RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
Factors influencing the provision of end-of-life care in critical care settings: development and testing of a survey instrument Kristen Ranse, Patsy Yates & Fiona Coyer Accepted for publication 20 October 2014
Correspondence to K. Ranse: e-mail:
[email protected] Kristen Ranse PhD RN Assistant Professor Disciplines of Nursing & Midwifery, Faculty of Health, University of Canberra, Australian Capital Territory, Australia Patsy Yates PhD RN Professor School of Nursing & Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland, Australia Fiona Coyer PhD RN Associate Professor School of Nursing & Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland, Australia
R A N S E K . , Y A T E S P . & C O Y E R F . ( 2 0 1 5 ) Factors influencing the provision of end-of-life care in critical care settings: development and testing of a survey instrument. Journal of Advanced Nursing 71(3), 697–709. doi: 10.1111/ jan.12576
Abstract Aim. To develop and psychometrically test a survey instrument to identify the factors influencing the provision of end-of-life care by critical care nurses. Background. Following a decision to withdraw life-sustaining treatment, critical care nurses remain with the patient and their family providing end-of-life care. Identification of factors influencing the provision of this care can give evidence to inform practice development and support nurses. Design. A cross-sectional survey of critical care nurses. Method. An online survey was developed, reviewed by an expert panel and pilot tested to obtain preliminary evidence of its reliability and validity. In May 2011, a convenience sample of critical care nurses (n = 392, response rate 25%) completed the survey. The analytical approach to data obtained from the 58 items measured on a Likert scale included exploratory factor analysis and descriptive statistics. Results. Exploratory factor analysis identified eight factors influencing the provision of end-of-life care: emotional support for nurses, palliative values, patient and family preferences, resources, organizational support, care planning, knowledge and preparedness. Internal consistency of each latent construct was deemed satisfactory. The results of descriptive statistics revealed a strong commitment to the inclusion of families in end-of-life care and the value of this care in the critical care setting. Conclusion. This paper reports preliminary evidence of the psychometric properties of a new survey instrument. The findings may inform practice development opportunities to support critical care nurses in the provision of endof-life care and improve the care that patients and their families receive. Keywords: critical care, end-of-life care, factors, instrument development, nursing, survey
© 2014 John Wiley & Sons Ltd
697
K. Ranse et al.
Why is this research needed? ● Previous research of critical care nurses has identified the perception that following a decision to withdraw or withhold life-sustaining treatment, critical care nurses are left to implement this decision and manage care of the patient at the end-of-life. ● Limited research has investigated the factors influencing the provision of end-of-life care by critical care nurses. ● No survey instrument was identified that could identify the key factors influencing the provision of end-of-life care by critical care nurses.
What are the key findings? ● This paper reports preliminary evidence of the psychomet-
et al. 2004, Heland 2006, Espinosa et al. 2010, Kirchhoff & Kowalkowski 2010, Ranse et al. 2012). Thus, the provision of end-of-life care in critical care settings comprises a large component of critical care nurses work. In the critical care environment, where goals of care are predominantly focused on cure, limited attention has been given to defining what is meant by end of life. In practice, however, end of life is often considered the period of life following a decision to withdraw or withhold life-sustaining treatment, as death of the patient is an expected outcome. Consistent with practice, in this study, end-of-life care is defined as the care period after a decision is made to withdraw or withhold life-sustaining treatment with death of the patient the expected outcome.
ric properties of a new survey instrument in one population. ● Eight factors influencing the provision of end-of-life care were identified: emotional support for nurses, palliative values, patient and family preferences, resources, organizational support, care planning, knowledge and preparedness. ● The findings reveal a strong commitment to the inclusion of families in end-of-life care and the value of this care in the critical care setting.
How should the findings be used to influence policy/ practice/research/education? ● The survey can be used to identify factors influencing the provision of end-of-life care in critical care units. ● The findings may inform the development of strategies to prepare and support nurses in the delivery of end-of-life care, improving the care that patients and their families receive.
Introduction Critical care units, developed in response to advances in medicine and technology, have enhanced the ability to save lives of critically ill and injured patients and delay death. Despite this curative focus, death remains inevitable and internationally, intensive care unit mortality rates of between 15–35% have been reported (Seymour 2001). The majority of patient deaths occur following a decision to withhold or withdraw life-sustaining treatment (Luce 1997, Bloomer et al. 2010, Le Conte et al. 2010). Previous studies have identified the perception that following this decision, the medical team withdraw from the care of the patient and the critical care nurse remains to enact this decision and provide care for the patient and their family (Halcomb 698
Background Previous research has provided some evidence of critical care nurses’ perceptions of barriers to and facilitators of end-of-life care (Kirchhoff & Beckstrand 2000, Beckstrand & Kirchhoff 2005, Beckstrand et al. 2006, Nelson et al. 2006). However, quantitative data analysis in these studies has been limited to descriptive statistics for individual survey items. Further research is needed to identify the key factors influencing the provision of end-of-life care for critical care nurses. Due to the limited previous research in this area and the broad range of social and personal factors that could potentially influence the provision of end-of-life care a conceptual framework was chosen to guide this study. Specifically, the Promoting Action on Research Implementation in Health Services (PARIHS) framework (Kitson et al. 1998, RycroftMalone et al. 2002) was used to capture the complexity of factors that can influence nursing practice. The PARIHS framework describes the factors influencing uptake of evidence in clinical practice (Kitson et al. 1998). The three key elements identified in the PARIHS framework: context, evidence and facilitation were considered to represent the complexity of factors that can influence nursing practice. In the PARIHS framework, context is defined as all aspects of the setting where ‘practice takes place’ (McCormack et al. 2002, p. 94) including the culture of the organization and the physical environment. Evidence is broadly defined to include all types of evidence that inform clinical practice, including research, nurses practice knowledge and patients experience and preference (Rycroft-Malone et al. 2004). In the PARIHS framework, facilitation is narrowly defined as the role of a person in making things easier for others (Rycroft-Malone et al. 2002). In this study, facilitation was more broadly defined as the process of making things © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
easier, including emotional support for nurses and the values and beliefs of individual nurses. The PARIHS framework provided a conceptual framework for this study to identify the factors influencing the provision of end-of-life care by critical care nurses. Identification of the factors influencing the provision of care provides a platform for practice development to improve the care that patients and their families receive and support nurses in the provision of this care.
