002) found that experts used more high relevance cues and made more accurate diagnoses.

Another factor related to the diagnostic task is the amount of data. Studies have shown that less information may be associated with higher accuracy (Lunney, 2001). With fewer data points, nurses are more likely to move to diagnostic-specific associations of data with interpretations and arrive at better diagnoses.

Regarding the complexity of diagnostic tasks, a few studies have confirmed the logic that increased complexity is associated with lower accuracy (Lunney, 2001). The more aspects of the human condition that are involved in the diagnostic task, the more complex the situation may be. Consider, for example, a woman who attends a clinic for management of diabetes, presenting with  a broken arm, and evidence that the broken arm may have occurred as a result of domestic violence. This is a much more complex situation to diagnose than just helping such a patient learn how to manage diabetes. If nurses in this clinic seldom encounter such cases, this woman would present a complex case for them. Lack of adequate resources to assist with data collection and interpretation will also increase the complexity of diagnosing such cases.

Situational Contexts. One situational factor that relates to accuracy is the adequacy of available resources. In health care settings, a high nurse-patient ratio is a factor that most likely affects accuracy; however, this factor has not been directly studied to date. High nurse-patient ratios have been shown to contribute to low quality of care (Bostick, Rantz, Flesner, & Riggs, 2006). With high nurse-patient ratios, nurses do not have time to form trusting relationships with patients, to collect valid and reliable data, or to think about diagnostic decisions.

With high nurse-patient ratios, nurses do not have time to form trusting relationships with patients, to collect valid and reliable data, or to think about diagnostic decisions. Another situational factor that has been studied is the environmental aspects of the setting. The environmental aspect of interruptions of nurses as they are working with patients has been studied through observations of nurses in clinical settings (Potter, Boxerman, Wolf, Marshall, Grayson, Sledge, & Evanoff, 2004; Potter Wolf, Boxerman, Grayson, Sledge, Donagan & Ebanoff, 2005; Hedberg & Larsson, 2004). These researchers reported that interruptions, such as staff inquiries or missing supplies, disturbed the continuity of nurses’ thinking processes while giving care, which most likely affected their decision making. The investigators concluded that environmental elements need to be taken into account in studies of nurses’ decision making.

Use of standardized nursing languages such as NANDA International (NANDA-I, 2007) is a third situational factor influencing accuracy of diagnosis. NANDA-I  provides the nurse with possible diagnoses to consider and the associated signs and symptoms for making accurate diagnoses. In a pioneering study conducted in Switzerland, Müller-Staub and colleagues (2007) demonstrated that teaching nurses about nursing diagnosis and how to accurately diagnose and document diagnoses, interventions, and outcomes had positive effects on the quality of care as measured by documented patient outcomes. In retrospective analyses of 123, 241 patient admissions in one hospital, Welton and Halloran (2005) concluded that use of nursing diagnoses was an independent predictor of patient hospital outcomes. When nursing diagnoses were added to the DRG model, the explanatory power of the discharge dataset improved by 30% to 146%, and was significantly associated with patient outcomes, such as length of stay and disposition to nursing homes. The effects of using standardized nursing languages on nurses’ power to help the children they serve and influence outcomes were measured in a pilot study with 12 nurses and 220 children in two groups (Lunney, Parker, Fiore, Cavendish, & Pulcini, 2004). One group of six nurses used NANDA-I (2007), the Nursing Interventions Classification (NIC) (Dochterman & Bulechek, 2004), and the Nursing Outcomes Classification (NOC, Moorhead, Johnson, & Maas, 2004). The other group used only the nursing terms that were included in the computerized software. There were no significant differences between the two groups on nurses’ power and children’s outcomes but, in post-study interviews with the nurses, the nurses reported use of nursing languages helped them to better focus their assessment processes and use these terms as a basis for discussions with children and families. In a study of the heuristics that nurses used in diagnosing, Ferrario (2003) concluded that use of standardized nursing terms may make the diagnostic thinking process more efficient.

