A Single-center Exploration of Attitudes to Deceased Organ Donation Over Time Among Healthcare Staff in Intensive Care

Abstract

Background.

Changes to deceased organ donation in the United Kingdom, including establishment of the specialist nurse for organ donation (SNOD) role, have resulted in increased numbers of donations. Have increasing numbers of donations altered attitudes among intensive care unit (ICU) healthcare professionals (ICU staff) to organ donation over time?

Methods.

A written survey of ICU staff at Nottingham University Hospitals National Health Service Trust was conducted across 2 wk in 2015, 2018, and 2020 (pre–COVID-19). Participants were asked to submit descriptors (words/phrases) they associated with 3 aspects of donation: donation after brain death (DBD), donation after circulatory death (DCD), and SNOD role. Three independent and blinded assessors categorized the descriptors as positive or negative in favorability. Thematic analysis was used to identify trends within each group of descriptors.

Results.

Across the 3 surveys, 281 responses were returned, containing a total of 2095 descriptors. Positive descriptors were found in 65% of DBD responses, 46% of DCD responses, and 92% of SNOD role. Over time, there was some evidence of increased polarization of opinion for DCD and to a smaller degree DBD. Attitude toward the SNOD role remained consistently highly favorable over time. Thematic analysis was correlated with the assessor favorability ratings to identify specific factors for positive or negative attitudes; this demonstrated the themes that were the most common causes of positive or negative attributions for each aspect of organ donation.

Conclusions.

ICU staff were found to be highly favorably positive toward the SNOD role, positive toward DBD, and negative toward DCD. Although we found broadly positive perceptions of the benefits of deceased organ donation, negative attitudes toward DCD centered on timescale and complexity of the donation process. Measurement of staff attitudes to organ donation may allow targeted interventions that support staff and improve patient and family care through the organ donation process.

Deceased organ donation practice has changed significantly in the United Kingdom after implementation of the Organ Donation Taskforce report in 2008 and introduction of the Specialist Nurse for Organ Donation (SNOD) role, Clinical Lead for Organ Donation role, and Organ Donation Committee.¹ The SNOD, although working in the intensive care unit (ICU), is employed externally by National Health Service (NHS) Blood and Transplant and has a clearly defined role in donor characterization, donor family assent to donation, donor optimization, and facilitation of organ and tissue retrieval.²

In the decade before the COVID-19 pandemic, donor numbers had nearly doubled, with increases in donation after brain(stem) death (DBD) and an even greater increase in donation after circulatory death (DCD).³ The rise in DCD was not without ethical, legal, and professional challenges.⁴

Reflecting but exceeding the wider UK experience, Nottingham University Hospitals NHS Trust (NUH), a large tertiary referral center in the East Midlands, has facilitated a sustained rise in the number of deceased organ donors from 4 in August 2007 (4 DBD, 0 DCD) to 39 in February 2019 (23 DBD, 16 DCD). In 2011, a survey of ICU staff from NUH and the surrounding Mid Trent Critical Care Network was carried out to explore attitudes to deceased organ donation.⁵ Results suggested a significant variation in attitudes toward DBD and DCD: 73% positive descriptors for DBD, compared with 25% for DCD. A causative factor may have been lesser familiarity with the process of DCD, which had only commenced in NUH in mid-2008. Notably, SNOD descriptors were overwhelmingly positive (84%). The authors recommended further study to assess ongoing development in attitudes.

This article represents the results of 3 subsequent attitude surveys of healthcare professionals working in Critical Care at NUH (ICU staff) toward deceased organ donation, performed in 2015, 2018, and 2020. It was postulated that with the increased donation numbers witnessed in NUH over time, this would be associated with positive changes in the attitudes of ICU staff, as a group, to DBD, DCD, and the SNOD role.