Factors influencing end-of-life care in critical care
Table 1 Domains and associated a priori sub-scales in the survey instrument. Domain
A priori sub-scales
Context
Organizational culture Resources Patient and family preferences Nursing education and experience Nurse practice knowledge Emotional support Nurses’ end-of-life care values and beliefs Coordination of care delivery
Evidence
Facilitation
Methods Aim The purpose of the study was to develop and psychometrically test a survey instrument to identify the factors influencing the provision of end-of-life care by critical care nurses.
Methodology A cross-sectional survey design was selected to investigate the provision of end-of-life care by critical care nurses. The development and testing of the survey instrument was undertaken in two stages: (1) instrument development; and (2) cross-sectional survey. The survey comprised of three sections: demographic information, factors influencing the provision of end-of-life care and the end-of-life care practices of critical care nurses. The focus of this paper is on the section of the survey instrument designed to identify the factors influencing the provision of end-of-life care by nurses in critical care settings. Stage one: instrument development Item generation. Construct validity refers to the degree to which an instrument measures the latent construct that it is supposed to be measuring (Polit & Beck 2006). The development of a robust survey instrument that is reliable and valid is a time-consuming process. However, investment of time in conceptualizing the domains of interest and generating items contributes to the validity of the instrument. The three key elements of the PARIHS framework: context, evidence and facilitation were identified as the key domains to be explored with items in the survey instrument that identify the respondents’ attitudes and beliefs and the perceptions of their workplace in relation to factors that may influence the provision of care to patients at the end-of-life and their families. The a priori sub-scales in each domain (Table 1) in the survey and the individual items in each a priori sub-scale, were empirically derived from the findings of an extensive literature review and a preliminary © 2014 John Wiley & Sons Ltd
descriptive qualitative study of critical care nurses’ end-of-life care experiences (Ranse et al. 2012). Some of the items from the Neonatal Palliative Care Attitude Scale (NiPCAS; Kain et al. 2009) deemed relevant to the adult critical care context were included in the survey following changes to reflect the adult patient population and the term end-of-life care was substituted for palliative care. Each item was measured on a 5-point Likert sub-scale where: 1 = strongly disagree; 2 = disagree; 3 = neutral; 4 = agree; 5 = strongly agree. Content validity testing. The initial survey comprising 35 items was sent to six Registered Nurses, independent to the research team and identified as having expertise in at least one of the following areas: end-of-life care, palliative care, critical care and/or instrument development. Each expert was asked to rate their agreement regarding the relevance of each survey item on a 4-point rating scale: 1 = not relevant; 2 = somewhat relevant; 3 = quite relevant; 4 = highly relevant (Lynn 1986, Davis 1992, Grant & Davis 1997, Polit & Beck 2006). Relevance ratings were used to calculate the content validity index for individual items (I-CVI) and the content validity index for the section overall survey (S-CVI). The initial S-CVI for the survey was 084. Items below I-CVI 080 were reviewed and four items were deleted as they were not considered relevant to end-of-life care in the adult critical care context and/or were similar to other items rated as relevant by the expert panel. Following review and deletion of items with low I-CVI, an S-CVI of 089 was obtained. The experts were asked to suggest additional items they felt relevant to the provision of end-of-life care in critical care settings and a further two items were added to the survey. The revised 33-item survey was returned to the expert panel for final review prior to distribution to the pilot test participants. Pilot testing. The 33-item survey was circulated to a convenience sample of 51 critical care nurses enroled in a postgraduate nursing subject at one higher education institution 699
K. Ranse et al.
and 17 responses to the survey were obtained (response rate 33%). This initial pilot test of the survey revealed that some of the a priori sub-scales were not robust measures of the constructs of interest. To improve construct validity, the instrument was significantly revised following further reflection on the findings of an earlier descriptive qualitative study (Ranse et al. 2012) and a further review of the literature of end-of-life care in critical care settings. Due to the significant adjustments made to the survey, which now comprised 55 items, a repeat pilot test and retest with a new sample was required to reassess the validity and reliability of the survey. The findings presented below refer to the results of this repeat pilot test and retest. In March 2011, pilot testing was undertaken with another convenience sample of critical care nurses enroled in a postgraduate subject at a higher education institution in Queensland, Australia to obtain preliminary assessment of the psychometric properties of the instrument and the feasibility of methods of data collection, management and analysis. An email containing a link to the online survey was circulated to 62 potential participants. A single reminder email was circulated and a response rate of 45% was obtained (n = 28). To ascertain test-retest reliability of the survey, the survey was recirculated to participants after 2 weeks. Eleven participants completed the survey on both occasions. Cronbach’s alpha was calculated to identify the internal consistency of items in each a priori sub-scale. Reliability coefficients of greater than 070 are considered satisfactory (Hair et al. 2006), although lower reliability coefficients may be considered satisfactory when measuring affective domains or in exploratory research (Hair et al. 2006, Field 2009). The Pearson product-moment correlation coefficient (r) was calculated to determine test-retest reliability of the survey instrument in obtaining consistent results on repeated measures (Polit & Beck 2008). Table 2 provides the results of internal consistency and test-retest reliability for each a priori sub-scale in the survey instrument. Some of the sub-scales had satisfactory internal consistency; however, two sub-scales in particular had very low internal consistency. The a priori sub-scales nurse practice knowledge and nurses’ end-of-life care values and beliefs obtained low cronbach’s alpha indicating that the items in these subscales were measuring discrete concepts rather than a single construct. Although deleting some items slightly improved internal consistency, all items were retained as it was considered that individually, the items measured an aspect of end-of-life care of interest to this study. In addition, in an attempt to improve the internal consistency of the a priori sub-scale nurses’ end-of-life care values and beliefs, three 700
Table 2 Pilot test – Internal consistency (n = 28) and test–retest reliability (n = 11) of the a priori scales in the survey instrument. A priori scale Organisational culture Resources Patient and family preferences Nursing education and experience Nurse practice knowledge Emotional support for nurses Nurses’ end-of-life care values and beliefs Coordination of care delivery
Number of items
a
r
7 8 6 10 8 7 5
058 060 079 075 041 086 039
061* 080** 068* 077** 086** 092** 085**
4
067
076**
*P < 005, two-tailed; **P < 001, two-tailed.