Diagnosis in Nursing is Complex

The discipline of nursing, with its focus on the health of human beings, may be the most complex science that exists (Webster, 1984). The complexity of diagnosing human responses is clearly illustrated in nursing case studies, such as those described in Lunney, 2001.  The complexity of the environments in which nurses work was substantiated in both the United States and Sweden. This environmental complexity adds to the challenge of achieving high accuracy (Bucknall, 2003; Hedberg & Larsson, 2004; Potter et al., 2004, 2005). Furthermore, lack of attention to the issue of diagnostic accuracy compounds the problem of low accuracy (Lunney, 1998). Accuracy needs to be a stated goal in order to achieve higher accuracy.

Implementation of Electronic Health Records

The implementation of electronic health records (EHRs) is imminent, with a goal in the United States of having all health records in an electronic format by 2015 (United States (U.S.) Department of Health & Human Resources, 2004). With implementation of electronic health records (EHRs), the relevance of accurate interpretations of patient data to the quality of nursing care will be greater than it is with paper records (Institute of Medicine [IOM], 2004; Olsson, Lymberts, & Whitehouse, 2004). EHRs provide better organization and ease of noting key information such as diagnoses, interventions, and outcomes, so improved continuity of care is expected (IOM, 2004). Nurses’ diagnoses will easily be identified for follow-up by other nurses and for data aggregation to describe and compare patients’ experiences across settings and localities. The accuracy of nurses’ diagnoses will also influence the effectiveness of interventions provided by health care professionals in other disciplines as they increasingly rely on a nurse’s diagnosis in selecting their specific intervention(s).

Accountability for Accuracy

Accuracy of nurses’ diagnoses is the foundation for achieving positive outcomes through use of nursing interventions... Because research studies document variance in nurses’ diagnoses, and variance means that some diagnoses are not accurate, nurses in both practice and education are encouraged to consider their accountability for accuracy of diagnoses. Levin, Lunney, and Krainovich-Miller (2005), for example, applied the five steps of evidenced-based medicine, as described by Sackett, Strauss, Richardson, Rosenberg, and Haynes (2000), to show how diagnostic accuracy in nursing can be improved through use of research evidence and patient preferences. A new PCD model (Population, Cue Cluster, Differential Diagnoses) was proposed for the first step of evidence-based practice, i.e., asking answerable questions. The five-step, evidence-based process, of (a) ask answerable questions, (b) find the best evidence to answer the questions, (c) appraise the validity of the evidence, (d) integrate the evidence with experience and patient preferences, and (e) evaluate the effectiveness of the first four steps, was explained as it pertains to accuracy of nurses’ diagnoses. Accuracy of nurses’ diagnoses is the foundation for achieving positive outcomes through use of nursing interventions, either with or without the use of standardized nursing diagnoses from NANDA-I or other diagnostic languages. When nurses act on their interpretations of data, they are acting on diagnoses, whether or not the diagnoses are stated.

Accountability for nurses’ accuracy has been limited in the past but may be improving. Prior to 2006, there was very little discussion about this issue in English language journals, whether these journals were primarily research-based, theory-based, or anecdotal. A 2006 CINAHL search using accuracy of nurses’ diagnoses as the search term, yielded eight articles, including the clinical study of accuracy previously mentioned that was described in two sources, three papers that summarized the research on critical thinking and accuracy, and three position papers, all of which were written by Lunney and others.  A 2006 Medline/PubMed search of the National Library of Medicine database using the same search term yielded 38 studies, but only eight of these were relevant and additional to those in the CINAHL search. These eight studies consisted of a 1980 study and seven studies related to pain. A 2006 Medline/PubMed search using nurses’ decision making yielded 2000 references, which included many unrelated articles. There were four additional studies of nurses’ data interpretations and 37 studies related to clinical decision making. In these 37 studies, the expectation that nurses interpretations vary was implied.

In April 2007, evidence of increased emphasis on accuracy was noted at a conference of the Association of Common European Nursing Diagnoses Interventions and Outcomes (ACENDIO). The keynote address by Daniel Pesut focused on the importance of clinical reasoning and identifying a keystone diagnosis. A keystone diagnosis is an alternate term for the most accurate diagnosis; it is the diagnosis that best matches the complexity of the patient’s story and provides guidance for patient outcomes and nursing interventions.  Various studies related to accuracy were also presented by researchers. For example,  Paquay, Wouters, Debaillie, and Geys, (2007) studied the convergent and discriminant validity of 101 NANDA nursing diagnoses for 1,952 patients with dependency problems in activities of daily living (ADL). Nursing diagnoses were formulated for 80% of the patients and most indicators of convergent and discriminant validity of the 101 NANDA diagnoses were highly significant. The authors are planning to study the validity of nurses’ diagnoses in relation to other domains besides ADL and the accuracy of nurses’ diagnoses in relation to diagnostic reasoning.