MATERIALS AND METHODS

Aims

This study explored attitudes toward DBD, DCD, and the SNOD, among ICU staff in a single center over time. Although not all directly comparable, these 3 areas were chosen for their familiarity and salience within contemporary UK deceased organ donation practice; the choice of broad topics also enabled subjects to register a diverse and non-predetermined range of responses.

The primary aims were the identification of differences in attitudes to these 3 aspects of donation and the quantification of trends over time. Secondary aims included the identification of recurring themes within responses.

Data Acquisition

Attitudes of staff toward organ donation were recorded using a written survey conducted across the 2 critical care units at NUH over conveniently chosen 2-wk periods in 2015, 2018, and 2020. Participants were recruited across a range of shifts; there was no selection for professional background, experience, or demographic, and no minimum experience of critical care or organ donation was stipulated. Surveys were distributed in person by an Intensive Care Medicine speciality trainee, SNOD, or medical student; explanation and instructions were standardized using a written proforma. Consent of participants was implied by voluntary completion of the proforma. The Health Research Authority decision toolkit indicated no requirement for Research Ethics Committee approval.⁶

The proforma was composed of 2 sections: a tick box demographic section (participant gender and professional group) followed by free-text components in which participants were asked “What words do you associate with Donation after Brainstem Death (DBD) […] Donation after Circulatory Death (DCD) [... and] the Specialist Nurse for Organ Donation (SNOD).” Responses were handwritten with no format or length specified. Proformas were anonymously returned via a box in each critical care unit.

Data Processing

Responses were obtained from proformas and collated into a spreadsheet. Single words or short phrases were transcribed directly (hereafter termed “descriptors”); longer phrases with multiple concepts were divided into 2 or 3 distinct descriptors to permit individual assessment. Data were tabulated alphabetically alongside study year and survey question, enabling isolation of separate data sets for each aspect of donation.

All proformas for which demographic information was completed were included in the analysis. Criteria for exclusion of individual responses were illegibility, a response of “none” (although “don’t know” or “no idea” were included), and responses that repeated the wording of questions without elaboration (eg, DBD, SNOD).

Favorability Analysis

Derivation of numerical data from the qualitative data set enabled comparison between subgroups. The study used the social psychology definition of an attitude: an evaluation of an object on a spectrum of favorability ranging from positive to negative. This provided a unified metric that could be applied across all subject responses. Each choice of words was assumed to be innately reflective of participants’ attitudes, and thus all responses were deemed eligible for an assessment of favorability.

Following the completion of the 2020 survey, a spreadsheet of the descriptors, combined from all years and questions, was individually submitted to 3 independent and blinded assessors: a healthcare lawyer, a data analyst, and a medical student. None had prior direct clinical experience in critical care or organ donation and were chosen to reflect a combination of medical and nonmedical backgrounds.

Assessors were asked to indicate whether they individually perceived each descriptor as a binary choice of either “positive” or “negative.” Each assessor worked independently. Completed assessments were compiled into a single spreadsheet, and functions were used to calculate agreement between the 3 sets of scores. Each response was thus denoted as “unanimous positive,” “majority positive,” “majority negative,” or “unanimous negative.”

Thematic Analysis

A thematic analysis was conducted by the research team, independent of the favorability analysis. Thematic analysis is a commonly used qualitative research tool in social sciences; the authors followed the reflexive method specified by the psychologists Braun and Clarke⁷ (as described further) to identify recurring themes across the 3 sets of descriptors.

The survey format produced a data set of individual phrases so a primary coding process before analysis was not required. An inductive approach was then used to organize descriptors into a series of themes, with every descriptor coded to a single theme. The themes were not predetermined before analysis; the researchers developed them in response to clusters of meaning interpreted within each data set. The 3 data sets were analyzed over 3 successive readings by 2 researchers (B.H. and L.H.), against which the emergent themes were tested: the objective was to produce a minimal number of coherent themes, each described by between 1 and 3 related terms.

