- Email: [email protected]
The Acceptability and Feasibility of Using Mortality Prediction Scores for Initiating End-of-Life Goals-of-Care Communication in the Adult Intensive Care Unit
Vol.
-
No.
- -
2019
Journal of Pain and Symptom Management
1
Brief Report
The Acceptability and Feasibility of Using Mortality Prediction Scores for Initiating End-of-Life Goals-of-Care Communication in the Adult Intensive Care Unit Shelly Orr, PhD, RN, CNE Virginia Commonwealth University School of Nursing, Richmond, Virginia, USA
Abstract Context. Uncertainties in prognosis remain a barrier to end-of-life (EOL) communication in the intensive care unit (ICU), thus strategies are needed for increasing the precision of prognosis and timeliness of EOL goals-of-care communication. Use of mortality prediction scores offers one approach to this issue. Objectives. This study evaluated the acceptability and feasibility of providers’ use of patient mortality prediction scores as part of routine practice to increase prognosis precision and timeliness of EOL communication as well as providers’ intentions to change practice related to EOL goals-of-care communication based on awareness of the scores. Methods. An explanatory mixed-methods approach was used to provide Sequential Organ Failure Assessment (SOFA) patient mortality prediction scores to ICU providers, who then completed an acceptability and feasibility questionnaire and participated in follow-up interviews conducted to further understand questionnaire responses and gain insight into their perceptions based on having SOFA scores. Results. Providers reported that using SOFA scores was acceptable and feasible, although there was some disagreement about effectiveness of SOFA scores for determining mortality risk. Providers with limited ICU experience were eager, and accepting of the scores while those with more experience found the scores to be an adjunct to their own intuition, although all acknowledged the benefit of looking at score trends. An important finding was the need to consider SOFA scores in relation to patient clinical context. Conclusion. Use of SOFA scores as a means to potentially increase EOL goals-of-care communication emerged as most beneficial and acceptable to providers with limited ICU experience. J Pain Symptom Manage 2019;-:-e-. Ó 2019 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved. Key Words Mortality risk, intensive care, end-of-life, goals of care, SOFA
Introduction Although health care teams recognize that profoundly ill patients in adult intensive care units (ICUs) may die, many families are caught by surprise when their loved one dies in a setting with the most advanced technology and intense care available. ICU deaths account for about 20% of patient deaths in U.S. hospitals, and this rate is increasing1 due in part to deficiencies in end-of-life (EOL) care
Address correspondence to: Shelly Orr, PhD, RN, CNE, Virginia Commonwealth University School of Nursing, 1100 East Leigh Street, Richmond, VA 23298-0567, USA. E-mail: [email protected] Ó 2019 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved.
communication that can compromise quality of EOL care2 and increase resource utilization.3,4 Previous studies suggest that EOL goals-of-care communication is infrequent among health care providers, patients, and families; often occurs late in the course of illness5,6; and relies on family members to act as patient surrogates in discussions.7 Prognostic uncertainty continues to be a source of distress for patients, their families, and health care providers.8 Increased severity of illness (SOI) scores
Accepted for publication: September 12, 2019.
0885-3924/$ - see front matter https://doi.org/10.1016/j.jpainsymman.2019.09.011
2
are associated with a significant increase in the relative risk of hospital death.9 Family meetings about EOL goals of care can improve family satisfaction with the EOL experience,10 but uncertainties in prognosis (e.g., SOI) are a barrier to EOL communication in the ICU.11 SOI mortality risk prediction scores are not routinely calculated, and there is little research examining their use for improving the timeliness of EOL goals-of-care communication in patients who are likely to die in the ICU. Multiple valid and reliable SOI scoring systems are available for predicting ICU mortality, but there is no consensus about how or when to use them in provider-patient/family communication about EOL. Moreover, evidence-based standards of care for EOL goals-of-care communication in adult ICUs do not exist, thus strategies for increasing the timeliness of discussions about EOL goals of care are needed.5,12 As an initial step to address this gap, the study reported here aimed to determine the acceptability and feasibility of using SOI mortality risk prediction scores for initiating EOL goals-of-care communication in the adult ICU.