items were added to the survey. Each of the a priori subscales obtained large correlation coefficients (>05) (Cohen 1988) indicating satisfactory test-retest reliability. Stage two: cross-sectional survey Population and sample. The population of interest in this study was Registered Nurses with experience in a critical care setting. A national critical care nursing association provided an accessible population. All members of this association that had agreed to be contacted for research purposes were invited to participate in this study. Data collection. In May 2011, members of a national critical care nursing association were emailed and invited to participate in the study by completing the online survey. A single reminder email was circulated 3 weeks following initial contact and the survey remained available online for a period of 6 weeks. Figure 1 provides a flowchart of participant recruitment and response to the survey. The survey software did not register partially completed survey responses, so the response rate (25%) only reflects those who completed the survey. Data analysis. Data were exported into PASW software version 18 and exploratory factor analysis using Principal Axis Factoring and Oblique rotation (direct oblimin) was undertaken on participant responses to the 58 items in the factors influencing provision of end-of-life care section of the survey instrument to identify the latent constructs in the instrument. Factorability of the data set was confirmed by obtaining a measure of sampling adequacy (MSA) greater than 050 for each item and obtaining a significant result on Bartlett’s test (Hair et al. 2006). Consistent with recommendations in the literature, multiple methods were used to determine the number of factors to be retained © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
Factors influencing end-of-life care in critical care
Potential participants n = 2300 Excluded Declined to be Contacted for research purposes n = 720
Invited to participate by email Initial invitation n = 1553 Reminder invitation n = 1580
Responded to survey n = 396
Increase in membership and members willing to be contacted for research purposes
Non-response to survey
Excluded Duplicate response n = 1 Had not provided endof-life care n = 3
Not currently working in clinical practice n = 8
Undelivered emails n = unknown
Responded for analysis n = 392 Response rate = 25%
Figure 1 Flowchart of study participants. including Kaiser’s criterion, a scree plot and the number of factors proposed a priori (Fabrigar et al. 1999, Costello & Osborne 2005, Hair et al. 2006, Bandalos & Finney 2010). Comparison of multiple factor solutions for the best fit of the data with regard to number of items loading on each factor, minimal cross-loadings and theoretical or conceptual meaningfulness was ultimately used to select the final factor solution (Floyd & Widaman 1995, Fabrigar et al. 1999, Costello & Osborne 2005, Bandalos & Finney 2010). The internal consistency of sub-scales identified by factor analysis was further assessed by calculating Cronbach’s alpha coefficient for the items in each sub-scale. A reliability coefficient greater than 070 is considered satisfactory (Hair et al. 2006). Descriptive statistics, including mean, standard deviation and frequency of participant response were calculated for each individual item in the survey. © 2014 John Wiley & Sons Ltd
Ethical considerations The University Human Research Ethics Committee provided approval for the conduct of this study, including the pilot tests and cross-sectional survey. A national critical care nursing association also reviewed the study proposal and provided administrative support through the distribution of an email invitation to potential participants.
Results Participant profile The mean age of participants was 431 years (SD 92) and 135% were male. Years of experience in nursing ranged from 2–43 years (mean = 209, SD 101), with 771% of participants having more than 10 years of experience in 701
K. Ranse et al.
nursing and 64% reporting more than 10 years of experience in the critical care setting (mean = 144 SD 82). In addition to extensive years of experience, the majority of participants had postgraduate qualifications in nursing (827%) and had completed or were currently undertaking formal education in critical care (903%).
than the broader organization as the other items in this factor. Similarly, the item ‘Patients at the end-of-life require little nursing care’ was removed from Factor 7 as it was conceptually a poor fit with the other items that loaded on this factor. The label given to each factor, number of items in the factor and Cronbach’s alpha following revision is provided in Table 5.