The results were formulated on a spreadsheet and functions were used to calculate the occurrence of each theme. A novel methodology was used to derive a numerical value for how favorable descriptors were, based on the “positive” or “negative” perceptions of the group of assessors. Favorability assessments for individual descriptors were correlated with each theme by assigning a numerical value corresponding to the assessor consensus: +2 or –2 for a unanimous positive or negative assessment and +1 or –1 for a majority positive or negative assessment. A mean value was calculated from the total for each theme to adjust for group size and provide a comparable scale of favorability between +2 (positive) and –2 (negative). This allowed a correlation of themes with a quantification of the favorability assessments.

RESULTS

Survey Completion

Data collection took place in October 2015, January 2018, and February 2020. There were a total of 281 participants: comprising 93 (2015), 108 (2018), and 80 (2020). Comparison by year demonstrated that a consistent majority of participants were female (81% in 2015, 72% in 2018; 71% in 2020). Nurses constituted the majority of participants in each year (84% in 2015, 58% in 2018, and 73% in 2020), see Figure 1. Survey response rates were estimated at 61% (2015), 62% (2018), and 42% (2020), calculated by comparing the number of responses to the total number of staff scheduled to be on shift during the study period.

FIGURE 1.

Survey participant profession by year (%).

A total of 2095 descriptors were assessed; this comprised 773 in 2015, 766 in 2018, and 556 in 2020. The total number of descriptors per aspect of organ donation was 729 for DBD, 592 for DCD, and 774 for SNOD. Adjusted for the number of survey participants, the mean number of DCD descriptors was consistently lower than for DBD or SNOD (Figure 2).

FIGURE 2.

Mean number of descriptors by subject and year.

Favorability Analysis

Judgments were unanimously positive for 59% of all descriptors (1129 of 2095) and unanimously negative for 23% (486 of 2095). Figures 3A and 3B outline the favorability assessments over the 3 surveys for DBD and DCD, respectively. Figure 3C outlines the favorability assessments for attitudes to the distinct concept of the SNOD role.

FIGURE 3.

A, Favorability assessment of DBD descriptors by year. B, Favorability assessment of DCD descriptors by year. C, Favorability assessment of SNOD descriptors by year. DBD, donation after brain death; DCD, donation after circulatory death; SNOD, specialist nurse for organ donation.

Favorability assessments of 729 DBD descriptors were categorized by assessor consensus and year. Positive assessments (unanimous and majority) were made of 65% of all descriptors. This pattern was consistent across all years, including when nonunanimous decisions were excluded (64% positive; 349 of 542).

Of 592 DCD descriptors, 46% of the assessments were positive (unanimous and majority). Between 2015 and 2020, there was a 6% absolute rise in the proportion receiving unanimous positive assessments; however, there was a corresponding 5% fall in the majority of positive assessments: the percentage of combined positive assessments was therefore consistent over the 3 surveys (45%, 48%, and 46%, respectively). Exclusion of nonunanimous assessments (all years) revealed a 44% positive favorability in DCD (202 of 461).

Favorability assessments for 774 SNOD descriptors were categorized by assessor consensus and year. Positive assessments (unanimous and majority) were made of 92% of all responses. This pattern was consistent across all years, and the positive proportion increased further when nonunanimous judgments were excluded (95%; 678 of 712).

Thematic Analysis

Organization of responses into themes (Table 1) was conducted for the 3 data sets, combining responses from all survey years. From the 729 DBD descriptors, 16 themes were identified; an initial analysis of the 592 DCD descriptors then demonstrated that the same themes could be applied. Two further reviews of each data set were used to define themes more precisely and reallocate responses as appropriate. The incidence of responses within each theme varied between the data sets, with a range of 11 to 119 descriptors per theme for DBD, and 7 to 123 for DCD. The frequency of themes was unevenly distributed, with the 5 most frequent themes representing 62% (451 of 729) of DBD descriptors and 65% of DCD descriptors (382 of 592).

TABLE 1.

