- Email: [email protected]
Expert critical care nurses’ use of pulmonary artery pressure monitoring
Original article
Expert critical care nurses’ use of pulmonary artery pressure monitoring Leanne M. Aitken
Critical care nurses make numerous complex decisions during their day-to-day practice. General themes in previous decision-making studies have included the influence of knowledge and previous experience, the increasing complexity of decisions made and the change in decisionmaking processes used as the nurse progresses from a novice to an expert practitioner. This paper reports one component of a study which used a concept attainment framework to determine what data were used by eight expert critical care nurses in relation to haemodynamic monitoring. Results indicated that pulmonary artery pressure monitoring was used to attain the concepts of preload, cardiac output and blood pressure. In addition, participants used few clinical assessment attributes, but collected a large number of attributes which they arranged around three to five central concepts and took a broad view of haemodynamic assessment. One participant did not display many of the decision-making features normally associated with an expert practitioner. In conclusion, expert critical care nurses process an immense amount of data in a short space of time. However, they may not use all available data. Evidence suggests not all nurses who practise in the field for a lengthy period reach the level of an expert. © Harcourt Publishers Ltd Leanne M Aitken, I.C. Cert, B HSc(Nurs)Hons, PhD, FRCNA, Director, Staff Development Unit, The Canberra Hospital Adjunct Associate Professor of Clinical Nursing, University of Canberra, Canberra, ACT, Australia Correspondence to: Leanne M. Aitken, PO Box 11, Woden ACT 2606 Australia. Tel: +61 2 6244 2437; Fax: +61 2 6244 3195; Email: leanne.aitken@act. gov.au (Requests for offprints to LMA) Manuscript accepted 9/3/00
Introduction Haemodynamic monitoring is an accepted and essential component in the assessment of a critically ill patient. It is necessary for accurate determination of underlying cardiac function, as well as measurement of response to therapeutic intervention. Interpretation and integration of this information requires the clinician to make many decisions in a rapid and effective manner. An understanding of how expert critical care nurses make these decisions offers the potential to both refine these practices and to teach the principles to novice critical care practitioners.
Background Previous research into decision-making by critical care nurses has not been extensive. There have
© 2000 Harcourt Publishers Ltd
been several descriptions of factors which influence critical care nursing decisions (Thompson & Sutton 1985; Baumann & Bourbonnais 1982, 1983). Knowledge and experience were identified as the most important factors influencing decision-making in this setting. These decisions were seen to be increasing in complexity with the gradual assumption of many responsibilities that previously belonged to physician. This theme of the expanding professional role has also been identified by Watson (1994) in his examination of nurses’ decision-making. Findings included the revelation that the majority of decisions were made based on experience while only a minority of subjects used research findings to make their decisions. One study (Sims & Fought 1989) found that critical care nurses made more correct than incorrect decisions. However, evidence suggests
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
209
Intensive and Critical Care Nursing
they frequently give incorrect reasons for their actions. This problem of incorrect reasons was not found to improve with experience. The series of work by Benner and colleagues (Benner et al. 1996, 1992; Benner 1984) applied and extended the Dreyfus Model of Skill Acquisition to depict the transition in nurses’ problem-solving processes as they move from practising as novices through to practising as experts. These authors describe, first, a change from reliance on abstract principles and rules to use of past, concrete experience; second, a shift from analytic rule-based thinking to intuition; third, a move towards viewing a situation as a complex whole with specific relevant parts rather than a compilation of equally relevant bits and finally, a passage from detached observer to involved practitioner. Research examining decision-making principles within critical care areas has also examined individual aspects of the nurses’ role in patient management. Aspects examined include pain assessment and management (Mardegan 1997; Guyton-Simmons & Ehrmin 1994; Ferrel et al. 1991; Guyton-Simmons & Mattoon 1991); blood pressure monitoring (Bruya & Demand 1985) and management of patients with tachydysrhythmias (Henry 1991a, b). These studies begin to describe how critical care nurses make clinical decisions during everyday practice. The results provide some general themes such as the influence of knowledge and previous experience, the increasing complexity of decisions made and the change in decision-making processes used as the nurse progresses from a novice to an expert practitioner. Of great concern is the trend that, although the nurses being studied were able to identify the appropriate decisions, they had difficulty providing the correct theoretical rationale for their decisions (Sims & Fought 1989; Baumann & Bourbonnais 1982).
Decision-making processes Decision-making processes have been investigated from two major perspectives, namely rationalist and interpretive. Rationalist studies make the broad assumption that there is some analytical or problem-solving activity going on in the mind of the decision-maker, and that activity can be studied and made explicit.
210
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
In contrast, the interpretive perspective has been largely based on Heideggerian phenomenology principles and frequently uses retrospective techniques. The aim is to describe processes and outcomes of a decision-making process and make explicit an understanding of practice (Tanner 1993).
Methodology This paper reports one component of a study which used a concept-attainment framework to determine what data were collected, how participants collected, processed and validated those data, as well as the strategies used to attain the desired outcome of the decision-making process in relation to haemodynamic monitoring. Specifically, this paper presents an overview of the attributes and concepts collected as well as the general principles of decision-making used by participants. Concept-attainment theory is one of the descriptive rationalist approaches available for studying decision-making processes. It describes the strategies which decision-makers use to acquire, retain and utilize information (known as attributes) in order to attain a hypothesis (known as a concept) and has its origins in work by Bruner et al. (1956). Attributes consist of any discriminably different features of an event or a condition which vary from time to time (Bruner et al. 1956). Within the clinical healthcare setting, attributes include the signs and symptoms of a patient, such as breath sounds and past history. A concept is a network of inferences, thoughts or ideas that is formed in the mind as a result of categorization of attributes (Bruner et al. 1956). Within healthcare, a concept normally takes the form of a diagnosis which is inferred on the basis of the presence of a range of signs and symptoms, or attributes. The final element in concept attainment theory – a strategy – is a series of processes or mental operations in the acquisition, retention and utilization of information (Bruner et al. 1956). A strategy does not refer to a conscious plan for achieving and utilizing information.
Design Registered nurses who met the criteria for an expert critical care nurse (Fig. 1) were invited to
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
1. Registered Nurse; 2. Critical care certificate / diploma; 3. More than five years’ critical care experience; 4. More than six months’ experience in current critical care unit; 5. Currently working at least two days per week; 6. Involved in the care of patients with pulmonary artery pressure monitoring at least once a week; 7. Consider themselves to be an expert in the area of haemodynamic monitoring.
Fig. 1
Criteria for expert critical care nurses.
participate in this study during education sessions conducted within five critical care units within two metropolitan teaching hospitals. Eight nurses volunteered to participate in the study. The participants in the study were asked to ‘think aloud’ during a two-hour period of care of a critically ill patient (Aitken & Mardegan, in press). This ‘thinking aloud’ occurred in the real rather than simulated setting. The choice of this setting was based on the over-riding concern that the simulated setting was unable to replicate truly the complexity and unpredictability of the real critical care setting (Fonteyn & Fisher 1995). The researcher believed the data obtained from a study conducted in the real setting would more accurately reflect the genuine way that expert critical care nurses make clinical decisions regarding the haemodynamic assessment and management of their patients.
Data collection Once a participant had consented to involvement in the study, a time for data-collection was arranged. Data-collection took place during either the day or the night and took into account factors such as potential workload within the critical care unit. Participants then received further instruction and practice in the method of ‘thinking aloud’. On the day of data-collection, the nurse co-ordinating the shift was made aware of the intended data-collection, no other staff were specifically advised of the data-collection, although they were potentially all exposed to the
© 2000 Harcourt Publishers Ltd
education sessions, and specific information was not refused if they posed a question. This process was used to ensure the practice setting was as near to normal as possible. An audio tape-recorder with a lapel microphone was attached to the participant for a two-hour period of care of a critically ill patient with a pulmonary artery catheter in situ. The period of taping usually commenced at approximately half past an hour to ensure two sets of routine observations were performed in addition to non-routine observations and management during the data-collection phase. All care continued as it usually would, including interaction with the patients, their family and other members of the health care team. During the ‘thinking aloud’, the participant was asked to describe how s/he was assessing and managing the patient. This included a description of what data were being collected and how the nurse was using this information to formulate hypotheses or diagnoses regarding the patient. It was emphasized that the participants did not need to provide an explanation or rationale for their thoughts or actions, that was the purpose of the follow-up interview and hence any ‘thinking aloud’ that was disjointed or seemed illogical could be explained at this time. Participants were advised of the interest in haemodynamic management, but no emphasis was placed on pulmonary artery pressure measurements. An emphasis was, however, placed on providing care that was as similar to their usual practice as possible.