Methods A mixed-methods design with multiple phases was used to address the aims of the study. The first phase was qualitative and consisted of a focus group for selection of an SOI scoring system. The second and third phases used a mixed-methods sequential explanatory design13 in which quantitative data were collected first followed by qualitative data. The rationale for using this mixed-methods approach was that it permitted the use of interviews to further explain and interpret findings from quantitative questionnaires. The study was completed in a medical ICU (MICU) at a large academic medical center. In this unit, two medical teams, each comprise an attending physician, a fellow physician, and a mix of interns, residents, acute care nurse practitioners (ACNPs), and physician assistants (PAs), provided patient care. The principal investigator (PI) had previously conducted research on this unit; therefore, the PI was familiar with the unit structure but had no direct oversight of staff. The study was approved by the relevant institutional review board. Written consent was obtained from all study participants for all phases of the study. In the first phase of the study, attending physicians, fellow physicians, ACNPs, and PAs working in the MICU were recruited for a focus group via electronic mail as they were deemed expert providers responsible for medical care of patients admitted to the unit. Based on an integrative review completed by the PI and feasibility of implementation in the study
Orr
Vol.
-
No.
- -
2019
setting, the PI presented four well-validated and reliable SOI scoring systems with corresponding references (Mortality Probability Model III,14e17 Acute Physiology and Chronic Health Evaluation IV,16,18 Sequential Organ Failure Assessment [SOFA],19e21 Simplified Acute Physiology Score III22,23) to the focus group participants in table format. After discussion among the participants about the perceived feasibility of using each SOI system, the PI requested that they reach consensus on a single system. They were also asked to complete a demographic form embedded in Research Electronic Data Capture (REDCap). The second phase of the study used a quantitative questionnaire to determine if participants could use SOI mortality prediction scores as part of their routine practice in the ICU and to evaluate their perceptions of acceptability and feasibility of using the scores. Questions examining acceptability were based on the knowledge that acceptability is a multifaceted construct that examines how well participants consider an intervention to be appropriate.24 For this study specifically, the PI assessed how well the scores were received and understood by the participants, how well they trusted the scores, and how appropriate they believed it was to use the scores in their daily practice. Questions examining feasibility were based on the knowledge that assessment of feasibility is helpful in determining whether an intervention is appropriate for further testing.25 For this study specifically, the PI assessed whether participants had enough time to incorporate the scores into their daily practice and initial information regarding whether receiving the scores was beneficial. All MICU fellows, residents, and intern physicians, as well as all ACNPs and PAs, were recruited on an ongoing basis via electronic mail and face to face for this portion of the study. Attending physicians were excluded because of the short length of their rotations in the ICU. The PI or research assistant (RA) calculated daily mortality risk percentages for MICU patients admitted under the care of study participants for 10 consecutive days, using the free web-based calculator available to the public. To ensure consistency and congruency with the chosen system’s published protocol, the PI developed a user’s manual for the PI and RA. Calculated scores were limited to three days per patient for study feasibility purposes. Laminated reference cards were given to providers during study enrollment for assistance with interpreting the scores. Mortality risk percentages were shared with participating providers on a card each morning before team rounding. After the 10-day period, participants received a link to a REDCap questionnaire asking about acceptability and feasibility of using mortality risk prediction scores as part of their workflow and practice.26 Results of the questionnaire were
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
3
Table 1 Topics for Focused Interviews 1. Tell me about any experience you have had with patients during their end of life. 2. What data, if any, do you currently use to guide your end-of-life care practice? 3. A few weeks ago, you were provided with Sequential Organ Failure Assessment (SOFA) mortality prediction scores for your patients. How did you feel about that? Specifically tell me about: a. what made you trust or distrust the score b. why you thought being provided with the score was a good thing or a bad thing c. what you did with the information d. how you think the scores could be incorporated into your daily routine as a provider e. any ways in which it may have changed your thinking about your patient’s prognosis 4. Tell me about any ways your practice may have changed regarding communication with patients and/or families about end-of-life goals-ofcare since your experience with using SOFA mortality prediction scores. If your practice has not changed, tell me about any intentions you have for changing (or not) based on the experience with the SOFA scores. 5. When caring for an ICU patient with a high risk for mortality, tell me about your perceived ability to impact their end-of-life experience?