Psychometric properties of the instrument Factorability of the data was confirmed with correlations greater than 030 identified between the majority of items and a significant result obtained on Bartlett’s test of sphericity [v2 (1653) = 736851, P < 0001]. Although 94% of completed surveys had a missing value on at least one item, the ratio of between 5–10 participants per item recommended in the literature was achieved (Hair et al. 2006). In addition, the Kaiser–Meyer–Oklin (KMO) measure of sampling adequacy was 081 and the KMO for all individual items was greater than 050. Sample size was deemed adequate for factor analysis (Field 2009). Principal axis factoring with oblique rotation (direct oblimin) was undertaken on the data. Kaiser’s criterion suggested a 16 factor solution, the scree plot indicated a six or eight factor solution and the sub-scales proposed a priori consisted of eight factors. Therefore, comparison of factor solutions using six, seven, eight, nine and 16 factors was undertaken. Nine and 16 factor solutions were deemed unsatisfactory as they contained several factors with only two or three items loading on individual factors. The eight factor solution most closely resembled the sub-scales proposed a priori and provided the best fit of the data in terms of conceptual and theoretical meaningfulness. Out of the 58 items, 42 had a pattern coefficient greater than 030 and were considered to contribute to one of the eight factors (Table 3). There were 259 (150%) non-redundant residuals with absolute values greater than 005. Moderate correlation (>030) was identified between factors one and four and factors two and eight (Table 4). The reliability of the sub-scales identified through EFA was assessed by calculating Cronbach’s alpha to determine the internal consistency of the sub-scales. Items in each factor were reviewed to ensure that they conceptually fit. Two items (‘Pressure area care should continue to be provided for a patient at the end-of-life’ and ‘My own personal experience of death have influenced the care I provide to patients at the end-of-life’) were removed from factor two as they were not considered to conceptually fit with the other items in this scale. The item ‘after a decision is made to withdraw treatment, I spend less time with the patient’ was removed from Factor 5 as it refers to the self, rather 702
Descriptive analyses The ten items that achieved the highest ranking based on their mean score, indicating the items that achieved the highest level of agreement as factors influencing the provision of end-of-life care in critical care settings are provided in Table 6. The top five items were items from the palliative values latent construct, while three of the items from the patient and family preferences scale were also represented in the ten most strongly agreed with items.
Discussion Exploratory factor analysis identified eight factors influencing the provision of end-of-life care: emotional support for nurses; palliative values; patient and family preferences; resources; organizational support; care planning; knowledge and preparedness and these constructs conceptually fit within the three key elements of the PARIHS framework. The key elements of the PARIHS framework: context, evidence and facilitation will give an organizing framework to discuss the key findings of this study.
Context The latent constructs organizational culture and resources were considered to represent contextual factors that influence the provision of end-of-life care in critical care settings. Items in the latent construct organizational culture reflected a context of nurse-led practice and did not include items capturing the attitudes of the medical team towards end-of-life care or the broader curative culture proposed a priori. The findings of this study give evidence of a practice context of nurse-led care after a decision is made to withhold or withdraw life-sustaining treatment. These findings are consistent with the perception that critical care nurses implement withdrawal of life-sustaining treatment and manage the provision of end-of-life care reported internationally (Halcomb et al. 2004, Heland 2006, Coombs & Long 2008, Espinosa et al. 2010, Kirchhoff & Kowalkowski 2010, Ranse et al. 2012). © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
Factors influencing end-of-life care in critical care
Table 3 Factor loadings for exploratory factor analysis with oblique rotation. Factor loadings Items My supervisors/managers ensure staff caring for patients at the EOL are supported My supervisors/managers provide guidance that assists me to provide EOLC After caring for a patient at the EOL, my colleagues will ask me if I am OK I feel supported when caring for a patient at the EOL My colleagues appreciate the stressors associated with caring for patients at the EOL When a patient dies in my Unit, counselling is available if I need it In my Unit, nurses view the provision of EOLC as an important component of critical care nursing Lack of staff results in no ‘down time’ between the death of a patient and the admission of the next patient In my Unit, nurses are expected to cope with death and dying EOLC is as important as curative care in the critical care environment I can influence the patient and family’s experience of EOLC EOL care should include care of the patient’s family It is a privilege to care for a patient at the EOL and their family My own personal experiences of death have influenced the care I provide to patients at the EOL and their family Analgesia should be titrated to keep the patient comfortable even if this hastens death Pressure area care should continue to be provided for a patient at the EOL It is OK to cry when a patient dies EOLC requires advanced nursing skills I have learnt from other nurses how to provide EOLC In my Unit, families are given adequate time to consider decisions to withhold/withdraw treatment for the patient In my Unit, family meetings with the health care team are held to resolve differing views about a patient’s treatment In my Unit, the patient’s preferences (or their likely preference according to family) for EOLC are considered In my Unit, families are involved in decisions about the dying patient