Even though wide variations in accuracy can be expected based on complex factors in each of the three categories of the (a) diagnostician; (b) diagnostic task, and (c) situational contexts, as well as the interactions of these three categories, the implications of low accuracy are significant. Widespread implementation of EHRs means that data will be aggregated to describe nursing care. Data that are based on low accuracy diagnoses will be misleading, if not useless. Nurses in practice (staff nurses, leaders, and administrators) and nursing educators need to be more diligent in promoting and measuring the accuracy of nurses’ diagnoses.

Strategies To Promote Accuracy of Nurses’ Diagnoses

Nurses both in practice and education can promote accuracy of nurses’ diagnoses by making changes as needed in the three types of factors that influence diagnostic accuracy: the diagnostician, nature of the diagnostic task, and the situational context. Strategies related to each of these factors are offered below. References, which provide further information on implementing each strategy, are also provided.

Diagnostician

Nurses and students can be encouraged to recognize ambiguity and find ways to address it,...nurse educators should help students think of themselves as developing diagnosticians. Strategies for nurses to develop as diagnosticians include assuming the image and role of being a diagnostician, accepting the ambiguity of clinical judgment in nursing, promoting the principle of working in partnership with patients and families, collaborating with interdisciplinary team members, and facilitating the use of critical thinking processes, including reflection. Brief examples of strategies in each category are provided below.

Strategies for nurses to assume the image and role of being a diagnostician can be identified and applied by all nurses. Beginning with the very first nursing course, nurse educators should help students think of themselves as developing diagnosticians. This image needs to be reinforced and supported in all nursing education experiences, whether they be in institutions of higher education or nursing practice settings, through the teaching methods and content selected. Teaching methods that require students to think, make decisions, and develop clinical judgments are needed. Teaching through lectures provides knowledge but does not help students think through how to use the knowledge in clinical settings. In contrast, problem-based learning, a technique that requires students to find the answers to problems, helps students develop problem-solving skills, for example, by asking students to identify the diagnosis (Amos & White, 1998; Beers, 2005; Williams, 2001).

Although some faculty fear that adequate content will not be presented when using problem-based learning, Beers (2005) reported that there were no significant differences in knowledge acquisition, as indicated by objective test scores, between two  groups of students who received the same content, but with one group receiving a lecture and the other receiving problem-based learning. Additionally Williams (2001) noted that problem-based learning, active learning, and self-directed learning provided additional learning advantages over lecture methods in clinical settings and in application of previous knowledge to actual clinical cases. One advantage is that it gives students opportunities to use the types of thinking that they later will need for clinical application of knowledge. Willingham (2007) a cognitive scientist, says that teachers should give students repeated opportunities to practice the types of thinking that will be needed in real situations. Another advantage is that students learn to answer questions posed by the teacher as well as to ask questions that can be answered through the use of research and other types of evidence.  Since problem-based learning is not always feasible without changing an entire curricula and may be costly (Beers, 2005), other strategies for active, reflective, and integrated learning should be considered. For example, in practice settings, staff nurses can assume the role of diagnostician and present the rationale for the diagnosis they have selected to be critiqued by their peers. Administrators can encourage nurses’ discussions of  intellectual, interpersonal, and technical aspects of nursing, rather than encouraging only task completion.