retrieved from REDCap in the form of descriptive statistics. Participants for this portion of the study were also asked to complete a demographic form embedded in REDCap. Finally, to further explain the results of the quantitative portion of the study and to learn more about the potential impact of using SOI mortality risk prediction scores, all participants who received daily mortality risk percentages and completed the acceptability and feasibility REDCap questionnaire were asked to participate in a follow-up interview as the third phase of the study. Those agreeing to participate were scheduled for one-hour face-to-face interviews with the PI. Although specific topics were covered during the interview (Table 1), the PI permitted the interview to move freely from topic to topic by allowing the participant’s cues to determine the flow.27 Each interview was digitally voice recorded and transcribed verbatim. To ensure that a comprehensive analysis framework was followed, the Consolidated Criteria for Reporting Qualitative Research checklist for interviews was used.28 A qualitative descriptive approach was used to analyze the interview data29,30 with a fluid process wherein the PI reviewed transcripts after every two to three interviews, determined emerging themes, then confirmed them with subsequent participants. The resulting themes across all transcripts were identified along with exemplars until theoretical saturation was obtained. To ensure that the resulting themes were credible and finalize them, the PI discussed the findings with experts who were familiar with the subject under study. Finally, the final themes and exemplars were examined to help explain the results of the acceptability and feasibility questionnaire.
Results Participant Characteristics A total of 15 providers participated in the study (Table 2). The age of participants ranged from 26 to 51, with most being female (60%) and white (93%).
The sample had five ACNP/PAs and 10 physicians, who had variable years of practice experience both in the ICU setting and in EOL. For the purpose of analysis, ACNPs and PAs were grouped together to protect the confidentiality of the one participating PA, and although it is acknowledged that their disciplines are not interchangeable, they shared the same clinical responsibilities in the study site.
Phase 1dFocus Group Five providers (two ACNP/PAs, two attending physicians, and one fellow physician) participated in the focus group to select an SOI instrument. After discussion, the providers selected the SOFA as the most feasible for use given its free and easy online access, limited number of variables, ability to provide admission and daily scores, and increasing the presence in the setting (and therefore, recognition among providers) because of its role in the updated sepsis guidelines.31 SOFA assigns one to four points to the following organ systems depending on the level of organ dysfunction: circulatory, respiratory, renal, hematologic, hepatic, and central nervous system, and then the total score is translated into an estimated mortality percentage.19 Data required for the calculation are typically collected on ICU admission and daily thereafter throughout the ICU stay, with the most abnormal values for each day being used for scoring.21
Phase 2dAcceptability and Feasibility Questionnaire Two of the ACNP/PAs who participated in the focus group also participated in the second phase of the study, in addition to a new group of providers, consisting of three additional ACNP/PAs, four interns, two residents, and one fellow physician (total N ¼ 12). This group received SOFA scores for 10 days for patients admitted under their care (total score N ¼ 145). These scores came from a total of 70 patients and were calculated using the online calculator (http://clincalc.com/icumortality/sofa.aspx). There was an average of 2.1 daily scores provided per patient because some patients were transferred out of the unit
4
Orr
Table 2 Demographic Characteristics of Study Participants Across All Study Phases (n ¼ 15) Variable Gender Female Male Race American Indian or Alaska Native Asian Black or African American Native Hawaiian or Other Pacific Islander White Other Health profession ACNP or PA Physician Intern Resident Fellow Attending Years of practice in an ICU setting Less than one year One year Two years Three years Four years Greater than four years Previous experience with EOL None Personal (i.e., loss of someone close to participant) Professional Coursework on EOL care Hands-on experience with patients during their EOL
Frequency/Percent 9 (60) 6 (40) 0 1 (6.7) 0 0 14 (93.3) 0 5 10 4 2 2 2
(33.3) (66.7) (40) (20) (20) (20)
4 (26.7) 0 1 (6.7) 4 (26.7) 2 (13.3) 4 (26.7) 1 (6.7) 9 (60) 11 (73.3) 5 (45.5) 9 (81.8)
ACNP ¼ acute care nurse practitioner; PA ¼ physician assistant; ICU ¼ intensive care unit; EOL ¼ end of life.
or died and thus did not have a full three days of calculations. In addition, some providers who were simultaneously enrolled in the study cared for the same patients, which generated some concurrent score calculations. On average, participants were provided with scores for 2.6 patients per day. When reviewing a 25% random sample of each other’s daily calculations for inter-rater reliability purposes, there were two SOFA scores in which the PI and RA had conflicting calculations; these were resolved by a simultaneous review of the various SOFA variables. Questionnaire data (Fig. 1) revealed that overall, the participants found the use of SOFA scores for their patients to be acceptable and feasible as part of routine workflow and practice. However, there was some disagreement among participants, particularly related to the use of the scores as an effective way for determining patient mortality risk, which warranted further exploration.