© 2014 John Wiley & Sons Ltd
1
2
3
4
5
6
7
h2
8
082
005
001
006
004
003
009
004
070
074
010
007
001
003
000
013
006
062
070
021
001
005
009
006
003
002
053
067
004
001
006
001
009
004
010
059
059
017
004
006
002
017
009
013
051
053
014
004
001
001
004
024
005
039
028
022
002
012
000
017
000
004
025
028
017
015
013
016
017
004
004
022
024
021
004
006
022
016
003
001
019
002
064
003
002
001
004
007
005
046
004
058
007
004
006
004
005
015
041
000
054
011
013
005
010
005
004
037
014
053
000
002
001
014
002
012
039
008
038
004
008
000
002
007
002
016
005
035
007
003
008
001
001
011
019
004
031
001
002
001
009
002
003
011
004 008 015
029 028 020
004 001 011
000 000 015
013 000 015
007 000 015
001 027 000
008 000 014
014 018 018
002
012
076
002
004
003
007
001
057
008
015
071
004
002
009
005
004
058
012
003
057
007
001
027
009
010
049
005
007
053
011
007
008
003
010
039
703
K. Ranse et al.
Table 3 (Continued). Factor loadings Items In my Unit, withdrawal of treatment is delayed whilst waiting for relatives of the patient to arrive When a patient dies in my Unit, families have sufficient time to spend with the patient I am often exposed to death in the critical care environment Abrupt withdrawal of mechanical ventilation may contribute to dyspnoea The physical environment of my Unit is ideal for providing EOLC My Unit is adequately equipped to support the comfort needs of the family during EOLC Private rooms are available to care for the patient at the EOL My Unit is adequately staffed for providing the needs of dying patients and their families When a patient dies in my Unit, nurses have sufficient time to spend with the family In my Unit, nurses that know the patient and family are allocated to care for the patient to maintain continuity. . . Following a decision to withhold or withdraw treatment, patients remain in my Unit for EOLC After a decision is made to withdraw treatment, less time is spent with the patient by the critical care medical team After a decision is made to withdraw treatment, nurses are left to manage the care of the patient In my Unit, EOLC is mostly allocated to junior nurses After a decision is made to withdraw treatment, I spend less time with the patient The medical staff support EOLC for dying patients in my Unit In my Unit, when a diagnosis with a likely poor outcome is made, families are informed of EOLC options In my Unit, the team expresses its opinions, values and beliefs about providing care to dying patients All members of the healthcare team in my Unit agree with and support EOLC when it is implemented In my Unit, a plan of care for the patient is clearly documented by the medical team In my Unit, the critical care team and the patient’s primary medical or surgical team agree on a plan of care
704
1
2
3
4
5
6
7
h2
8
008
012
050
005
010
002
004
001
029
019
006
033
027
015
004
021
010
042
004
017
021
002
020
002
005
004
013
003
019
019
004
014
007
003
012
014
009
004
006
077
005
007
002
002
054
014
001
007
073
002
001
011
000
053
001
013
004
071
010
001
002
005
048
013
018
021
059
003
004
003
013
056
023
022
027
037
010
001
021
037
062
019
005
003
028
018
012
014
012
029
007
022
006
027
008
009
002
001
018
000
001
001
006
075
013
010
008
061
003
002
005
001
072
015
003
009
057
009
012
002
004
032
015
010
014
021
013
021
015
002
032
016
028
012
035
008
001
017
004
0117
055
007
009
045
000
004
030
002
017
053
007
002
048
020
004
001
001
004
051
001
003
038
008
009
007
019
000
047
018
004
038
001
009
017
011
024
043
001
010
040
006
003
021
006
014
042
000
004
034
© 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
Factors influencing end-of-life care in critical care
Table 3 (Continued). Factor loadings 1
Items The provision of EOLC requires emotional detachment I have received in-service education that assists me to support and communicate with families of dying patients I have received in-service education that assists me to provide care and manage patient symptoms at the EOL I draw on evidence from the literature to provide EOLC to patients and their families My postgraduate nursing education included EOLC content relevant to the critical care context Patients at the EOL require little nursing care My undergraduate nursing education included EOLC content There are policies/guidelines to assist in the delivery of EOLC in my Unit When withdrawing mechanical ventilation at the EOL, ventilator weaning should occur. . . Intravenous fluids should be provided to maintain hydration for the patient at the EOL I feel adequately prepared to care for patients at the EOL I don’t know what to say to the family of a patient at the EOL I have limited knowledge of symptom management for patient’s at the EOL I feel a sense of personal failure when a patient dies a
2
3
4
5
6
7
h2
8
004
022
006
002
021
024
012
009
017
015
020
013
005
006
026
057
028
060
011
023
016
008
004
022
054
033
057
002
008
025
002
007
010
043
013
033
016
004
005
005
004
004
043
001
026
007 012
021 001
017 012
020 013
021 004
016 001
034 028
001 002
031 014
016
001
000
011
003
026
028
004
026
001
003
005
000
004
012
027
001
009
002
002
003
005
010
006
023
003
008
002
006
002
003
007
008
008
067
052
000
006
004
005
002
002
005
064
045
006
007
007
002
010
005
007
052
029
015
014
009
013
011
006
008
029
022
085
070
079
081
064
080
066
069
Factor loadings >030 are in boldface. EOL, end of life; EOLC, end-of-life care.
Table 4 Factor correlation matrix. Factor 1 2 3 4 5 6 7 8
1
2 100 001 020 032 023 029 013 020
3 001 100 022 005 004 002 008 033
4 020 022 100 019 014 021 000 017
The responses of some participants in this study were consistent with previous research that has identified the perception that critical care and acute care environments are not adequately designed or resourced for the provision of end-of-life care (Davidson et al. 2003, Nordgren & Olsson © 2014 John Wiley & Sons Ltd
5 032 005 019 100 013 024 010 007
6 023 004 014 013 100 002 007 014
7 029 002 021 024 002 100 005 002
8 013 008 000 010 007 005 100 013
020 033 017 007 014 002 013 100
2004, Wotton et al. 2005, Beckstrand et al. 2008, Jackson et al. 2010, McKeown et al. 2010, Bloomer et al. 2013). Although variation in participant agreement relating to resource items was noted in this study, nurses can use their initiative to create a context to positively influence the 705
K. Ranse et al.
Table 5 Final factors, labels and cronbach’s alphas. Factor
Label
Items
a
1 2 3 4 5 6 7 8
Emotional support for nurses Palliative values Patient and family preferences Resources Organisational culture Care planning Knowledge Preparedness
6 5 6 5 3 6 4 3
085 073 079 081 069 080 069 069
Table 6 The 10 most strongly agreed items undertaken based on their mean score (n = 392). Item
M (SD)
EOLC should include care of the patient’s family (n = 391) Analgesia should be titrated to keep the patient comfortable even if this hastens death (n = 391) EOLC is as important as curative care in the critical care environment (n = 389) It is a privilege to care for a patient at the EOL and their family I can influence the patient and family’s experience of EOLC (n = 391) Pressure area care should continue to be provided for a patient at the EOL (n = 391) It is OK to cry when a patient dies (n = 390) In my Unit, families are involved in decisions about the dying patient In my Unit, withdrawal of treatment is delayed whilst waiting for relatives of the patient to arrive In my Unit, family meetings with the health care team are held to resolve differing views about a patient’s treatment M, mean; life care.