Example themes and corresponding descriptors

Theme	Example descriptors
Life saving; beneficial	“Helps others survive when you are dead,” “New life,” “Important”
Timescale of donation process	“Patient taken to theater 5 minutes after death,” “Time for family,” “Clock watching”
Complexity of donation process	“Complex retrieval process,” “multi-agency working”
Physiological/pathological process	“Kidneys, liver, body tissue,” “No function to brain stem”
Emotional burden/loss	“Sad for patient and family,” “Scary for family,” “Feel like a vulture,” “Stressful”

Many themes appeared with a comparable relative frequency between the data sets (Table 2), these included: “life saving; beneficial,” “complexity of donation process,” “physiological/pathological process,” and “emotional burden/loss.” More frequent themes for DBD compared with DCD included “certainty; understanding; predictability” (29 versus 8), “respect; compassionate care” (35 versus 13), and the “diagnosis of death” (32 versus 8), of which 31 descriptors referred directly to brainstem death testing. Themes occurring with greater frequency for DCD compared with DBD included the “timescale of donation process” (123 versus 27), “uncertainty; confusion” (40 versus 23), and “intrusive; depersonalizing” (22 versus 11).

TABLE 2.

Full list of emergent themes from DBD, DCD, and SNOD thematic analysis in descending frequency, top 5 most frequent themes in each category in bold

DBD themes	DCD themes	SNOD themes
Life saving; beneficial	Time scale of donation process	Empathetic; caring
Complexity of donation process	Life saving; beneficial	Specialist knowledge/experience
Psychological/pathological process	Complexity of donation process	Process/coordination of donation
Emotional burden/loss	Physiological/pathological process	Beneficial
Virtuous; praiseworthy	Emotional burden/loss	Support (nonspecific)
Family experience/dialog	Uncertainty; confusion	Family support/liaison
Respect; compassionate care	Virtuous; praiseworthy	Virtuous; praiseworthy
Diagnosis of death	Family experience/dialog	Communication skills
Certainty; understanding; predictability	Chance/magnitude of donation success	Hardworking; commitment
Timescale of donation process	Intrusive; depersonalizing	Availability; welcoming
Hope; closure; legacy	Respect; compassionate care	Staff support/advice
Uncertainty; confusion	Hope; closure; legacy	Advocate for patient
Chance/magnitude of donation success	Choice; decision	Ambiguity of role
Healthcare professionals	Certainty; understanding; predictability	Complex decisions/situation; diplomacy
Choice; decision	Diagnosis of death	End-of-life care; respect
Intrusive; depersonalizing	Healthcare professionals	Calm; reassuring
		Inappropriate; brutal
		Advocate for donation

DBD, donation after brain death; DCD, donation after circulatory death; SNOD, specialist nurse for organ donation.

Some themes centered on innately favorable or unfavorable concepts (eg, virtuous; praiseworthy or emotional burden/loss). However, other themes were more complex; this was elucidated by their correlation with a quantification of the favorability assessments. By assigning each response a numerical value corresponding to the panel consensus (+2 or –2 for a unanimous positive or negative assessment and +1 or –1 for a majority positive or negative assessment), a mean value was calculated from the total for each theme to adjust for group size and provide a comparable scale of favorability between +2 (positive) and –2 (negative). Results are shown in Figure 4A and B for the DBD and DCD data sets using a diverging color-coded treemap to depict the relative frequency and favorability of the 10 most frequently occurring themes. Figure 4C displays the SNOD data set by mean favorability score.

FIGURE 4.

A, Ten most frequent DBD themes by frequency (area) and favorability (color). B, Ten most frequent DCD themes by frequency (area) and favorability (color). C, Ten most frequent SNOD themes by frequency (area) and favorability (color). DBD, donation after brain death; DCD, donation after circulatory death; SNOD, specialist nurse for organ donation.