In reference to accepting the ambiguities of clinical judgments in nursing, nursing students need to be told, and nurses need to be reminded, that because they are helping people with complex responses to health problems and life processes, errors in diagnosing are possible and must be acknowledged. Nurses and students can be encouraged to recognize ambiguity and find ways to address it, for example, by seeking consultation from other health care providers, validating impressions with patients and families, and becoming acquainted with current, professional literature. Lunney (2001) and Sandstrom (2006) used small group discussions of written case studies in which the group members discussed different possible diagnoses to help students develop tolerance for ambiguity as they noted that other learners, too, had difficulty identifying the best diagnosis. Use of case studies can enable learners to comprehend the complexity of diagnosing human responses as they learn to analyze, synthesize, evaluate, and apply knowledge. Practicing nurses can support tolerance for ambiguity by accepting the difficulties that new nurses’ experience with data interpretations and helping them to identify variations that occur with given contextual factors such as age and culture. Nurse administrators can support nurses in dealing with ambiguity by marketing the value of using nursing diagnoses, along with the complexities of doing so, as they interact with other disciplines, other administrators, and the public.

Patients are more likely to work toward the achievement of improved health outcomes if they agree with the nurses’ diagnoses and proposed interventions. Educators and practicing nurses can apply the principle of working in partnership with patients and families and the principle of working in collaboration with interdisciplinary team members, as standards of care. Partnership processes with patients and families are essential for diagnosing human responses. Patients are more likely to work toward the achievement of improved health outcomes if they agree with the nurses’ diagnoses and proposed interventions. Leenerts and Teel (2006) developed a promising strategy to achieve partnership, called Self Care Talk. This relational conversation method is a combination of the four specific communication skills of (a) listening with intent (e.g., listening for habits, concerns, and needs related to self care), affirming emotions (e.g., conveying recognition and respect for emotions), creating relational images (e.g., helping patients to create positive images of past experiences that they can relate to current experiences), and using planned enactment (e.g., asking patients to describe plans for self care). Collaboration with interdisciplinary team members is essential because other health providers may have knowledge and insights about patients and families that are not known by nurses.

The need for critical thinking skills has been explained by many nurse theorists and researchers as previously cited. Knowledge of tools for critical thinking can facilitate the making of clinical judgments. Journal writing is a tool that can be used by educators to help students learn to reflect on their critical thinking in clinical decision making (Lunney, 2001). For experienced nurses, the development and use of appropriate heuristics can  help them draw upon their past experiences to strengthen their clinical judgments. The representativeness heuristic, described by  Brannon and Carson (2003) and  Thompson (2003) is one such heuristic. The representativeness heuristic is a characteristic of experienced clinicians in which “detailed analyses and probability assessments are replaced with representations that include computations of similarity to previous experiences, evaluation of associations and exemplars, and attributions of causality” (Ferrario, 2003, p. 44). This heuristic increases the efficiency and effectiveness of diagnostic reasoning. Some heuristics, however, such as overconfidence have negative effects on clinical judgment. Thompson (2003) provided strategies to combat overconfidence and other types of biases that develop with experience. Developing the ability to find and use appropriate research for a given care situation is one such strategy. Another strategy is helping nurses to “know what they do not know and revise estimates of correctness accordingly” (p. 233), which is known as calibration. Two ways to increase calibration is to think of reasons that a decision might be wrong and to identify other possible explanations besides the current explanation being considered. When overconfidence is addressed by others, it needs to be discussed with respect and care so nurses can maintain their self esteem while improving their accuracy as diagnosticians.

The questioning of students and nurses in relation to the care of patients for purposes of stimulating critical thinking is a method that can easily be incorporated into lecture, clinical supervision, and preceptorships (Browne & Keeley, 2007; Rubenfeld & Scheffer, 2006). The use of higher level questions, i.e. questions which require learners to analyze, synthesize, evaluate, or apply knowledge, as compared with questions that ask only for a recall of knowledge, also promotes critical thinking (Phillips & Duke, 2001). Lecturers, for example, can use these higher level questions throughout a class to stimulate students’ thinking about the related issues and responding to the teacher’s questions, rather than merely writing what the teacher has said. Use of higher level questions can be used throughout the curriculum to develop the critical thinking skills of the students.

Nature of the Diagnostic Task

One strategy to manage the complexity of clinical situations in nursing is identification of  diagnoses, interventions, and outcomes that are common with specific populations, such as diabetic patients, and development of standards of practice based on this information that will provide decisional support for nurses. A consensus validation research process can be used, for example, to identify the common diagnoses, interventions, and outcomes for local (geographical) populations.  (Carlson, 2006) reported that this strategy helped nurse stakeholders reduce the number of NANDA-I (2007) nursing diagnoses to consider from 197 to about 20 that were relevant to a given population and local setting. 