SD,
47 (05) 46 (06) 45 (07) 45 (07) 45 (07) 44 (07) 42 (08) 42 (07) 4. 2 (07) 41 (08)
standard deviation; EOL, end of life; EOLC, end-of-
provision of end-of-life care in critical care settings. Thus, it is perhaps not surprising that the provision of end-of-life care by nurses can occur in isolation to the broader practice context of curative care and environmental constraints.
Evidence Exploratory factor analysis of participant responses to the survey identified three latent constructs that conceptually fit within the element evidence: knowledge, preparedness and patient and family preferences. The items comprising the latent construct knowledge did not assess nurses’ actual knowledge of end-of-life care, but rather opportunities for 706
knowledge acquisition through continuing professional education. Participants in this study generally agreed that they felt adequately prepared for the provision of end-of-life care despite having limited access to opportunities for knowledge acquisition. Similarly, Kirchhoff and Kowalkowski (2010) report that almost half (497%) of critical care nurses in their study were very confident in their ability to give end-of-life care, despite limited education and training. It should be noted that the majority of participants in both studies were experienced critical care nurses and previous research has identified that actual experience in the provision of end-of-life care contributes to nurse practice knowledge in this area (Heland 2006, Fridh et al. 2009, Zomorodi & Lynn 2010). In this study, the majority of participants agreed or strongly agreed with statements reflecting inclusion of patient and family preference in their unit. The view of participants in this study is consistent with a patient and family centred approach to care at the end-of-life (Truog et al. 2008). Inclusion of families in end-of-life decision-making and the actual provision of end-of-life care is seen as important to enable a good death (Borbasi et al. 2005).
Facilitation A conceptual analysis of facilitation identified variation in description and use of the term in the literature (Harvey et al. 2002). EFA of participant responses to the survey instrument identified the latent constructs palliative values, emotional support for nurses and care planning that conceptually fit within the element facilitation as defined in this study. There was strong agreement by participants to the items in the survey instrument that loaded on the latent construct palliative values and these items reflected values consistent with a palliative approach. These findings indicate that participants in this study have a strong commitment to the inclusion of families in end-of-life care and the value of this care in the critical care setting. A shift in the focus of care at the end of life to the family (Halcomb et al. 2004, Fridh et al. 2009) and reports of end-of-life care being considered a privilege in the critical care setting (Calvin et al. 2007, Fridh et al. 2009, Vanderspank-Wright et al. 2011) have been identified in previous studies. The latent construct of emotional support for nurses captured nurses’ perceptions of the support and assistance they received from their colleagues and managers in the workplace when providing end-of-life care. Strong agreement was noted with the item referring to collegial support. Informal support through talking with colleagues has been © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT
widely reported in previous research (Jones & FitzGerald 1998, Kirchhoff et al. 2000, Halcomb et al. 2004, Shorter & Stayt 2010). The items in the care planning latent construct reflected support for end-of-life care and agreement on a plan of care among the healthcare team. Adequate guidance through a written plan of care or standardised order form may facilitate the provision of good end-of-life care (Truog et al. 2008), particularly for nurses with less experience in the provision of end-of-life care in this setting.
Study limitations The PARIHS framework was selected as an organizing framework in this study to identify the factors that may influence the provision of end-of-life care by critical care nurses. Although the key elements of the framework were thought to capture the complexity of factors influencing end-of-life care practice in the critical care context, some of the sub-elements identified in the PARIHS framework were not considered relevant. Instead, a range of factors influencing practice for each key element was proposed a priori. Eight sub-scales identified through EFA were considered to have conceptual and theoretical fit with the factors proposed a priori and the three key elements of the PARIHS framework, providing preliminary evidence of the construct validity of the survey. The sub-scales identified through EFA are derived from participant responses to the items included in the survey. Although a rigorous process of item generation and instrument development was undertaken, other latent constructs of factors influencing end-of-life care may exist and may not have been identified. However, the findings presented here are considered to have conceptual meaningfulness. Replication of this study using a new sample, alternative methods of evaluating test-retest reliability and exploratory or confirmatory factor analysis would give additional evidence of the reliability and validity of the survey. This study was limited to a non-probability convenience sample of critical care nurses and those that chose to participate may be characteristically different to those who chose not to participate. It should be noted that the response rate to the pilot test was 45% (n = 28) and only 11 participated in the test-retest. Despite this small sample size, the pilot test provided some preliminary evidence of the psychometric properties of the survey instrument. Although a larger sample size was obtained in the cross-sectional survey (n = 392), the survey was distributed through the membership database of an Australian nursing association and due to privacy restrictions, the researchers were unable track undelivered emails or © 2014 John Wiley & Sons Ltd
Factors influencing end-of-life care in critical care
attribute lack of participation due to not currently working in clinical practice. Therefore, the response rate of 25% underestimates the actual response rate.
Conclusion This study has provided preliminary evidence of the psychometric properties of a new instrument to identify factors influencing the provision of end-of-life care by critical care nurses. The survey may be used in critical care units to identify factors influencing the provision of end-of-life care. Analysis of participant responses to the survey may inform the development of strategies to prepare and support nurses in the delivery of end-of-life care, improving the care that patients and their families receive. The survey does not attempt to define good end-of-life care or consider the perspectives of recipients of care or other healthcare professionals involved in the provision of end-of-life care in critical care settings. Future research could address this gap. Further testing of the instrument in a new sample using confirmatory factor analysis or structural equation modelling would provide additional evidence of the psychometric properties of the instrument.