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
211
Intensive and Critical Care Nursing
An interview was conducted with each participant following transcription of the data obtained during the period of ‘thinking aloud’. This semi-structured interview focused on the processes verbalized in the ‘thinking aloud’. The researcher asked a series of questions aimed at clarification and validation of the data. The participant was encouraged to provide additional clarifying information where they thought appropriate. This interview was audiotaperecorded.
Data analysis Data analysis consisted of initial development of concept maps for each participant based on the attributes and concepts they identified relating to haemodynamic assessment and management. Later phases of analysis used only the data which were directly related to pulmonary artery pressure monitoring, and involved identification of a range of features including hypothesis generation and deactivation, validation of the accuracy of both attributes and hypotheses and the decision-making strategies used. The first phase of this analysis, the identification of the attributes and concepts related to haemodynamic assessment and management, is reported in this paper. This analysis consisted of development of a concept map of the decision-making process used by each participant (Novak & Gowin 1984). The concept maps were developed based on the attributes and concepts identified from both the ‘thinking aloud’ period of data collection and the follow-up interview. As the transcripts were analysed each newly identified attribute or concept was added to the concept map. Generally, each new attribute or concept was related to information already within the map and so the relationship could be identified immediately through the use of appropriate lines and arrows. Where relationships were not readily apparent at the time of identification the attribute or concept was added without relational links; these links were added at a later stage as the relationship was clarified. Concept maps were used to provide an effective means of analysing and presenting a large volume of data collected over a relatively lengthy time frame. At the same time as the concept maps were developed, a list of all the
212
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
haemodynamic concepts and associated attributes used by each subject was composed. This allowed identification of the number of attributes each subject used in the attainment of any given concept. It also allowed identification of the types of attributes, for example invasive haemodynamic monitoring or clinical assessment, that were used.
Internal reliability Internal reliability of the data analysis and concept mapping was maximized by involving an expert critical care nurse academic in the analysis of data from the first two participants. Both the researcher and the nurse academic observed the participant during the ‘thinking aloud’ session, reviewed the transcript from this phase of data collection and determined any questions to be asked during the follow-up interview. The follow-up interview was performed by the researcher. Analysis was performed independently. Independent analysis of data from the first and second participant provided 83% and 91% agreement respectively in the concepts identified and 82% and 87% agreement respectively in the attributes identified. On both occasions, 100% agreement was achieved with discussion.
Ethical considerations Permission to conduct this study was successfully sought from the Research and Ethics Committees of a university and two teaching hospitals. Informed consent was obtained from each of the nurse participants. While patients did not fill the role of a subject in the study, they did require significant consideration. An explanation of the process, including a rationale for the nurse’s constant talking, was provided. This explanation was given regardless of the patient’s conscious state and repeated at regular intervals. If the patient indicated the constant talking was annoying them, or if the nurse thought they were becoming agitated by the constant talking, provision was made to terminate data-collection. This option was never used, possibly because the talking was not unlike that involved in clinical teaching which is a regular feature of bedside activity. The decision for starting and stopping the ‘thinking
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
aloud’ remained with the nurse caring for the patient. Areas of ‘thinking aloud’ which the nurse felt the patient should not hear, or had the potential to be distressing to the patient, was described away from the bedside, usually at the site where documentation and health care team discussions took place. If gaps occurred in the data as a result of this practice, the follow-up interview provided a forum for clarification. If family members were visiting the patient at any time during data collection they were given the same explanation and considerations as the patient.
Results Eight expert critical care nurses participated in this study. The majority of participants were female (87.5%) and had an average of 6.2 years of critical care experience (range 5–9 years), with that experience being gained in between one and four critical care units. All participants had undertaken hospital-based post-registration critical care programmes between three and eight years ago and had been practising in their current unit for an average of 4.7 years (range 3.25–9 years). Seven participants worked four or five
Fig. 2
days per week with the remaining participant working two days a week. All participants cared for patients with pulmonary artery catheters in situ on average twice a week or more. Identification of the attributes and concepts used by each participant in the area of haemodynamic assessment and management enabled the development of individualized concept maps (for an example, see Fig. 2). These concept maps were remarkably similar in their general pattern for seven of the eight participants, with only one participant (participant ‘h’) appearing to make decisions using notably different strategies (Fig. 3). Analysis of the transcripts and concepts maps for attributes and concepts pertaining specifically to pulmonary artery pressure monitoring revealed some interesting trends. The pulmonary capillary wedge pressure (PCWP) was used as an attribute by all participants in the study, while the pulmonary artery diastolic (PAD) pressure was used by only five of the eight participants. The pulmonary artery systolic (PAS) and pulmonary artery mean (PAM) pressures were not used by any of the participants in this study to assess and manage the haemodynamic status of their patients.
Concept map (Participant ‘f’).
© 2000 Harcourt Publishers Ltd
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
213
Intensive and Critical Care Nursing
Fig. 3
Concept map (Participant ‘h’).
Similarities existed between each of the participants when determining which concepts they attained with the use of PCWP and PAD pressure. These concepts included preload, cardiac output and blood pressure. In no instance were these three concepts attained with exclusive reliance on pulmonary artery pressure monitoring.
Preload The concept of preload dealt with the issues related to the fluid status of the patient and was referred to by various terms including ‘Filling State/Status’ and ‘Fluid Load/Status’ in addition to the generic ‘Preload’. All participants placed a strong emphasis on the PCWP as an attribute indicating the adequacy of the concept. A range of other attributes were used by participants to attain the concept of preload (Tables 1 & 4).
Cardiac output Cardiac output was used to describe the general concept of blood flow or perfusion throughout the body and was identified by all eight
214
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
participants. However, one participant described cardiac output as being two steps removed from the attribute of PCWP and consequently has not been included in this analysis (Tables 2 & 4). For the remaining seven participants, the concept of cardiac output was always one step removed from the attributes of PAD and PCWP, with the input usually via preload. One of the participants referred to this concept as ventricular muscle function, however the content of the explanations provided during the follow-up interview suggested that the term was being used to describe a similar concept to that generally described as cardiac output.
Blood pressure The final concept attained by the majority of participants with the use of pulmonary artery pressure monitoring was blood pressure, which was perceived as an important influence on the perfusion of blood throughout the patient, particularly in relation to individual organs such as the kidneys. Despite all participants identifying the concept of blood pressure, one participant did not describe a link, even remotely,
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
Table 1 Attributes used by participants to attain the concept of preload
Table 2 Attributes used by participants to attain the concept of cardiac output
Attribute
Attribute
Number of participants who used attribute
PCWP CVP PAD Type of fluid given Urine output Heart function Amount of fluid given Glycerol trinitrate Bleeding Renal function Temperature Fluid balance Past history Haemofiltration Frusemide Constriction/dilation Peripheral oedema Mucous membranes Haemoglobin Chest X-ray Respiratory status Serum sodium Heart rate Cardiac rhythm Left ventricular pressure Size of respiratory swing on PA trace
8 8 5 5 5 4 3 3 3 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Number of participants who used attribute
Preload Cardiac rhythm Cardiac rate Afterload Stroke volume Blood pressure Ventricular function Myocardial ischaemia Myocardial work Respiratory status Hypoxia Atrio-ventricular synchrony Temperature LVSWI Ventilation Negative inotropic effect Dopamine Urine output Pre-operative condition Mitral incompetence Lactate/pH Myocardial depressant factor Pedal pulses
7 6 6 4 3 3 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1
Abbreviation: LVSWI, Left ventricular stroke work index.
Abbreviations: PCWP, pulmonary capillary wedge pressure; CVP, central venons pressure; PAD, pulmonary artery diastolic.
between the attributes of PCWP or PAD and the concept of blood pressure, while two other participants described the concept as being at least two steps removed from the attributes of PCWP or PAD. The remaining five participants used the attributes of PCWP or PAD to attain the concept of blood pressure, although these attributes were always one step removed from blood pressure via the concept of preload. These attributes were combined with a range of other attributes to attain the concept (Tables 3 & 4).
Accuracy of pulmonary artery pressure monitoring In addition, all participants outlined a concept related to the accuracy of pulmonary artery pressure monitoring in which they described the various attributes which influenced the accuracy and therefore perceived usefulness of the pressure measurements obtained. Two
© 2000 Harcourt Publishers Ltd
participants used the attributes of PCWP and PAD pressure to attain additional concepts including tamponade, dilation/constriction, urine output and respiratory function.
Clinical assessment attributes Identification of the above attributes also included recognition of whether expert critical care nurses used clinical assessment attributes to support or validate pressure monitoring concepts or whether their reliance for assessment of these concepts was primarily on ‘numbers’. Analysis indicated that pressure measurement values were relied on almost exclusively and clinical assessment was a very small feature of the haemodynamic assessment of a critically ill patient. Participants used between zero and two clinical assessment attributes to attain the concepts of preload, cardiac output and blood pressure (Table 5). These included attributes such as peripheral oedema, condition of the mucous membranes, pedal pulses, conscious state and urine output.