Additional decision-support tools continue to be developed to assist nurses with clinical decision making. Some of the current tools are concept mapping (Ferrario, 2003; 2004a; 2004b), N-CODES (O’Neill, Dluhy, & Chin, 2005; O’Neill, Dluhy, Hansen, & Ryan, 2006), decision analysis (Narayan, Corcoran-Perry, Drew, Hoyman, & Lewis, 2003), and software known as CHOICE (Ruland, & Bakken, 2003). Educators and administrators should explore the possibilities of using one or more of these decision-support tools as they work to strengthen nurses’ diagnostic skills

Situational Context

In health care settings, there are many situational practices that can be modified to increase accuracy of diagnosis. Some of these practices include appropriate selection of software for EHRs, development of environments that support nurse-patient partnerships, provision of feedback to nurses about the patient outcomes based on their diagnoses, adoption of policies that support the goal of accuracy, changing of admission protocols, marketing to the public nurses’ diagnoses of human responses, and funding of resources to address the common diagnoses, interventions, and outcomes for the populations they serve. The following paragraphs will provide hints and resources to assist in utilizing these practices.

Software structures and processes need to be examined for their ability to support the nursing process. The structure, for example, should include separate screens for assessment data and the diagnostic process so that the adequacy of data support for diagnoses can be retrospectively examined. There must be sufficient space in the sections for assessment data and diagnoses so that research-based diagnoses and their data support can be added or changed in accordance with research findings. Also important is the availability of electronic libraries to provide evidence for the effectiveness of nursing interventions. The nursing concepts in NANDA-I, NIC and NOC have had extensive research support. Processes that need to be examined are the linkages of assessment data, diagnoses, interventions, and nursing outcomes.

In relation to nurse-patient partnerships, nurses can be the leaders in health care settings for implementation of partnership models of patient care. For example, health care agencies can encourage nurses to set up contracts with patients and encourage patient and family signatures as an indication of partnership (similar to informed consent) in addressing the diagnoses for which they agree. Patient and family agreement with a nurse’s diagnosis can help the patient and family understand themselves and increase their willingness to work with health care personnel to improve their health status.

Provision of feedback to nurses about the patient outcomes based on their diagnoses, especially in acute care settings, will help nurses to experience the satisfaction of making accurate diagnoses and providing the appropriate associated interventions. Use of the NOC with its five-point scales will help nurses to see positive outcomes that occur before patients leave acute care units (Moorhead et al., 2004). It would also be helpful to create organizational feedback systems so nurses could evaluate the appropriateness of their diagnoses and interventions based on the patient’s long term outcomes.

...health care systems should be teaching and supporting nurses to rule in and rule out specific competing diagnoses... Adoptions of policies that support the goal of diagnostic accuracy are important because agency policies provide organizational frameworks for nursing practice. If policies and procedures suggest that nurses should focus on tasks, such as distributing medications, with little support for thinking about patient data for purposes of accurate diagnoses, one can expect the rates of accuracy will be lower. The evidence presented in this article indicates that health care systems should be teaching and supporting nurses to rule in and rule out specific competing diagnoses instead of collecting patient data just to complete assessment forms.

Admission protocols should be changed if they suggest that nurses must document a specific number of nursing diagnoses at the end of the initial assessment. First, accurate diagnoses are made when sufficient data is accumulated to support the diagnoses, not necessarily at the end of an admission assessment. It may not be possible to identify the most appropriate problems to address upon admission. Second, initially there may be no nursing care problems or nursing diagnoses in some patients. This is illustrated by a clinical study of nurses’ accuracy in diagnosing the psychosocial problems of 160 patients in three hospitals in which a staff nurse and two clinical nurse specialist assessed the patients and found that 28 (17.5%) of the patients had no psychosocial problems that needed to be addressed by nurses (Lunney, et al., 1997).