Phase 3dFace-to-Face Interviews Seven of 12 participants who received daily SOFA scores participated in a follow-up interview. A summary of their responses is presented in Table 3,
Vol.
-
No.
- -
2019
organized according to main themes that corresponded with the purpose of this study, then divided into subthemes that were identified throughout the transcriptions. Resulting themes and corresponding exemplars helped to better understand results from the acceptability and feasibility questionnaire and to learn more about the potential impact of using SOI mortality risk prediction scores for timelier EOL goals-of-care communication.
Discussion Although the PI acknowledges her own perceptions regarding EOL communication and prognostication, the purposeful involvement of three additional nonEOL researchers as mentors throughout the study was done with the intent to minimize any bias in study design and analysis and support reflexivity. Regardless, it remains important to address all study limitations. The single ICU used for this study may limit the generalizability to other types of critical care units, including similar ICUs. In addition, the small nature of this study provides acceptability and feasibility information only and further limits generalizability. The PI also acknowledges that the single coder used to examine the interview data is an additional limitation. However, despite these limitations, this study has provided useful information and a foundation for future work. Increasing the timeliness of EOL goals-of-care communication in the adult ICU is warranted to ensure that care is in alignment with the wishes of the patient. In addition to the benefits for the patient, proactive communication reduces anxiety, depression, and post-traumatic stress disorder for family members whose loved one dies in the ICU.32 Thus, the use of mortality risk scores as a facilitator to such communication has the potential to improve care. However, the idea is not that a specific SOFA mortality risk prediction score would be used as a cut-off point for mandating goals-of-care communication; nor would that specific score likely be shared with the family. Instead, when used in context, can it heighten providers’ awareness, and therefore, prompt earlier EOL goals-of-care conversations? Implementation of future research using mortality risk scores in the ICU may fail if providers’ perceptions are not considered first.33 This small-scale study addressed this need, revealing that providers found SOFA feasible and acceptable for calculating mortality risk prediction scores, easy to use as part of the daily workflow, and widely accepted and trusted. Nonetheless, interview data indicated that context must be considered and SOFA scores alone should not be used to initiate EOL goals-of-care communication
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
5
I clearly understood what the SOFA mortality prediction score meant. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
8
I trust that the SOFA mortality prediction scores that were provided to me were accurate. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
I had enough time to incorporate the mortality prediction score into the plan of care for my patients. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
It is appropriate for me to know an accurate prediction of my patient’s mortality risk. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
Knowing my patient’s mortality prediction score made me think about the prognosis more. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
Using SOFA mortality prediction scores is an effective way for determining my patient’s mortality risk. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
I believe it would be beneficial to use the results of SOFA mortality prediction calculations on a daily basis. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
Fig. 1. Summary of acceptability and feasibility questionnaire findings. SOFA ¼ Sequential Organ Failure Assessment.
6
Table 3 Summary of Focused Interview Findings Main Theme Effects on clinical decision making
Exemplars
Context: The most substantial of all themes identified, participants reported the need to consider SOFA scores in relation to patient context. Respondents suggested that a number alone should not determine mortality risk and whether a goals-of-care conversation should occur, as there could be contextual issues (i.e., intubation simply for a procedure, not patient condition) related to the score being elevated Level of experience: Participants with limited ICU experience (less than one year), either personally or professionally, were eager and accepting of SOFA mortality risk prediction scores provided to them. According to their reports, this was due to the ability of SOFA scores to detect nuances that they did not always see in the clinical picture alone and because they experienced instances in which the SOFA scores they received were more accurate than their subjective assessment. In addition, those with less experience indicated that high mortality risk scores pushed them to have earlier goals-of-care conversations than they would have had if they had not been provided with the score. In contrast, participants with more ICU experience indicated that scores confirmed their own judgment or intuition. Because the scores matched their subjective assessments (i.e., trends in laboratories and vitals and previous experiences with similar diagnoses), they trusted the scores. However, some participants said they might distrust the scores if they differed drastically from their own assessments. Those with more experience voiced concerns that although those with less experience could benefit from having the scores, they might not consider the whole picture and would prematurely change their level of care or have goals-of-care conversations Trends: All participants spoke of the value of looking at trends in clinical data to examine the big picture because in their view, an admission SOFA score alone would not be as helpful because it was only a snapshot. Participants followed trends in daily scores as a way for determining whether treatment interventions were useful or not More than mortality prediction: Most participants also noted that there were benefits of using SOFA scores beyond just risk of mortality. For example, they used individual organ system scores within the SOFA calculation to identify specific areas in which additional intervention might or might not be beneficial. This theme is related to the theme of context in that the individual components of SOFA can provide information on which body system(s) may be causing an increase in mortality risk. In addition, two of the participants suggested using SOFA scores on the general inpatient units as a means to potentially prevent transfer to an ICU if it was not in alignment with the patient’s goals of care Approval of SOFA: All participants indicated they were at least somewhat familiar with SOFA. They indicated that SOFA could be