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
215
Intensive and Critical Care Nursing
Table 3 Attributes used by participants to attain the concept of blood pressure
Table 4 Number of attributes used by participants to attain concepts
Attribute
Concept
Number of participants who used attribute
Preload Glycerol trinitrate Dilation/constriction Cardiac output Sodium nitroprusside Conscious level Pain Ventricular muscle function Afterload Dopamine Pancuronium Midazolam Morphine Cardiac rhythm Atrioventricular synchrony Heart rate Peripheral pulses Past history Pre-operative blood pressure Renal function Urine output Individual needs/considerations Surgeon’s preference
5 4 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
There was a tendency to treat clinical observation and the measurements gained from pulmonary artery pressure monitoring as being only slightly, if at all, related to each other. This was demonstrated by one participant who had the following comments during the ‘thinking aloud’ data collection period: I think it’s just compatible with life, no wonder he’s not weeing very much … You know this guy’s amazing when you consider his blood pressure, if only we didn’t have that Swan in, he’d be up on that ward, … I mean he wouldn’t be peeing very well.
These comments were made while the expert critical care nurse was caring for a 100-kg man two days after aortic valve replacement and coronary artery bypass grafts. The patient had a cardiac index of 1.7 litres/min/metre2, CVP of 20 cm H2O, PCWP of 26 cm H2O, urine output had been 70 ml total over the previous five hours, blood pressure was 140/64 mmHg with a mean blood pressure of 84 mmHg and a heart rate of between 110 and 140 beats per minute in atrial fibrillation. The patient had extremely cold feet and was fluctuating in level of consciousness,
216
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
Preload Cardiac output Blood pressure
Number of attributes used Average
Range
6.88 7.57 7.40
3–11 4–12 2–11
becoming disorientated on a regular basis. This participant’s ‘thinking aloud’ gave the impression that s/he believed there was complete disparity between the clinical assessment and the ‘numbers’ rather than recognizing that they should support each other and be used as such.
Decision-making process In addition to the identification of specific attributes and concepts, the process of data analysis facilitated description of some general points relevant to most participants including: • Seven of the expert critical care nurses studied did not think in a linear fashion, but rather their decision-making revolved around three to five central concepts. Although the principles of haemodynamic monitoring did not change for these participants, the concepts which received most emphasis and the pathways or inferences between attributes and concepts did change, indicating individualized care for each of the patients. Participant ‘h’ (Fig. 3) demonstrated a significantly different pattern in the concept map with the links between various attributes and concepts not forming any regular patterns. • Throughout the ‘thinking aloud’, the expert critical care nurses moved from one concept to another and back again with great frequency and apparent ease. • Participants took a broad view of haemodynamic assessment and described management as a balancing act between a variety of different attributes and concepts. • The timing of haemodynamic assessment and management was determined not only by the needs of the patient, but also by nonpatient-related or outside factors such as research, medical rounds, meal breaks and the end of a shift.
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
Table 5 Clinical observation attributes used by participants to attain concepts Concept
Number of clinical observation attributes used Average Range
Preload Cardiac output Blood pressure
1.38 0.29 1.20
0–2 0–1 0–2
• There were some assessment features which were noted during the ‘thinking aloud’ that did not seem to be incorporated in the hypotheses or rationales provided, for example left ventricular stroke work index (LVSWI). • Most participants were able to describe very clearly what was concerning them about a patient’s condition or why they considered a particular hypothesis or diagnosis was relevant, with only Participant ‘h’ using the explanation of ‘gut feeling’.
Discussion Eight registered nurses, who were each perceived to be an expert in the haemodynamic assessment and management of critically ill patients, participated in this study. Seven out of the eight participants involved in the study demonstrated decision-making practices that would normally be associated with expert practice. Possibly the most surprising feature of the current study was the numerous decisions that were made by all participants throughout the two-hour period of data collection. Participants identified between 11 and 25 concepts, with between three and six of these being directly related to pulmonary artery pressure monitoring. These concepts form the focus of this paper. Pulmonary artery pressure monitoring was linked to the concepts of preload, afterload and blood pressure by the majority of participants. Of particular interest was the wide range of other attributes which most participants linked to these concepts; this suggests thorough assessment of the patients by this group of expert critical care nurses, although limitations related to clinical assessment are discussed later. The presumption of thorough assessment is consistent with that put forward by Itano (1989), who found that experts collected more data than non-experts and
© 2000 Harcourt Publishers Ltd
attributed this to their increased knowledge and experience. However, the majority of the published literature examining this issue is contrary to this presumption and will now be considered. No significant correlation has been found between thoroughness of data collection and accuracy of data interpretation (Elstein et al. 1990). Indeed, increased accuracy in the decisionmaking scenario has been found with limited amounts of data (Cianfrani 1984; Gordon 1980). There is general agreement that not only does the expert practitioner search the problem environment for data which are of most value, but also that the expert requires less data to attain an accurate decision (Elstein et al. 1978; Kleinmuntz 1968). The value of each attribute acquired, rather than the number of attributes acquired, is considered to be the most important issue (Matthews & Gaul 1979). There is also some evidence to suggest the amount of data acquired is at least somewhat dependent on the acuity of the situation, with nurses collecting fewer data in the more acute situation (Henry 1991b). These findings may represent the use of effective ‘chunking’ by the decision-maker. The type of attributes acquired by participants is also worth considering. This study suggested that participants placed a significant emphasis on objective attributes such as those obtained from invasive monitoring, laboratory investigations and therapeutic interventions. These attributes are beneficial to the decision-maker in that they have a normal or acceptable range that is determined by an external statement and therefore are helpful in attaining concepts with maximal correctness. In contrast, subjective attributes such as clinical assessment lack such a specific defining statement and, consequently, may be perceived as less valuable by the decision maker. The lack of clinical assessment attributes used by the participants in this study raises both methodological and theoretical questions. It is possible that the method of data collection used in this study, namely ‘thinking aloud’, did not accurately measure the clinical assessment which expert critical care nurses undertook. Many clinical assessment attributes, such as peripheral colour and condition of the mucous membranes, were ongoing and required no specific attention unless abnormalities were noted. Hence, it may not have been until these
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
217
Intensive and Critical Care Nursing
attributes became abnormal, such as pale peripheries and dry mucous membranes, that the participant consciously thought about and noted the attribute. When the paucity of clinical assessment attributes was identified by the researcher during analysis, four concept maps were returned to the relevant participants, who were asked to comment on whether they felt the map accurately reflected the way they believed they made decisions or if they perceived significant inaccuracies. None of these four participants identified the absence of clinical assessment as an issue. This additional finding suggests that little emphasis is placed on clinical assessment attributes by experts in the critical care environment. This issue needs addressing through the educational programs which exist within the specialty. The feature of the participants revolving their decision-making around several concepts allows these central concepts to serve as an organizing framework for the decision-making problem under consideration, thereby overcoming some of the limitations of memory capacity (Elstein et al. 1990). This framework potentially provides a base for pattern recognition, allowing the use of heuristics, ‘chunking’ or intuition. The absence of a formal linear process should not be surprising, according to Fonteyn (1991). She points out that although practitioners, particularly nurses and physicians, are taught to reason in a linear fashion (e.g. the nursing process) during their formal education and initial clinical exposure, this practice is discontinued by the expert. Experts have been described as being highly selective in their data acquisition and only when the situation appears unfamiliar do experts search for additional less highly organised or relevant data (GuytonSimmons & Ehrmin 1994; Guyton-Simmons & Matton 1991). This patterning of expert decision-making is in stark contrast to that displayed by participant ‘h’ (Fig. 3). The haphazard patterning in the decision-making strategies employed by this participant was suggestive of a practitioner operating at a level less than that of an expert (Joseph & Patel 1990; Benner 1984). This participant appeared unable to identify clearly relevant relationships between attributes and concepts, instead collecting a large number of potentially irrelevant attributes. Expert critical
218
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
care nurses will be able to assess patients adequately by selecting relevant cues which lead to appropriate decisions (Henry 1991a). This ability, recognized as a heuristic (Fisher & Fonteyn 1995; Tversky & Kahneman 1977), is essential to the practice of an expert and has been described as the ‘notion of salience’ or the ability to ignore selectively the less important aspects of a situation or place emphasis on the more important aspects (Benner et al. 1996). The decision-making practices displayed by participant ‘h’ are described as being characteristic of an experienced nurse who, despite her/his years of experience, has failed to gain the level of an expert nurse (Benner et al. 1996). This raises questions about the criteria used for selection of expert participants in this study and suggests a combination of years of experience and self-perception are not adequate in determining an expert. The feature of participants moving frequently and easily between concepts is also in keeping with the practices of an expert (Benner et al. 1992). Participants were able to deal simultaneously with multiple aspects of management, they experienced no difficulty in managing rapidly changing clinical situations and allocated appropriate time and resources to a variety of facets of care with attention to their varying relevance and priority. Participants also took a broad view of haemodynamic assessment, describing management as a balancing act between a variety of different attributes and concepts. This feature of ‘balancing’ management involved assessing not only what was currently occurring, but also anticipating the potential consequences of current treatment, optimizing the benefits and limiting the side-effects and has been described by previous authors (Stannard et al. 1996). This feature should be interpreted with caution given the lack of clinical assessment which was identified in this study. The expert critical care nurse also appeared to balance the stability of the patient against the desire to have the most current data available for critical decision-making occasions. This allowed flexibility in the timing of haemodynamic assessment and management and was achieved by considering both the needs of the patient and non-patient-related factors such as research, medical rounds and meal breaks.