Nurse administrators should also consider marketing the use of nurses’ diagnoses of human responses to the public. This will help the people who use the health care agency know what to expect from the nurses and work more effectively with the nurses to identify the responses to health problems and life processes for which they desire nursing assistance. It will also reinforce the behaviors of nurses who focus more diligently on making accurate diagnoses of patients’ responses. 

Finally, adequate agency funding of resources is necessary for nurses to accurately diagnose and treat to their patients.  Needed resources include reference materials related to human response diagnoses, such as the latest books by NANDA-I (2007), research reports to guide evidenced-based practice (Levin, et al., 2005), opportunities for continuing education, the time and support to allow nurses to think critically and use clinical reasoning skills, and the availability of other providers to help identify the most appropriate diagnoses. Continuing education may be needed to bring nurses up-to-speed and keep them up-to-date in these areas. Time for nurses to think and to collaborate with patients, families, and other health providers is an important resource for improving diagnostic accuracy among nurses.

Summary

This appeal for nurses in practice and education to address the accuracy of nursing diagnoses is clearly supported by the research evidence, the complexity of diagnosis in nursing, and the impending implementation of EHRs. The accountability of nurses to address accuracy has been weak in the past but may be improving. Strategies to further demonstrate accountability by improving accuracy have been provided in this article.

Author

Margaret Lunney, RN, PhD
E-mail: Margaret.lunney@gmail.com

Margaret Lunney is a Professor and Graduate Nursing Programs Coordinator at the College of Staten Island, the City University of the New York (CUNY), and Doctoral Faculty at CUNY’s Graduate Center. For 25 years, her research and professional activities have focused on community health nursing, critical thinking, and the concept of accuracy of nurses’ diagnoses, measurement of accuracy, and use of standardized nursing languages. Among the honors she received are the 2001 Distinguished Nurse Researcher Award from the New York State Nurses’ Foundation and a 2007 Fulbright Award as research consultant and lecturer in Japan. She received her M.S. in nursing from Hunter College, CUNY, and her PhD in Nursing Science from New York University.


© 2008 OJIN: The Online Journal of Issues in Nursing
Article published January 31, 2008

References

Amos, E., & White, M. J. (1998). Teaching tools: Problem-based learning. Nurse Educator, 23(2), 11-14.

Beers, G. W. (2005). The effect of teaching method on objective test scores: Problem-based learning versus lecture. Journal of Nursing Education, 44, 305-309.

Benner, P. A. (1984). From novice to expert: Excellence and power in clinical nursing practice. Menlo Park, CA: Addison Wesley.

Bostick, J. E., Rantz, M. J., Flesner, M. K., & Riggs, C. J. (2006). Systematic review of studies of staffing and quality in nursing homes.  Journal of American Medical Directors Association, 7, 366-376.

Brannon, L. A., & Carson, K. L. (2003). The representative heuristic: Influence on nurses’ decision making. Applied Nursing Research, 16, 201-204.

Browne, M. N., & Keeley, S. M. (2007). Asking the right questions: A guide to critical thinking. (8th ed.). Upper Saddle River, NJ: Pearson Prentice Hall. 

Bucknall, T. (2003). The clinical landscape of critical care: nurses’ decision-making. Journal of Advanced Nursing, 43, 310-319.

Carlson, J. (2006). Abstract: Consensus validation process: A standardized research method to identify and link the relevant NANDA, NIC and NOC terms for local populations. International Journal of Nursing Terminologies & Classification, 17, 23-24.

Carnevali, D. L., & Thomas, M. J. (1993). Diagnostic reasoning and treatment decision making. Philadelphia: Lippincott.

Dochterman, J. M., Bulechek, G. M. (2004). Iowa Intervention Project: Nursing Interventions Classification [NIC] (4th ed.). St. Louis: Mosby.

Ebright, P., Patterson, E., Chalko, B., & Render, M. (2003). Understanding the complexity of registered nurse work in acute care settings. Journal of Nursing Administration, 33, 630-638.

Ferrario, C.G. (2003). Experienced and less experienced nurses’ diagnostic reasoning: Implications for fostering students’ critical thinking. International Journal of Nursing Terminologies & Classification, 14, 41-52.

Ferrario, C. G. (2004a). Developing nurses’ critical thinking skills with concept mapping. Journal of Nurses in Staff Development, 20, 261-267.