Acknowledgements We extend our thanks to the nurses who contributed to this study by providing expert review of the survey instrument or responding to the online survey.
Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest No conflict of interest.
Author contributions All authors have agreed on the final version and meet at least one of the following criteria [recommended by the ICMJE (http://www.icmje.org/ethical_1author.html)]:
• •
substantial contributions to conception and design, acquisition of data or analysis and interpretation of data; drafting the article or revising it critically for important intellectual content. 707
K. Ranse et al.
Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site.
References Bandalos D.L. & Finney S.J. (2010) Factor analysis: exploratory and confirmatory. In The Reviewers Guide to Quantitative Methods in the Social Sciences (Hancock G.R. & Mueller R.O., eds), Routledge, Hoboken NJ, pp. 93–114. Beckstrand R.L. & Kirchhoff K.T. (2005) Providing end-of-life care to patients: critical care nurses’ perceived obstacles and supportive behaviors. American Journal of Critical Care 14(5), 395–403. Beckstrand R.L., Callister L.C. & Kirchhoff K.T. (2006) Providing a ‘good death’: critical care nurses’ suggestions for improving end-of-life care. American Journal of Critical Care 15(1), 38–45. Beckstrand R.L., Smith M.D., Heaston S. & Bond A.E. (2008) Emergency nurses’ perceptions of size, frequency and magnitude of obstacles and supportive behaviours in end-of-life care. Journal of Emergency Nursing 34(4), 290–300. Bloomer M.J., Triuvoipati R., Tsiripillis M. & Botha J.A. (2010) End of life management of adult patients in an Australian metropolitan intensive care unit: a retrospective observational study. Australian Critical Care 23(1), 13–19. Bloomer M.J., Morphet J., O’Connor M., Lee S. & Griffiths D. (2013) Nursing care of the family before and after a death in the ICU—An exploratory pilot study. Australian Critical Care 26(1), 23–28. Borbasi S., Wotton K., Redden M. & Chapman Y. (2005) Letting go: a qualitative study of acute care and community nurses’ perceptions of a ‘good’ versus a ‘bad’ death. Australian Critical Care 18(3), 104–113. Calvin A.O., Kite-Powell D.M. & Hickey J.V. (2007) The neuroscience ICU nurse’s perceptions about end-of-life care. Journal of Neuroscience Nursing 39(3), 143–150. Cohen J. (1988) Statistical Power Analysis for the Behavioral Sciences, 2nd edn. Lawrence Erlbaum, Hillsdale, NJ. Coombs M. & Long T. (2008) Managing a good death in critical care: can health policy help? Nursing in Critical Care 13(4), 208–214. Costello A.B. & Osborne J.W. (2005) Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Practical Assessment Research & Evaluation 10 (7), 1–9. Davidson P.M., Introna K., Daly J., Paull G., Jarvis R., Angus J., Wilds T., Cockburn J., Dunford M. & Dracup K. (2003) Cardiorespiratory nurses’ perceptions of palliative care in nonmalignant disease: data for the development of clinical practice. American Journal of Critical Care 12(1), 47–53. Davis L.L. (1992) Instrument review: getting the most from a panel of experts. Applied Nursing Research 5(4), 194–197. Espinosa L., Young A., Symes L., Haile B. & Walsh T. (2010) ICU nurses’ experiences in providing terminal care. Critical Care Nursing Quarterly 33(3), 273–281. Fabrigar L.R., Wegener D.T., MacCallum R.C. & Strahan E.J. (1999) Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods 4(3), 272–299.
708
Field A. (2009) Discovering Statistics using SPSS, 3rd edn. SAGE, London. Floyd F.J. & Widaman K.F. (1995) Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment 7(3), 286–299. Fridh I., Forsberg A. & Bergbom I. (2009) Doing one’s utmost: nurses’ descriptions of caring for dying patients in an intensive care environment. Intensive and Critical Care Nursing 25(5), 233–241. Grant J.S. & Davis L.L. (1997) Selection and use of content experts for instrument development. Research in Nursing & Health 20(3), 269–274. Hair J.F., Black W.C., Babin B.J., Anderson R.E. & Tatham R.L. (2006) Multivariate Data Analysis, 6th edn. Pearson Prentice Hall, Upper Saddle River, NJ. Halcomb E., Daly J., Jackson D. & Davidson P. (2004) An insight into Australia nurses’ experience of withdrawal/withholding of treatment in the ICU. Intensive and Critical Care Nursing 20(4), 214–222. Harvey G., Loftus-Hills A., Rycroft-Malone J., Titchen A., Kitson A., McCormack B. & Seers K. (2002) Getting evidence into practice: the role and function of facilitation. Journal of Advanced Nursing 37(6), 577–588. Heland M. (2006) Fruitful or futile: intensive care nurses’ experiences and perceptions of medical futility. Australian Critical Care 19(1), 25–31. Jackson A., Purkis J., Burnham E., Hundt G.L. & Blaxter L. (2010) Views of relatives, carers and staff on end-of-life care pathways. Emergency Nurse 17(10), 22–26. Jones T. & FitzGerald M. (1998) Withdrawal of life-support treatment: the experience of critical care nurses. Australian Critical Care 11(4), 117–121. Kain V., Gardner G. & Yates P. (2009) Neonatal Palliative Care Attitude Scale: development of an instrument to measure the barriers to and facilitators of palliative care in neonatal nursing. Pediatrics 123(2), e207–e213. Kirchhoff K.T. & Beckstrand R.L. (2000) Critical care nurses’ perceptions of obstacles and helpful behaviours in providing endof-life care to dying patients. American Journal of Critical Care 9(2), 96–105. Kirchhoff K.T. & Kowalkowski J.A. (2010) Current practices for withdrawal of life support in intensive care units. American Journal of Critical Care 19(6), 532–541. Kirchhoff K.T., Spuhler V., Walker L., Hutton A., Vaughan Cole B. & Clemmer T. (2000) Intensive Care Nurses’ experiences with end-of-life care. American Journal of Critical Care 9(1), 36–42. Kitson A., Harvey G. & McCormack B. (1998) Enabling the implementation of evidence based practice: a conceptual framework. Quality in Health Care 7(3), 149–158. Le Conte P., Riochet D., Batard E., Volteau C., Giradudeau B., Arnaudet I., Labastire L., Levraut J., Thys F., Lauque D., Piva C., Schmidt J., Trewick D. & Potel G. (2010) Death in emergency departments: a multicentre cross-sectional survey with analysis of withholding and withdrawing life support. Intensive Care Medicine 36(5), 765–772. Luce J.M. (1997) Making decisions about the forgoing of lifesustaining therapy. American Journal of Respiratory Critical Care Medicine 156(6), 1715–1718.