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
The final feature which expert critical care nurses displayed in this study, was their ability to describe clearly what was concerning them about a patient’s condition or why they considered a particular hypothesis or diagnosis was relevant. It should be clarified that the issue of accuracy of rationales was not assessed during this study, only the provision of the rationale. The ability of the participants to provide explanations or rationales is in line with some literature (GuytonSimmons & Mattoon 1991; Thompson & Sutton 1985). This is in contrast to a series of studies in which experts used the terms of ‘gut feeling’ or ‘intuition’ when they were unable to provide a precise rationale for their actions but believed their actions were based on prior experience (Benner et al. 1996; Benner 1984). Participant ‘h’ in the current study was the only participant to use the explanation of ‘gut feeling’ rather than being able to provide a rationale for some actions. In view of the fact that participant ‘h’ displayed few of the decisionmaking characteristics of an expert practitioner, this point raises the question of whether ‘gut feeling’ is truly a characteristic of the expert practitioner or rather a term used when the practitioner is unable to articulate a rationale. It is possible that some form of pattern recognition is the initial stimulus that causes the expert practitioner to be concerned about a clinical situation, but with probing s/he can articulate specific signs and symptoms which cause the concern. A further explanation for the absence of examples of intuition in this study is that intuition is a form of reasoning which occurs below the level of conscious awareness; the current study used a data collection technique which required the decision-making processes to be at a conscious level. Hence, intuition may have been used by participants in this study, but eluded the data collection process. This issue was not explored in detail in the current study but is raised as an area which deserves further study with more participants.
Conclusion In conclusion, expert critical care nurses use pulmonary artery pressure monitoring to attain the concepts of preload, cardiac output and blood pressure. They process an immense amount of data in a short space of time and appear to be
© 2000 Harcourt Publishers Ltd
able to move easily from one subject to another and back again, developing priorities as they progress through their work. Experts take into account a wide range of factors, using multiple attributes to determine each hypothesis or diagnosis. There is evidence to suggest some types of attributes, for example those based on clinical assessment, do not receive the emphasis that is appropriate. Evidence also exists to suggest that not all nurses who practise in the field for a long period of time reach the level of an expert. This in itself is not a problem, the problem comes from the fact that these practitioners learn all the ‘right words’ from being in the environment but do not have the supporting knowledge. The potential therefore exists for other practitioners, especially those new to the critical care area, to hear the ‘right words’ and rely on them at a level greater than may be appropriate. References Aitken L M Mardegan K J 2000 Using ‘thinking aloud’ as a means of data collection in the natural setting, Western Journal of Nursing Research, in press Baumann A, Bourbonnais F 1982 Nursing decision making in critical care areas. Journal of Advanced Nursing 7: 435–446 Baumann A, Bourbonnais F 1983 Decision-making in a crisis situation. Canadian Nurse 79(5): 23–25 Benner P 1984 From Novice to Expert: Excellence and Power in Clinical Nursing Practice. Addison-Wesley, Menlo Park, California Benner P, Tanner C, Chesla C 1992 From beginner to expert: gaining a differentiated clinical world in critical care nursing. Advances in Nursing Science 14(3): 13–28 Benner P, Tanner C A, Chesla C A 1996 Expertise in Nursing Practice: Caring, Clinical Judgment, and Ethics. Springer, New York Bruner J S, Goodnow J J, Austin G A 1956 A Study of Thinking. John Wiley & Sons, New York Bruya M A, Demand J K 1985 Nursing decision making in critical care: traditional versus invasive blood pressure monitoring. Nursing Administration Quarterly 9(4): 19–31 Cianfrani K L 1984 The influence of amounts and relevance of data on identifying health problems. In: Kim M J, McFarland, G K, McLane A M, (eds) Classification of Nursing Diagnoses: Proceedings of the Fifth National Conference. CV Mosby St Louis Elstein A S, Shulman L S, Sprafka S A 1990 Medical problem solving: a ten-year retrospective. Evaluation And The Health Professions 13(1): 5–36 Elstein A S, Shulman L S, Sprafka S A 1978 Medical Problem Solving: An Analysis of Clinical Reasoning. Harvard University Press, Cambridge, Massachusetts
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
219
Intensive and Critical Care Nursing
Ferrell B R, Eberts M T, McCaffery M, Grant M 1991 Clinical decision making and pain. Cancer Nursing 14(6): 289–297 Fisher A, Fonteyn M 1995 An exploration of an innovative methodological approach for examining nurses’ heuristic use in clinical practice. Scholarly Inquiry for Nursing Practice 9(3): 263–276 Fonteyn M E 1991 Implications of clinical reasoning studies for critical care nursing. Focus on Critical Care 18(4): 322–327 Fonteyn M, Fisher A 1995 Use of think aloud method to study nurses’ reasoning and decision making in clinical practice settings. Journal of Neuroscience Nursing 27(2): 124–128 Gordon M 1980 Predictive strategies in diagnostic tasks. Nursing Research 29(1): 39–45 Guyton-Simmons J, Ehrmin J T 1994 Problem solving in pain management by expert intensive care nurses. Critical Care Nurse 14(5): 37–44 Guyton-Simmons J, Mattoon M 1991 Analysis of strategies in the management of coronary patients pain. Dimensions of Critical Care Nursing 10(1): 21–27 Henry S B 1991a Clinical decision making of critical care nurses managing computer-simulated tachydysrhythmias. Heart & Lung 20(5): 469–477 Henry S B 1991b Effect of level of patient acuity on clinical decision making of critical care nurses with varying levels of knowledge and experience. Heart & Lung 20(5): 478–485 Itano J K 1989 A comparison of the clinical judgment process in experienced registered nurses and student nurses. Journal of Nursing Education 28(3): 120–126 Joseph G M, Patel V L 1990 Domain knowledge and hypothesis generation in diagnostic reasoning. Medical Decision Making 10(1): 31–46
220
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
Kleinmuntz B 1968 The processing of clinical information by man and machine. In: Kleinmuntz B, Formal Representation of Human Judgement (ed) John Wiley & Sons, New York, pp 149–186 Mardegan K J 1997 The clinical judgement of expert nurses when managing the post-operative pain of those who have undergone coronary artery bypass graft surgery. MN Thesis, Royal Melbourne Institute of Technology, Melbourne Matthews C A, Gaul A L 1979 Nursing diagnosis from the perspective of concept attainment and critical thinking. Advances in Nursing Science 2(1): 17–26 Novak J D, Gowin C B 1984 Learning how to learn. Cambridge University Press, Cambridge Sims K A, Fought S G 1989 Clinical decision making in critical care. Critical Care Nursing Quarterly 12(3): 79–84 Stannard D, Puntillo K, Miaskowski C, Gleeson S, Kehrle K, Nye P 1996 Clinical judgment and management of postoperative pain in critical care patients. American Journal of Critical Care 5(6): 433–441 Tanner C A 1993 Rethinking clinical judgment. In: Diekelmann N L, Rather M L, Transforming RN Education: Dialogue and Debate. National League for Nursing, San Francisco Thompson D R, Sutton T W 1985 Nursing decision making in a coronary care unit. International Journal of Nursing Studies 22(3): 259–266 Tversky A, Kahneman D 1977 Judgment under uncertainty: heuristics and biases. In: Johnson-Laird P N, Watson P C (eds) Thinking: Readings in Cognitive Science. Cambridge University Press, Cambridge Watson S 1994 An exploratory study into a methodology for the examination of decision making by nurses in the clinical area. Journal of Advanced Nursing 20: 351–360
© 2000 Harcourt Publishers Ltd
Expert critical care nurses’ use of pulmonary artery pressure monitoring Leanne M. Aitken
Critical care nurses make numerous complex decisions during their day-to-day practice. General themes in previous decision-making studies have included the influence of knowledge and previous experience, the increasing complexity of decisions made and the change in decisionmaking processes used as the nurse progresses from a novice to an expert practitioner. This paper reports one component of a study which used a concept attainment framework to determine what data were used by eight expert critical care nurses in relation to haemodynamic monitoring. Results indicated that pulmonary artery pressure monitoring was used to attain the concepts of preload, cardiac output and blood pressure. In addition, participants used few clinical assessment attributes, but collected a large number of attributes which they arranged around three to five central concepts and took a broad view of haemodynamic assessment. One participant did not display many of the decision-making features normally associated with an expert practitioner. In conclusion, expert critical care nurses process an immense amount of data in a short space of time. However, they may not use all available data. Evidence suggests not all nurses who practise in the field for a lengthy period reach the level of an expert. © Harcourt Publishers Ltd Leanne M Aitken, I.C. Cert, B HSc(Nurs)Hons, PhD, FRCNA, Director, Staff Development Unit, The Canberra Hospital Adjunct Associate Professor of Clinical Nursing, University of Canberra, Canberra, ACT, Australia Correspondence to: Leanne M. Aitken, PO Box 11, Woden ACT 2606 Australia. Tel: +61 2 6244 2437; Fax: +61 2 6244 3195; Email: leanne.aitken@act. gov.au (Requests for offprints to LMA) Manuscript accepted 9/3/00
Introduction Haemodynamic monitoring is an accepted and essential component in the assessment of a critically ill patient. It is necessary for accurate determination of underlying cardiac function, as well as measurement of response to therapeutic intervention. Interpretation and integration of this information requires the clinician to make many decisions in a rapid and effective manner. An understanding of how expert critical care nurses make these decisions offers the potential to both refine these practices and to teach the principles to novice critical care practitioners.