Ferrario, C. G. (2004b). Developing clinical reasoning strategies: Cognitive shortcuts. Journal of Nurses in Staff Development, 20, 229-235.

Gambrill, E. (2005). Critical thinking in clinical practice: Improving the quality of judgments and decisions (2nd ed.). Hoboken, NJ: John Wiley & Sons.

Gordon, M. (1994). Nursing diagnosis: Process and application (3rd ed). St. Louis: Mosby.

Hedberg, B., & Larsson, U. S. (2004). Environmental elements affecting the decision making process in nursing practice. Journal of Clinical Nursing, 13, 316-324.

Hicks, F. D. Merritt, S. L., & Elstein, A. S. (2003). Critical thinking and clinical decision making in critical care nursing: A pilot study. Heart & Lung, 32, 169-180.

Institute of Medicine. (2004). Keeping patients safe. Washington DC: National Academy Press.

Junnola, T., Eriksson, E., Salantera, S., & Lauri, S. (2002). Nurses’ decision-making in collecting information for the assessment of patients’ nursing problems. Journal of Clinical Nursing, 11, 186-196.

Leenerts, M. H., & Teel, C. S. (2006). Relational conversation method for creating partnerships: Pilot study. Journal of Advanced Nursing, 54, 467-476.

Levin, R., Lunney, M., & Krainovich-Miller, B. (2005). Improving diagnostic accuracy using an evidenced-based nursing model. International Journal of Nursing Terminologies & Classification, 15, 114-122.

Lunney, M. (1990). Accuracy of nursing diagnoses: Concept development. Nursing Diagnosis, 11(1), 12 -17.

Lunney, M. (1998). Where are we now? Accuracy of nurses’ diagnoses: Foundations of NANDA, NIC and NOC. Nursing Diagnosis: The Journal of Nursing Language and Classification, 9, 83-85.

Lunney M. (2001). Critical thinking and nursing diagnosis: Case studies and analyses. Philadelphia: NANDA International.

Lunney, M., Karlik, B., Kiss, M., & Murphy, P. (1997). Accuracy of nurses’ diagnoses of psychosocial responses. Nursing Diagnosis, 8, 157-166.

Lunney, M., Parker, L., Fiore, L., Cavendish, R., & Pulcini, J. (2004). Feasibility of studying the effects of using NANDA, NIC and NOC on children’s health outcome. CIN: Computers, Informatics, Nursing, 22, 316-325.

McCabe, C. (2004). Nurse-patient communication: An exploration of patients’ experiences. Journal of Clinical Nursing, 13, 41-49.

McCaffrey, M., & Ferrell, B. R. (1997). Nurses’ knowledge of pain assessment and management: How much progress have we made? Journal of Pain Symptom Management, 14, 175-188.

Moorhead, S., Johnson, M., & Maas, M. (2004). Iowa Outcomes Project: Nursing Outcomes Classification [NOC] (3rd ed.). St. Louis: Mosby.

Müller-Staub, M., Needham, I., Odenbreit, M., Lavin, M. A., & van Achterberg, T. (2007). Improved quality of nursing documentation: Results of a nursing diagnoses, interventions and outcomes study. International Journal of Nursing Terminologies and Classifications,18,  5-17.

NANDA International. (2007). Nursing diagnoses: Definitions and classification, 2007-2008. Philadelphia: Author.

Narayan, S. M., Corcoran-Perry, S., Drew, D., Hoyman, K., & Lewis, M. (2003). Decision analysis as a tool to support an analytical pattern of reasoning. Nursing & Health Sciences, 5, 229-243.

Olsson, S., Lymberts, A., & Whitehouse, D. (2004). European Commission activities in ehealth. International Journal of Circumpolar Health, 63, 310-316.

O’Neill, E. S., Dluhy, N. M., & Chin, E. (2005). Modeling novice clinical reasoning for a computerized decision support system. Journal of Advanced Nursing, 49, 68-77.

O’Neill, E. S., Dulhy, N. M., Hansen, A. S., & Ryan, J. R. (2006). Coupling the N-CODES system with actual nurse decision-making. CIN: Computers, Informatics, Nursing, 24, 28-34.