© 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY: INSTRUMENT DEVELOPMENT Lynn M.R. (1986) Determination and quantification of content validity. Nursing Research 35(6), 382–385. McCormack B., Kitson A., Harvey G., Rycroft-Malone J., Titchen A. & Seers K. (2002) Getting evidence into practice: the meaning of ‘context’. Journal of Advanced Nursing 38(1), 94–104. McKeown A., Cairns C., Cornbleet M. & Longmate A. (2010) Palliative care in the intensive care unit: an interview-based study of the team perspective. International Journal of Palliative Nursing 16(7), 334–338. Nelson J.E., Angus D.C., Weissfeld L.A., Puntillo K.A., Danis M., Deal D., Levy M. & Cook D.J. (2006) End-of-life care for the critically ill: a national intensive care unit survey. Critical Care Medicine 34(10), 2547–2553. Nordgren L. & Olsson H. (2004) Palliative care in a coronary care unit: a qualitative study of physicians’ and nurses’ perceptions. Journal of Clinical Nursing 13(2), 185–193. Polit D.F. & Beck C.T. (2006) The content validity index: are you sure you know what’s being reported? Critique and recommendations. Research in Nursing & Health 29(5), 489–497. Polit D.F. & Beck C.T. (2008) Nursing Research: Generating and Accessing Evidence for Nursing Practice, 8th edn. Wolters Kluwer/Lippincott, Williams & Wilkins, Philadelphia, PA. Ranse K., Yates P. & Coyer F. (2012) End-of-life care in the intensive care setting: a descriptive exploratory qualitative study of nurses’ beliefs and practices. Australian Critical Care 25(1), 4–12. Rycroft-Malone J., Kitson A., Harvey G., McCormack B., Seers K., Titchen A. & Estabrooks C. (2002) Ingredients for change:
Factors influencing end-of-life care in critical care
revisiting a conceptual framework. Quality & Safety in Health Care 11(2), 174–180. Rycroft-Malone J., Seers K., Titchen A., Harvey G., Kitson A. & McCormack B. (2004) What counts as evidence in evidencebased practice? Journal of Advanced Nursing 47(1), 81–90. Seymour J. (2001) Critical Moments: Death and Dying in Intensive Care. Open University Press, Buckingham, UK. Shorter M. & Stayt L.C. (2010) Critical care nurses’ experiences of grief in an adult intensive care unit. Journal of Advanced Nursing 66(1), 159–167. Truog R.D., Campbell M.L., Curtis J.R., Haas C.E., Luce J.M., Rubenfeld G.D., Rushton C.H. & Kaufman D.C. (2008) Recommendations for end-of-life care in the intensive care unit: a consensus statement by the American College of Critical Care Medicine. Critical Care Medicine 36(3), 953–963. Vanderspank-Wright B., Fothergill-Bourbonnais F., Brajtman S. & Gagnon P. (2011) Caring for patients and families at end of life: the experiences of nurses during withdrawal of life-sustaining treatment. Dynamics 22(4), 31–35. Wotton K., Borbasi S. & Redden M. (2005) When all else has failed: nurses’ perception of factors influencing palliative care for patients with end-stage heart failure. Journal of Cardiovascular Nursing 20(1), 18–25. Zomorodi M. & Lynn M.R. (2010) Critical care nurses’ values and behaviours with end-of-life care: perceptions and challenges. Journal of Hospice and Palliative Nursing 12(2), 89–96.
The Journal of Advanced Nursing (JAN) is an international, peer-reviewed, scientific journal. JAN contributes to the advancement of evidence-based nursing, midwifery and health care by disseminating high quality research and scholarship of contemporary relevance and with potential to advance knowledge for practice, education, management or policy. JAN publishes research reviews, original research reports and methodological and theoretical papers. For further information, please visit JAN on the Wiley Online Library website: www.wileyonlinelibrary.com/journal/jan Reasons to publish your work in JAN:
• High-impact forum: the world’s most cited nursing journal, with an Impact Factor of 1·527 – ranked 14/101 in the 2012 ISI Journal Citation Reports © (Nursing (Social Science)).
• Most read nursing journal in the world: over 3 million articles downloaded online per year and accessible in over 10,000 libraries worldwide (including over 3,500 in developing countries with free or low cost access).
• • • •
Fast and easy online submission: online submission at http://mc.manuscriptcentral.com/jan. Positive publishing experience: rapid double-blind peer review with constructive feedback. Rapid online publication in five weeks: average time from final manuscript arriving in production to online publication. Online Open: the option to pay to make your article freely and openly accessible to non-subscribers upon publication on Wiley Online Library, as well as the option to deposit the article in your own or your funding agency’s preferred archive (e.g. PubMed).
© 2014 John Wiley & Sons Ltd
709