Background Previous research into decision-making by critical care nurses has not been extensive. There have
© 2000 Harcourt Publishers Ltd
been several descriptions of factors which influence critical care nursing decisions (Thompson & Sutton 1985; Baumann & Bourbonnais 1982, 1983). Knowledge and experience were identified as the most important factors influencing decision-making in this setting. These decisions were seen to be increasing in complexity with the gradual assumption of many responsibilities that previously belonged to physician. This theme of the expanding professional role has also been identified by Watson (1994) in his examination of nurses’ decision-making. Findings included the revelation that the majority of decisions were made based on experience while only a minority of subjects used research findings to make their decisions. One study (Sims & Fought 1989) found that critical care nurses made more correct than incorrect decisions. However, evidence suggests
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
209
Intensive and Critical Care Nursing
they frequently give incorrect reasons for their actions. This problem of incorrect reasons was not found to improve with experience. The series of work by Benner and colleagues (Benner et al. 1996, 1992; Benner 1984) applied and extended the Dreyfus Model of Skill Acquisition to depict the transition in nurses’ problem-solving processes as they move from practising as novices through to practising as experts. These authors describe, first, a change from reliance on abstract principles and rules to use of past, concrete experience; second, a shift from analytic rule-based thinking to intuition; third, a move towards viewing a situation as a complex whole with specific relevant parts rather than a compilation of equally relevant bits and finally, a passage from detached observer to involved practitioner. Research examining decision-making principles within critical care areas has also examined individual aspects of the nurses’ role in patient management. Aspects examined include pain assessment and management (Mardegan 1997; Guyton-Simmons & Ehrmin 1994; Ferrel et al. 1991; Guyton-Simmons & Mattoon 1991); blood pressure monitoring (Bruya & Demand 1985) and management of patients with tachydysrhythmias (Henry 1991a, b). These studies begin to describe how critical care nurses make clinical decisions during everyday practice. The results provide some general themes such as the influence of knowledge and previous experience, the increasing complexity of decisions made and the change in decision-making processes used as the nurse progresses from a novice to an expert practitioner. Of great concern is the trend that, although the nurses being studied were able to identify the appropriate decisions, they had difficulty providing the correct theoretical rationale for their decisions (Sims & Fought 1989; Baumann & Bourbonnais 1982).
Decision-making processes Decision-making processes have been investigated from two major perspectives, namely rationalist and interpretive. Rationalist studies make the broad assumption that there is some analytical or problem-solving activity going on in the mind of the decision-maker, and that activity can be studied and made explicit.
210
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
In contrast, the interpretive perspective has been largely based on Heideggerian phenomenology principles and frequently uses retrospective techniques. The aim is to describe processes and outcomes of a decision-making process and make explicit an understanding of practice (Tanner 1993).
Methodology This paper reports one component of a study which used a concept-attainment framework to determine what data were collected, how participants collected, processed and validated those data, as well as the strategies used to attain the desired outcome of the decision-making process in relation to haemodynamic monitoring. Specifically, this paper presents an overview of the attributes and concepts collected as well as the general principles of decision-making used by participants. Concept-attainment theory is one of the descriptive rationalist approaches available for studying decision-making processes. It describes the strategies which decision-makers use to acquire, retain and utilize information (known as attributes) in order to attain a hypothesis (known as a concept) and has its origins in work by Bruner et al. (1956). Attributes consist of any discriminably different features of an event or a condition which vary from time to time (Bruner et al. 1956). Within the clinical healthcare setting, attributes include the signs and symptoms of a patient, such as breath sounds and past history. A concept is a network of inferences, thoughts or ideas that is formed in the mind as a result of categorization of attributes (Bruner et al. 1956). Within healthcare, a concept normally takes the form of a diagnosis which is inferred on the basis of the presence of a range of signs and symptoms, or attributes. The final element in concept attainment theory – a strategy – is a series of processes or mental operations in the acquisition, retention and utilization of information (Bruner et al. 1956). A strategy does not refer to a conscious plan for achieving and utilizing information.
Design Registered nurses who met the criteria for an expert critical care nurse (Fig. 1) were invited to
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
1. Registered Nurse; 2. Critical care certificate / diploma; 3. More than five years’ critical care experience; 4. More than six months’ experience in current critical care unit; 5. Currently working at least two days per week; 6. Involved in the care of patients with pulmonary artery pressure monitoring at least once a week; 7. Consider themselves to be an expert in the area of haemodynamic monitoring.
Fig. 1
Criteria for expert critical care nurses.
participate in this study during education sessions conducted within five critical care units within two metropolitan teaching hospitals. Eight nurses volunteered to participate in the study. The participants in the study were asked to ‘think aloud’ during a two-hour period of care of a critically ill patient (Aitken & Mardegan, in press). This ‘thinking aloud’ occurred in the real rather than simulated setting. The choice of this setting was based on the over-riding concern that the simulated setting was unable to replicate truly the complexity and unpredictability of the real critical care setting (Fonteyn & Fisher 1995). The researcher believed the data obtained from a study conducted in the real setting would more accurately reflect the genuine way that expert critical care nurses make clinical decisions regarding the haemodynamic assessment and management of their patients.
Data collection Once a participant had consented to involvement in the study, a time for data-collection was arranged. Data-collection took place during either the day or the night and took into account factors such as potential workload within the critical care unit. Participants then received further instruction and practice in the method of ‘thinking aloud’. On the day of data-collection, the nurse co-ordinating the shift was made aware of the intended data-collection, no other staff were specifically advised of the data-collection, although they were potentially all exposed to the
© 2000 Harcourt Publishers Ltd
education sessions, and specific information was not refused if they posed a question. This process was used to ensure the practice setting was as near to normal as possible. An audio tape-recorder with a lapel microphone was attached to the participant for a two-hour period of care of a critically ill patient with a pulmonary artery catheter in situ. The period of taping usually commenced at approximately half past an hour to ensure two sets of routine observations were performed in addition to non-routine observations and management during the data-collection phase. All care continued as it usually would, including interaction with the patients, their family and other members of the health care team. During the ‘thinking aloud’, the participant was asked to describe how s/he was assessing and managing the patient. This included a description of what data were being collected and how the nurse was using this information to formulate hypotheses or diagnoses regarding the patient. It was emphasized that the participants did not need to provide an explanation or rationale for their thoughts or actions, that was the purpose of the follow-up interview and hence any ‘thinking aloud’ that was disjointed or seemed illogical could be explained at this time. Participants were advised of the interest in haemodynamic management, but no emphasis was placed on pulmonary artery pressure measurements. An emphasis was, however, placed on providing care that was as similar to their usual practice as possible.