For DBD, the most frequent themes predominantly comprised positive descriptors. The opposite pattern was seen for DCD and corresponded with 2 factors: the increased frequency of generally unfavorable themes (eg, certainty, understanding, predictability) and more negative associations for certain themes, such as the “timescale of donation process” (mean favorability score –0.05 versus 1.17) and the “family experience/dialog” (0.41 versus 1.22).

Thematic analysis was repeated for the 774 SNOD descriptors; 18 themes were identified, ranging from 3 to 128 responses per theme. As with the other data sets, the 5 most frequently occurring themes accounted for 63% of all descriptors (486 of 774). The most frequently occurring theme was “empathetic; caring,” followed by “specialist knowledge/experience.” Results are shown in Figure 4C and Table 2.

Only 4 SNOD themes had a mean favorability <1.60, all of which were of low frequency (range, 5–13). Among these were a recognition of the complex situations managed by the SNOD team (score –0.33) and perceptions of the role itself (score –0.38); the latter was based on uncertainty of the role or objections to the title (eg, "terrible acronym").

DISCUSSION

Worldwide, deceased organ donation is almost universally preceded by ICU admission. Yet intensive care staff are known to have a range of attitudes to organ donation, which may result in implicit or explicit bias in their practice and impact the donation process.^5,8

We conducted a pragmatic single-center survey of ICU staff attitudes, as a group, to deceased organ donation over time. Our hypothesis was that increased donation numbers within the ICU over time would lead to greater familiarity and hence more favorable attitudes among staff. Instead despite an increase in deceased donors from our ICUs over the proceeding 13 y (DBD 4 [2007/8], 23 [2019/20]; DCD 0 [2007/8], 16 [2019/20]), attitudes remained remarkably unchanged from what our original 2011 survey had discovered.⁵ We found that our ICU staff associated DBD with positive descriptions and the SNOD role with highly positive descriptions, but that DCD was associated with more negative descriptions. We conclude that this demonstrates more favorable collective attitudes to DBD and the SNOD role and less favorable attitudes to DCD.

The only change we observed between survey periods was an increase in the proportion of descriptors for DCD (and to a smaller degree DBD), for which the assessors were unanimous in their rating of favorability. This suggests a less ambiguous use of language over time and possibly increased polarization of attitudes, although a stronger conclusion would be beyond the scope of this study.

A previous review of attitudinal data collected from healthcare staff showed that there exists a strong correlation between positive attitudes to organ donation and high donation performance in the country in which they work, although findings were not presented separately for DCD and DBD donation.⁸ This contrasts with the unchanged negative attitudes to DCD in our study despite the observed increase in DCD numbers over time. A literature review of the views of healthcare staff on organ donation found that the proportion of doctors with positive attitudes to organ donation was higher than that of nurses and that there was minimal investigation into the attitudes of other allied health professionals.⁹ We found no evidence of this difference between doctor and nurse attitudes in our survey or when comparing with allied health professional participants, although sample size was small for the allied health professional group.

Our thematic analysis demonstrates that ICU staff have broadly positive perceptions of the benefits of deceased organ donation and the praiseworthiness of organ donors. Attitudes to DBD most frequently concerned explicitly positive concepts (eg, gift of life) with negative responses generally contained to themes of death, disease, and bereavement. Overtly positive themes were less frequently associated with DCD; instead, there was a greater emphasis on the likelihood and impact of “unsuccessful” donations that do not proceed (frequent disappointment for families).

Process themes of organ donation, including referral and donor management, are featured frequently for both pathways. Regarding DBD, broadly positive terms such as “organized” and “controlled” were most common; this may reflect both the clear national guidance on the neurological diagnosis of death and the interval of relative stability that follows. Additional local factors may also include an above-average familiarity with DBD as a tertiary neurosurgery and Major Trauma service.

Negative attitudes to DCD are not readily explained by lesser familiarity; referral numbers and family approaches for DCD now exceed DBD both at NUH and nationally.³ Negative attitudes to DCD frequently focused on the timescale and complexity of the process. The most common descriptor was “rushed.” Within United Kingdom DCD practice, donation does not proceed after family consent on 30% of occasions because of prolonged time to asystole.³ The associated uncertainty and disappointment is therefore an inherent challenge.