Paquay, L., Wouters, R., Debaillie, R., & Geys, L. (2007, April). Validity of nursing home diagnoses in Flemish home care nursing. Paper presented at the ACENDIO Conference, Amsterdam, Netherlands.

Phillips, N., & Duke, M. (2001).The questioning skills of clinical teachers and preceptors: A comparative study. Journal of Advanced Nursing, 33, 523-529.

Potter, P., Boxerman, S., Wolf, L., Marshall, J., Grayson, D., Sledge, J., & Evanoff, B. (2004). Mapping the nursing process: A new approach for understanding the work of nursing. Journal of Nursing Administration, 34, 101-109.

Potter, P., Wolf, L., Boxerman, S., Grayson, D., Sledge, J., Dunagan, C., & Evanoff, B. (2005). Understanding the cognitive work of nursing in the acute care environment. Journal of Nursing Administration, 35, 327-335.

Puntillo, K., Neighbor, M., O’Neill, N., & Nixon, R. (2003). Accuracy of emergency nurses in assessment of patients’ pain. Pain Management in Nursing, 4, 71-75.

Redden, M., & Wotton, K. (2001). Clinical decision making by nurses when faced with third-space fluid shift: How well do they fare? Gastrointestinal Nursing, 24, 182-191.

Reischman, R., & Yarandi, H. N. (2002). Critical care cardiovascular nurse expert and novice diagnostic cue utilization. Journal of Advanced Nursing, 39, 24-34.

Rubenfeld, M. G., & Scheffer, B. K. (2006). Critical thinking TACTICS for nurses. Boston: Jones & Bartlett.

Ruland, C. M., & Bakken, S. (2003). Developing, implementing, and evaluating decision support systems for shared decision making in patient care: A conceptual model and case illustration. Journal of Biomedical Informatics, 35, 313-321.

Sackett, D.L., Strauss, S.E., Richardson, W.S., Rosenberg, W., & Haynes, R.B. (2000). Evidenced-based medicine: How to teach and practice EBM. (2nd ed.) Edinburgh, UK: Churchill Livingston.

Sandstrom, S. (2006). Use of case studies to teach diabetes and other chronic illnesses to nursing students. Journal of Nursing Education, 45, 229-232.

Tanner, C.A. (2006). Thinking like a nurse: A research-based model of clinical judgment in nursing. Journal of Nursing Education, 45, 204-211.

Thompson, C. (2003). Clinical experience as evidence in evidence-based practice. Journal of Advanced Nursing, 43, 230-237.

U.S. Department of Health & Human Services. (2004). HHS fact sheet--HIT report at a glance.Retrieved July 21, 2007 from www.hhs.gov/news/press/2004pres/20040721.html

Webster G. (1984). Nomenclature and classification systems development. In M.J. Kim, G.K. McFarland, & A.M. McLean (Eds.), Classification of Nursing Diagnoses: Proceedings of the Fifth National Conference (pp. 14-25). St. Louis: Mosby.

Welton, J., & Halloran, E. (2005). Nursing diagnosis, diagnosis-related groups, and hospital outcomes. Journal of Nursing Administration, 35, 541-549.

Wilkinson, J.M. (2001). Nursing process and critical thinking (3rd ed.). Upper Saddle River: Prentice Hall Health.

Williams, B. (2001). The theoretical link between problem-based learning and self directed learning for continuing professional education. Teaching in Higher Education, 6(1), 85-98.

Willingham, D. T. (2007). Critical thinking: Why is it so hard to teach? American Educator, 31(2), 8-19.

ACKNOWLEDGEMENT

This article is based on a paper presented at the third conference of NANDA-International, NIC and NOC in Philadelphia, March 15-17, 2006. The author thanks Dr. Judy Carlson, Nurse Researcher at Tripler Army Medical Center in Honolulu, Hawaii, for her thoughtful critique of this paper and ongoing collaboration toward knowledge development of accuracy of nurses’ diagnoses.

Citation: Lunney, M., (Jan. 31, 2008)  "Critical Need to Address Accuracy of Nurses’ Diagnoses" OJIN: The Online Journal of Issues in Nursing. Vol. 13 No. 1.