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
211
Intensive and Critical Care Nursing
An interview was conducted with each participant following transcription of the data obtained during the period of ‘thinking aloud’. This semi-structured interview focused on the processes verbalized in the ‘thinking aloud’. The researcher asked a series of questions aimed at clarification and validation of the data. The participant was encouraged to provide additional clarifying information where they thought appropriate. This interview was audiotaperecorded.
Data analysis Data analysis consisted of initial development of concept maps for each participant based on the attributes and concepts they identified relating to haemodynamic assessment and management. Later phases of analysis used only the data which were directly related to pulmonary artery pressure monitoring, and involved identification of a range of features including hypothesis generation and deactivation, validation of the accuracy of both attributes and hypotheses and the decision-making strategies used. The first phase of this analysis, the identification of the attributes and concepts related to haemodynamic assessment and management, is reported in this paper. This analysis consisted of development of a concept map of the decision-making process used by each participant (Novak & Gowin 1984). The concept maps were developed based on the attributes and concepts identified from both the ‘thinking aloud’ period of data collection and the follow-up interview. As the transcripts were analysed each newly identified attribute or concept was added to the concept map. Generally, each new attribute or concept was related to information already within the map and so the relationship could be identified immediately through the use of appropriate lines and arrows. Where relationships were not readily apparent at the time of identification the attribute or concept was added without relational links; these links were added at a later stage as the relationship was clarified. Concept maps were used to provide an effective means of analysing and presenting a large volume of data collected over a relatively lengthy time frame. At the same time as the concept maps were developed, a list of all the
212
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
haemodynamic concepts and associated attributes used by each subject was composed. This allowed identification of the number of attributes each subject used in the attainment of any given concept. It also allowed identification of the types of attributes, for example invasive haemodynamic monitoring or clinical assessment, that were used.
Internal reliability Internal reliability of the data analysis and concept mapping was maximized by involving an expert critical care nurse academic in the analysis of data from the first two participants. Both the researcher and the nurse academic observed the participant during the ‘thinking aloud’ session, reviewed the transcript from this phase of data collection and determined any questions to be asked during the follow-up interview. The follow-up interview was performed by the researcher. Analysis was performed independently. Independent analysis of data from the first and second participant provided 83% and 91% agreement respectively in the concepts identified and 82% and 87% agreement respectively in the attributes identified. On both occasions, 100% agreement was achieved with discussion.
Ethical considerations Permission to conduct this study was successfully sought from the Research and Ethics Committees of a university and two teaching hospitals. Informed consent was obtained from each of the nurse participants. While patients did not fill the role of a subject in the study, they did require significant consideration. An explanation of the process, including a rationale for the nurse’s constant talking, was provided. This explanation was given regardless of the patient’s conscious state and repeated at regular intervals. If the patient indicated the constant talking was annoying them, or if the nurse thought they were becoming agitated by the constant talking, provision was made to terminate data-collection. This option was never used, possibly because the talking was not unlike that involved in clinical teaching which is a regular feature of bedside activity. The decision for starting and stopping the ‘thinking
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
aloud’ remained with the nurse caring for the patient. Areas of ‘thinking aloud’ which the nurse felt the patient should not hear, or had the potential to be distressing to the patient, was described away from the bedside, usually at the site where documentation and health care team discussions took place. If gaps occurred in the data as a result of this practice, the follow-up interview provided a forum for clarification. If family members were visiting the patient at any time during data collection they were given the same explanation and considerations as the patient.
Results Eight expert critical care nurses participated in this study. The majority of participants were female (87.5%) and had an average of 6.2 years of critical care experience (range 5–9 years), with that experience being gained in between one and four critical care units. All participants had undertaken hospital-based post-registration critical care programmes between three and eight years ago and had been practising in their current unit for an average of 4.7 years (range 3.25–9 years). Seven participants worked four or five
Fig. 2
days per week with the remaining participant working two days a week. All participants cared for patients with pulmonary artery catheters in situ on average twice a week or more. Identification of the attributes and concepts used by each participant in the area of haemodynamic assessment and management enabled the development of individualized concept maps (for an example, see Fig. 2). These concept maps were remarkably similar in their general pattern for seven of the eight participants, with only one participant (participant ‘h’) appearing to make decisions using notably different strategies (Fig. 3). Analysis of the transcripts and concepts maps for attributes and concepts pertaining specifically to pulmonary artery pressure monitoring revealed some interesting trends. The pulmonary capillary wedge pressure (PCWP) was used as an attribute by all participants in the study, while the pulmonary artery diastolic (PAD) pressure was used by only five of the eight participants. The pulmonary artery systolic (PAS) and pulmonary artery mean (PAM) pressures were not used by any of the participants in this study to assess and manage the haemodynamic status of their patients.
Concept map (Participant ‘f’).
© 2000 Harcourt Publishers Ltd
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
213
Intensive and Critical Care Nursing
Fig. 3
Concept map (Participant ‘h’).
Similarities existed between each of the participants when determining which concepts they attained with the use of PCWP and PAD pressure. These concepts included preload, cardiac output and blood pressure. In no instance were these three concepts attained with exclusive reliance on pulmonary artery pressure monitoring.
Preload The concept of preload dealt with the issues related to the fluid status of the patient and was referred to by various terms including ‘Filling State/Status’ and ‘Fluid Load/Status’ in addition to the generic ‘Preload’. All participants placed a strong emphasis on the PCWP as an attribute indicating the adequacy of the concept. A range of other attributes were used by participants to attain the concept of preload (Tables 1 & 4).
Cardiac output Cardiac output was used to describe the general concept of blood flow or perfusion throughout the body and was identified by all eight
214
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
participants. However, one participant described cardiac output as being two steps removed from the attribute of PCWP and consequently has not been included in this analysis (Tables 2 & 4). For the remaining seven participants, the concept of cardiac output was always one step removed from the attributes of PAD and PCWP, with the input usually via preload. One of the participants referred to this concept as ventricular muscle function, however the content of the explanations provided during the follow-up interview suggested that the term was being used to describe a similar concept to that generally described as cardiac output.
Blood pressure The final concept attained by the majority of participants with the use of pulmonary artery pressure monitoring was blood pressure, which was perceived as an important influence on the perfusion of blood throughout the patient, particularly in relation to individual organs such as the kidneys. Despite all participants identifying the concept of blood pressure, one participant did not describe a link, even remotely,
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
Table 1 Attributes used by participants to attain the concept of preload
Table 2 Attributes used by participants to attain the concept of cardiac output
Attribute
Attribute
Number of participants who used attribute
PCWP CVP PAD Type of fluid given Urine output Heart function Amount of fluid given Glycerol trinitrate Bleeding Renal function Temperature Fluid balance Past history Haemofiltration Frusemide Constriction/dilation Peripheral oedema Mucous membranes Haemoglobin Chest X-ray Respiratory status Serum sodium Heart rate Cardiac rhythm Left ventricular pressure Size of respiratory swing on PA trace
8 8 5 5 5 4 3 3 3 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Number of participants who used attribute
Preload Cardiac rhythm Cardiac rate Afterload Stroke volume Blood pressure Ventricular function Myocardial ischaemia Myocardial work Respiratory status Hypoxia Atrio-ventricular synchrony Temperature LVSWI Ventilation Negative inotropic effect Dopamine Urine output Pre-operative condition Mitral incompetence Lactate/pH Myocardial depressant factor Pedal pulses
7 6 6 4 3 3 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1
Abbreviation: LVSWI, Left ventricular stroke work index.
Abbreviations: PCWP, pulmonary capillary wedge pressure; CVP, central venons pressure; PAD, pulmonary artery diastolic.
between the attributes of PCWP or PAD and the concept of blood pressure, while two other participants described the concept as being at least two steps removed from the attributes of PCWP or PAD. The remaining five participants used the attributes of PCWP or PAD to attain the concept of blood pressure, although these attributes were always one step removed from blood pressure via the concept of preload. These attributes were combined with a range of other attributes to attain the concept (Tables 3 & 4).
Accuracy of pulmonary artery pressure monitoring In addition, all participants outlined a concept related to the accuracy of pulmonary artery pressure monitoring in which they described the various attributes which influenced the accuracy and therefore perceived usefulness of the pressure measurements obtained. Two
© 2000 Harcourt Publishers Ltd
participants used the attributes of PCWP and PAD pressure to attain additional concepts including tamponade, dilation/constriction, urine output and respiratory function.
Clinical assessment attributes Identification of the above attributes also included recognition of whether expert critical care nurses used clinical assessment attributes to support or validate pressure monitoring concepts or whether their reliance for assessment of these concepts was primarily on ‘numbers’. Analysis indicated that pressure measurement values were relied on almost exclusively and clinical assessment was a very small feature of the haemodynamic assessment of a critically ill patient. Participants used between zero and two clinical assessment attributes to attain the concepts of preload, cardiac output and blood pressure (Table 5). These included attributes such as peripheral oedema, condition of the mucous membranes, pedal pulses, conscious state and urine output.