Of interest, the international academic controversy surrounding brain death or the diagnosis of death by circulatory criteria was not expressed in the responses of our ICU staff; no responses questioned the neurological diagnosis of death, for example. DBD was associated with quality end-of-life care (more relaxed and dignified) and positive family experiences (family have time to say goodbye). By contrast, DCD elicited the most negatively worded attitudes: for example, “sometimes feel like a ‘vulture’ waiting to take to theater.” This finding of less favorable attitudes to DCD in our ICU staff has helped to focus ongoing staff educational and development packages relating to organ donation at NUH.

Given the dramatic and sustained reduction in deceased organ donation in the United Kingdom caused by the COVID-19 pandemic,^3,10 greater recognition and response to ICU staff attitudes to organ donation is essential. The timing of our surveys was such that ICU staff attitudes after the pandemic were not captured within our data, but this would be a key area of ongoing research both locally and in the wider intensive care community.

Limitations

The single-center design of our study may restrict generalizability. Additionally, we did not examine an individual ICU staff member’s change in attitude over time. Some of the surveys were distributed by SNODs, and despite the anonymous and voluntary nature of the survey, this may have introduced bias into the SNOD responses.

Our survey was pragmatically designed to be minimally prescriptive and to decrease researcher influence, allowing a broad range of responses. However, the varying number of responses per subject meant that subjects were unequally represented within the data set. The methodology required assumptions about the definition and nature of attitudes, specifically regarding their expression, representation as data, and analysis in binary terms of favorability. It was assumed that favorability can be adequately implied from language, even devoid of context. Such issues offer no easy solutions and are inherent to any comparable qualitative research; these are therefore important caveats but should not necessarily undermine the validity of the conclusions.

Analysis of responses required derivation of quantifiable results from qualitative data. Although less well established within medical research, combined quantitative and qualitative methodologies were used within 42% of previous UK studies of healthcare worker attitudes to organ donation.⁹ Validated tools exist for the measurement of attitudes to donation within general populations¹¹; however, predetermined questions and nonspecificity to ICU staff excluded their use here. Favorability analysis was quantified using an external mechanism in preference to self-assessed scores, which may be insensitive to unconscious attitudes, inaccurately reported, or affected by observer-expectancy effects, whereby participant behavior is altered to conform to perceptions of an externally desired outcome.¹² Therefore, although our favorability analysis used an unvalidated approach, our hope was that the triangulation of data through a blinded and independent group of assessors minimized researcher bias. Our 3 blinded assessors were not experienced in ICU or donation so bias and misunderstanding of terminology were possible.

CONCLUSION

Our single-center survey of ICU staff attitudes to deceased organ donation, carried out 3 times during 5 y, demonstrates persisting negativity to DCD despite increases in DCD donation numbers in our ICU. This contradicts the hypothesis that increasing familiarity by ICU staff with DCD will result in more favorable attitudes to donation. The study indicates overwhelming positivity to the SNOD role and broad support of DBD in our local ICU staff, but comparison of ICUs with different levels of donation activity may be of value. Staff education is recommended to improve familiarity and confidence in providing care at the point of withdrawal of life-sustaining treatment for DCD. We suggest that it may be beneficial to explore the emotional connotations of DCD that does not proceed and how expectations are framed for both families and staff.

In summary, although we found broadly positive perceptions of the benefits of deceased organ donation, negative attitudes centered on the timescale and complexity of the donation process in DCD. Measurement of staff attitudes to organ donation may allow targeted interventions that support staff and improve patient and family care through the organ donation process.

ACKNOWLEDGMENTS

The authors would like to recognize and thank the staff who contributed to help collect the surveys, those who completed the survey, the 3 panel assessors, and the donors and their families.