There was a patient, for example, that was kind of middle of the road, so probably around a 50% mortality risk, but they had a procedure done and they were intubated and were put on 100% on the ventilator. They were bronched or something
One time when I should have trusted it and did not, we had one patient who we had jump one day in his score and the only thing that really changed was that his bilirubin had gone up. I was like I don’t know, he was still pretty well and then of course that ended up not going very well It allowed me to take a moment and be like I think we should have that talk instead of waiting a day or two to see what happened kind of
Orr
There is so much noise from day to day and it’s like when you weigh yourself. You know, like I gained six pounds in a day. When I look at the [SOFA] trend, there are similar things
Maybe use SOFA even prior to ICU to see if escalation is actually warranted or not? You hang on to that person on the floor to the last moment when then they are like crashing and it’s like, we have got to get them down there right now or they are going to die in this moment. Whereas, maybe if you had someone that said look, they have been getting worse, maybe you could have coordinated that better
Vol. - -
If your clinician doesn’t put it into context with the whole clinical picture, or they don’t know the clinical picture, it could be very dangerous. Over-
No.
Feasibility and acceptability
Subthemes
2019
-
resuscitation, under-resuscitation, premature goals of care conversations, et cetera
Vol. No. - -
Being provided the score was a great thing because I did not feel like I had the time to calculate myself on my patients
2019
EOL care planning and practice
It was awesome
Alright, because we fix things. We don’t help them break slower or something, we try to fix everything and it’s just a sensitive subject. People come to us to be fixed, you know, I guess so, it’s just, it can be very tough
So, I don’t always like to use numbers because it gives them [family] that either false sense of security or e but if someone’s asking for a hard and fast number, at least then I can say, based on these scores, it changes from day to day, but today there’s a really high mortality
I think it was generally well received. The people [team] seemed open to it
Mortality Prediction for EOL Communication
used in calculating SOI scores throughout hospitalization rather than just at admission. However, they highlighted the importance of educating users on SOFA scores because they could fall into the hands of a less-experienced provider. They suggested that SOFA scores should be in the EHR in a format that required users to click on it to learn the score’s conversion to a mortality risk prediction percentage and individual system scores Time as a consideration: Many participants acknowledged the time required for someone to calculate SOFA scores and indicated because of this issue, they would not be able to perform the calculations daily. They suggested that SOFA scores should be autogenerated in the EHR Promising opportunity: Most participants revealed they were excited about their participation in the study, appreciated getting the scores each day, and looked forward to seeing how the results of the study might change current practice Fixing everyone: Most of the physician participants spoke of their desire to fix everyone, indicating that knowing that death was coming, they found it extremely difficult to move away from a curative approach; one participant said this was facing reality. In contrast, two ACNPs highlighted their years of bedside nursing as the likely key to their ability to move from curing to caring with those very sick patients early in the course of their admission. In general, participants with more experience indicated that they were already having conversations with families about poor prognosis on a consistent basis despite the difficulties. Although the scores gave them confidence in their assessment of the patient’s risk, they did not change the timeliness of goals-of-care conversations Communication with patient/family: Many participants verbalized the importance of sensitivity when discussing mortality risk with patients and/or their families. A few participants emphasized that numbers (i.e., mortality risk percentages) should not be shared with families as they can distract them and provide false reassurance. In addition, they indicated that communication with family about patient prognosis could positively affect the EOL care provided because it often provided the opportunity to discuss whether the patient’s wishes were being honored Collaboration with team: Many participants indicated that they discussed SOFA scores with the interdisciplinary team during daily rounds, which was well received and provided an opportunity to talk with peers about different treatment or care options that could influence the patient’s trajectory. In addition, many felt that having the scores should be a formal part of daily rounds Looking at the big picture: A couple of the participants suggested that providers’ focus should go beyond the ICU admission to include the patient’s quality of life even after the ICU stay. They highlighted the need to have goals-of-care conversations with all patients with life-limiting disease regardless of whether their risk for mortality during hospitalization was low
There was another one, this was a couple of weeks ago, not cancer but similar. Lots of co-morbidities coming into this. The score was pretty low. I think it was like 10 or 15 percent. I was like, I don’t know. I mean, yes, this hospitalization, sure, fine, but the next one and the next one and the next one and the six before this, she’s on a bad path
SOFA ¼ Sequential Organ Failure Assessment; ICU ¼ intensive care unit; SOI ¼ severity of illness; EHR ¼ electronic health record; ACNPs ¼ acute care nurse practitioners; EOL ¼ end of life.