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
215
Intensive and Critical Care Nursing
Table 3 Attributes used by participants to attain the concept of blood pressure
Table 4 Number of attributes used by participants to attain concepts
Attribute
Concept
Number of participants who used attribute
Preload Glycerol trinitrate Dilation/constriction Cardiac output Sodium nitroprusside Conscious level Pain Ventricular muscle function Afterload Dopamine Pancuronium Midazolam Morphine Cardiac rhythm Atrioventricular synchrony Heart rate Peripheral pulses Past history Pre-operative blood pressure Renal function Urine output Individual needs/considerations Surgeon’s preference
5 4 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
There was a tendency to treat clinical observation and the measurements gained from pulmonary artery pressure monitoring as being only slightly, if at all, related to each other. This was demonstrated by one participant who had the following comments during the ‘thinking aloud’ data collection period: I think it’s just compatible with life, no wonder he’s not weeing very much … You know this guy’s amazing when you consider his blood pressure, if only we didn’t have that Swan in, he’d be up on that ward, … I mean he wouldn’t be peeing very well.
These comments were made while the expert critical care nurse was caring for a 100-kg man two days after aortic valve replacement and coronary artery bypass grafts. The patient had a cardiac index of 1.7 litres/min/metre2, CVP of 20 cm H2O, PCWP of 26 cm H2O, urine output had been 70 ml total over the previous five hours, blood pressure was 140/64 mmHg with a mean blood pressure of 84 mmHg and a heart rate of between 110 and 140 beats per minute in atrial fibrillation. The patient had extremely cold feet and was fluctuating in level of consciousness,
216
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
Preload Cardiac output Blood pressure
Number of attributes used Average
Range
6.88 7.57 7.40
3–11 4–12 2–11
becoming disorientated on a regular basis. This participant’s ‘thinking aloud’ gave the impression that s/he believed there was complete disparity between the clinical assessment and the ‘numbers’ rather than recognizing that they should support each other and be used as such.
Decision-making process In addition to the identification of specific attributes and concepts, the process of data analysis facilitated description of some general points relevant to most participants including: • Seven of the expert critical care nurses studied did not think in a linear fashion, but rather their decision-making revolved around three to five central concepts. Although the principles of haemodynamic monitoring did not change for these participants, the concepts which received most emphasis and the pathways or inferences between attributes and concepts did change, indicating individualized care for each of the patients. Participant ‘h’ (Fig. 3) demonstrated a significantly different pattern in the concept map with the links between various attributes and concepts not forming any regular patterns. • Throughout the ‘thinking aloud’, the expert critical care nurses moved from one concept to another and back again with great frequency and apparent ease. • Participants took a broad view of haemodynamic assessment and described management as a balancing act between a variety of different attributes and concepts. • The timing of haemodynamic assessment and management was determined not only by the needs of the patient, but also by nonpatient-related or outside factors such as research, medical rounds, meal breaks and the end of a shift.
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
Table 5 Clinical observation attributes used by participants to attain concepts Concept
Number of clinical observation attributes used Average Range
Preload Cardiac output Blood pressure
1.38 0.29 1.20
0–2 0–1 0–2
• There were some assessment features which were noted during the ‘thinking aloud’ that did not seem to be incorporated in the hypotheses or rationales provided, for example left ventricular stroke work index (LVSWI). • Most participants were able to describe very clearly what was concerning them about a patient’s condition or why they considered a particular hypothesis or diagnosis was relevant, with only Participant ‘h’ using the explanation of ‘gut feeling’.
Discussion Eight registered nurses, who were each perceived to be an expert in the haemodynamic assessment and management of critically ill patients, participated in this study. Seven out of the eight participants involved in the study demonstrated decision-making practices that would normally be associated with expert practice. Possibly the most surprising feature of the current study was the numerous decisions that were made by all participants throughout the two-hour period of data collection. Participants identified between 11 and 25 concepts, with between three and six of these being directly related to pulmonary artery pressure monitoring. These concepts form the focus of this paper. Pulmonary artery pressure monitoring was linked to the concepts of preload, afterload and blood pressure by the majority of participants. Of particular interest was the wide range of other attributes which most participants linked to these concepts; this suggests thorough assessment of the patients by this group of expert critical care nurses, although limitations related to clinical assessment are discussed later. The presumption of thorough assessment is consistent with that put forward by Itano (1989), who found that experts collected more data than non-experts and
© 2000 Harcourt Publishers Ltd
attributed this to their increased knowledge and experience. However, the majority of the published literature examining this issue is contrary to this presumption and will now be considered. No significant correlation has been found between thoroughness of data collection and accuracy of data interpretation (Elstein et al. 1990). Indeed, increased accuracy in the decisionmaking scenario has been found with limited amounts of data (Cianfrani 1984; Gordon 1980). There is general agreement that not only does the expert practitioner search the problem environment for data which are of most value, but also that the expert requires less data to attain an accurate decision (Elstein et al. 1978; Kleinmuntz 1968). The value of each attribute acquired, rather than the number of attributes acquired, is considered to be the most important issue (Matthews & Gaul 1979). There is also some evidence to suggest the amount of data acquired is at least somewhat dependent on the acuity of the situation, with nurses collecting fewer data in the more acute situation (Henry 1991b). These findings may represent the use of effective ‘chunking’ by the decision-maker. The type of attributes acquired by participants is also worth considering. This study suggested that participants placed a significant emphasis on objective attributes such as those obtained from invasive monitoring, laboratory investigations and therapeutic interventions. These attributes are beneficial to the decision-maker in that they have a normal or acceptable range that is determined by an external statement and therefore are helpful in attaining concepts with maximal correctness. In contrast, subjective attributes such as clinical assessment lack such a specific defining statement and, consequently, may be perceived as less valuable by the decision maker. The lack of clinical assessment attributes used by the participants in this study raises both methodological and theoretical questions. It is possible that the method of data collection used in this study, namely ‘thinking aloud’, did not accurately measure the clinical assessment which expert critical care nurses undertook. Many clinical assessment attributes, such as peripheral colour and condition of the mucous membranes, were ongoing and required no specific attention unless abnormalities were noted. Hence, it may not have been until these
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
217
Intensive and Critical Care Nursing
attributes became abnormal, such as pale peripheries and dry mucous membranes, that the participant consciously thought about and noted the attribute. When the paucity of clinical assessment attributes was identified by the researcher during analysis, four concept maps were returned to the relevant participants, who were asked to comment on whether they felt the map accurately reflected the way they believed they made decisions or if they perceived significant inaccuracies. None of these four participants identified the absence of clinical assessment as an issue. This additional finding suggests that little emphasis is placed on clinical assessment attributes by experts in the critical care environment. This issue needs addressing through the educational programs which exist within the specialty. The feature of the participants revolving their decision-making around several concepts allows these central concepts to serve as an organizing framework for the decision-making problem under consideration, thereby overcoming some of the limitations of memory capacity (Elstein et al. 1990). This framework potentially provides a base for pattern recognition, allowing the use of heuristics, ‘chunking’ or intuition. The absence of a formal linear process should not be surprising, according to Fonteyn (1991). She points out that although practitioners, particularly nurses and physicians, are taught to reason in a linear fashion (e.g. the nursing process) during their formal education and initial clinical exposure, this practice is discontinued by the expert. Experts have been described as being highly selective in their data acquisition and only when the situation appears unfamiliar do experts search for additional less highly organised or relevant data (GuytonSimmons & Ehrmin 1994; Guyton-Simmons & Matton 1991). This patterning of expert decision-making is in stark contrast to that displayed by participant ‘h’ (Fig. 3). The haphazard patterning in the decision-making strategies employed by this participant was suggestive of a practitioner operating at a level less than that of an expert (Joseph & Patel 1990; Benner 1984). This participant appeared unable to identify clearly relevant relationships between attributes and concepts, instead collecting a large number of potentially irrelevant attributes. Expert critical
218
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
care nurses will be able to assess patients adequately by selecting relevant cues which lead to appropriate decisions (Henry 1991a). This ability, recognized as a heuristic (Fisher & Fonteyn 1995; Tversky & Kahneman 1977), is essential to the practice of an expert and has been described as the ‘notion of salience’ or the ability to ignore selectively the less important aspects of a situation or place emphasis on the more important aspects (Benner et al. 1996). The decision-making practices displayed by participant ‘h’ are described as being characteristic of an experienced nurse who, despite her/his years of experience, has failed to gain the level of an expert nurse (Benner et al. 1996). This raises questions about the criteria used for selection of expert participants in this study and suggests a combination of years of experience and self-perception are not adequate in determining an expert. The feature of participants moving frequently and easily between concepts is also in keeping with the practices of an expert (Benner et al. 1992). Participants were able to deal simultaneously with multiple aspects of management, they experienced no difficulty in managing rapidly changing clinical situations and allocated appropriate time and resources to a variety of facets of care with attention to their varying relevance and priority. Participants also took a broad view of haemodynamic assessment, describing management as a balancing act between a variety of different attributes and concepts. This feature of ‘balancing’ management involved assessing not only what was currently occurring, but also anticipating the potential consequences of current treatment, optimizing the benefits and limiting the side-effects and has been described by previous authors (Stannard et al. 1996). This feature should be interpreted with caution given the lack of clinical assessment which was identified in this study. The expert critical care nurse also appeared to balance the stability of the patient against the desire to have the most current data available for critical decision-making occasions. This allowed flexibility in the timing of haemodynamic assessment and management and was achieved by considering both the needs of the patient and non-patient-related factors such as research, medical rounds and meal breaks.
© 2000 Harcourt Publishers Ltd
Use of pulmonary & artery pressure monitoring
The final feature which expert critical care nurses displayed in this study, was their ability to describe clearly what was concerning them about a patient’s condition or why they considered a particular hypothesis or diagnosis was relevant. It should be clarified that the issue of accuracy of rationales was not assessed during this study, only the provision of the rationale. The ability of the participants to provide explanations or rationales is in line with some literature (GuytonSimmons & Mattoon 1991; Thompson & Sutton 1985). This is in contrast to a series of studies in which experts used the terms of ‘gut feeling’ or ‘intuition’ when they were unable to provide a precise rationale for their actions but believed their actions were based on prior experience (Benner et al. 1996; Benner 1984). Participant ‘h’ in the current study was the only participant to use the explanation of ‘gut feeling’ rather than being able to provide a rationale for some actions. In view of the fact that participant ‘h’ displayed few of the decisionmaking characteristics of an expert practitioner, this point raises the question of whether ‘gut feeling’ is truly a characteristic of the expert practitioner or rather a term used when the practitioner is unable to articulate a rationale. It is possible that some form of pattern recognition is the initial stimulus that causes the expert practitioner to be concerned about a clinical situation, but with probing s/he can articulate specific signs and symptoms which cause the concern. A further explanation for the absence of examples of intuition in this study is that intuition is a form of reasoning which occurs below the level of conscious awareness; the current study used a data collection technique which required the decision-making processes to be at a conscious level. Hence, intuition may have been used by participants in this study, but eluded the data collection process. This issue was not explored in detail in the current study but is raised as an area which deserves further study with more participants.
Conclusion In conclusion, expert critical care nurses use pulmonary artery pressure monitoring to attain the concepts of preload, cardiac output and blood pressure. They process an immense amount of data in a short space of time and appear to be
© 2000 Harcourt Publishers Ltd
able to move easily from one subject to another and back again, developing priorities as they progress through their work. Experts take into account a wide range of factors, using multiple attributes to determine each hypothesis or diagnosis. There is evidence to suggest some types of attributes, for example those based on clinical assessment, do not receive the emphasis that is appropriate. Evidence also exists to suggest that not all nurses who practise in the field for a long period of time reach the level of an expert. This in itself is not a problem, the problem comes from the fact that these practitioners learn all the ‘right words’ from being in the environment but do not have the supporting knowledge. The potential therefore exists for other practitioners, especially those new to the critical care area, to hear the ‘right words’ and rely on them at a level greater than may be appropriate. References Aitken L M Mardegan K J 2000 Using ‘thinking aloud’ as a means of data collection in the natural setting, Western Journal of Nursing Research, in press Baumann A, Bourbonnais F 1982 Nursing decision making in critical care areas. Journal of Advanced Nursing 7: 435–446 Baumann A, Bourbonnais F 1983 Decision-making in a crisis situation. Canadian Nurse 79(5): 23–25 Benner P 1984 From Novice to Expert: Excellence and Power in Clinical Nursing Practice. Addison-Wesley, Menlo Park, California Benner P, Tanner C, Chesla C 1992 From beginner to expert: gaining a differentiated clinical world in critical care nursing. Advances in Nursing Science 14(3): 13–28 Benner P, Tanner C A, Chesla C A 1996 Expertise in Nursing Practice: Caring, Clinical Judgment, and Ethics. Springer, New York Bruner J S, Goodnow J J, Austin G A 1956 A Study of Thinking. John Wiley & Sons, New York Bruya M A, Demand J K 1985 Nursing decision making in critical care: traditional versus invasive blood pressure monitoring. Nursing Administration Quarterly 9(4): 19–31 Cianfrani K L 1984 The influence of amounts and relevance of data on identifying health problems. In: Kim M J, McFarland, G K, McLane A M, (eds) Classification of Nursing Diagnoses: Proceedings of the Fifth National Conference. CV Mosby St Louis Elstein A S, Shulman L S, Sprafka S A 1990 Medical problem solving: a ten-year retrospective. Evaluation And The Health Professions 13(1): 5–36 Elstein A S, Shulman L S, Sprafka S A 1978 Medical Problem Solving: An Analysis of Clinical Reasoning. Harvard University Press, Cambridge, Massachusetts
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
219
Intensive and Critical Care Nursing
Ferrell B R, Eberts M T, McCaffery M, Grant M 1991 Clinical decision making and pain. Cancer Nursing 14(6): 289–297 Fisher A, Fonteyn M 1995 An exploration of an innovative methodological approach for examining nurses’ heuristic use in clinical practice. Scholarly Inquiry for Nursing Practice 9(3): 263–276 Fonteyn M E 1991 Implications of clinical reasoning studies for critical care nursing. Focus on Critical Care 18(4): 322–327 Fonteyn M, Fisher A 1995 Use of think aloud method to study nurses’ reasoning and decision making in clinical practice settings. Journal of Neuroscience Nursing 27(2): 124–128 Gordon M 1980 Predictive strategies in diagnostic tasks. Nursing Research 29(1): 39–45 Guyton-Simmons J, Ehrmin J T 1994 Problem solving in pain management by expert intensive care nurses. Critical Care Nurse 14(5): 37–44 Guyton-Simmons J, Mattoon M 1991 Analysis of strategies in the management of coronary patients pain. Dimensions of Critical Care Nursing 10(1): 21–27 Henry S B 1991a Clinical decision making of critical care nurses managing computer-simulated tachydysrhythmias. Heart & Lung 20(5): 469–477 Henry S B 1991b Effect of level of patient acuity on clinical decision making of critical care nurses with varying levels of knowledge and experience. Heart & Lung 20(5): 478–485 Itano J K 1989 A comparison of the clinical judgment process in experienced registered nurses and student nurses. Journal of Nursing Education 28(3): 120–126 Joseph G M, Patel V L 1990 Domain knowledge and hypothesis generation in diagnostic reasoning. Medical Decision Making 10(1): 31–46
220
Intensive and Critical Care Nursing ( 2 0 0 0 )
16, 209–220
Kleinmuntz B 1968 The processing of clinical information by man and machine. In: Kleinmuntz B, Formal Representation of Human Judgement (ed) John Wiley & Sons, New York, pp 149–186 Mardegan K J 1997 The clinical judgement of expert nurses when managing the post-operative pain of those who have undergone coronary artery bypass graft surgery. MN Thesis, Royal Melbourne Institute of Technology, Melbourne Matthews C A, Gaul A L 1979 Nursing diagnosis from the perspective of concept attainment and critical thinking. Advances in Nursing Science 2(1): 17–26 Novak J D, Gowin C B 1984 Learning how to learn. Cambridge University Press, Cambridge Sims K A, Fought S G 1989 Clinical decision making in critical care. Critical Care Nursing Quarterly 12(3): 79–84 Stannard D, Puntillo K, Miaskowski C, Gleeson S, Kehrle K, Nye P 1996 Clinical judgment and management of postoperative pain in critical care patients. American Journal of Critical Care 5(6): 433–441 Tanner C A 1993 Rethinking clinical judgment. In: Diekelmann N L, Rather M L, Transforming RN Education: Dialogue and Debate. National League for Nursing, San Francisco Thompson D R, Sutton T W 1985 Nursing decision making in a coronary care unit. International Journal of Nursing Studies 22(3): 259–266 Tversky A, Kahneman D 1977 Judgment under uncertainty: heuristics and biases. In: Johnson-Laird P N, Watson P C (eds) Thinking: Readings in Cognitive Science. Cambridge University Press, Cambridge Watson S 1994 An exploratory study into a methodology for the examination of decision making by nurses in the clinical area. Journal of Advanced Nursing 20: 351–360
© 2000 Harcourt Publishers Ltd