7
8
without determining what contributed to the score. Therefore, careful consideration must be given to ensure providers using the scores are educated on how it is calculated and interpreted and its limitations. Integration of SOFA scoring into the electronic health record would enhance the feasibility of its use in clinical practice as otherwise providers would have limited time to calculate the scores. The use of SOFA mortality risk scores for potentially increasing EOL goals-of-care communication emerged as most beneficial for providers with limited ICU experience, possibly as a result of their lack of clinical intuition. Because more experienced providers indicated that the scores only provided confirmation of their subjective assessments of patient mortality risk, the use of scores may not affect the probability of earlier goals-of-care communication. Nonetheless, experienced providers were initiating early conversations, perhaps due in part to the acuity of illness and frequent EOL occurrences in the MICU, and should be commended. The discrepancy in how inexperienced vs. experienced providers viewed the use of SOFA scores, and the limitations of this study, warrants further research to systematically investigate the effects of incorporating mortality risk scores into everyday clinical evaluation on subsequent EOL outcomes. Because using mortality risk prediction scores to prompt timelier EOL goals-of-care communication is only a single action in a complex clinical picture, other strategies that assist providers in effectively communicating with patients and/or their families during these difficult times are of utmost importance and should be included in future research as well.
Orr
Vol.
-
No.
- -
2019
care in the intensive care unit. Crit Care Med 2008;36: 1138e1146. 3. Quenot JP, Rigaud JP, Prin S, et al. Impact of an intensive communication strategy on end-of-life practices in the intensive care unit. Intensive Care Med 2012;38:145e152. 4. Curtis JR, Rubenfeld GD. Improving palliative care for patients in the intensive care unit. J Palliat Med 2005;8: 840e854. 5. Curtis JR, Engelberg RA, Bensink ME, Ramsey SD. Endof-life care in the intensive care unit: can we simultaneously increase quality and reduce costs? Am J Respir Crit Care Med 2012;186:587e592. 6. McNeely PD, H ebert PC, Dales RE, et al. Deciding about mechanical ventilation in end-stage chronic obstructive pulmonary disease: how respirologists perceive their role. CMAJ 1997;156:177e183. 7. Curtis JR, White DB. Practical guidance for evidencebased ICU family conferences. Chest 2008;134:835e843. 8. Etkind SN, Koffman J. Approaches to managing uncertainty in people with life-limiting conditions: role of communication and palliative care. Postgrad Med J 2016;92: 412e417. 9. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100: 1619e1636. 10. Kamel G, Paniagua M, Uppalapati A. Palliative care in the intensive care unit: are residents well trained to provide optimal care to critically ill patients? Am J Hosp Palliat Care 2015;32:758e762. 11. Billings JA. The end-of-life family meeting in intensive care. Part I: indications, outcomes, and family needs. J Palliat Med 2011;14:1042e1050. 12. Curtis JR. Palliative and end-of-life care for patients with severe COPD. Eur Respir J 2008;32:796e803. 13. Creswell JW, Plano Clark VL. Designing and conducting mixed methods research, 2nd ed. Thousand Oaks: Sage Publications, 2011.
Disclosures and Acknowledgments The author acknowledges the input of Elaine Amella, PhD, RN, FGSA, FAAN, Martina Mueller, PhD, Susan Newman, PhD, RN, CRRN, and Clareen Wiencek, PhD, RN, CNP, ACHPN, into the design and conduct of the study. The author also thanks Deborah B. McGuire, PhD, RN, FAAN, for her helpful review and editorial input. This work was supported by Sigma Theta Tau InternationaleeGamma Omega Chapter (Nursing Science and Practice grant). The author declares no conflicts of interest.
References 1. Zomorodi M, Bowen G. Value-behavior congruency when providing end-of-life care in the intensive care unit. J Hosp Palliat Nurs 2010;12:295e302. 2. Glavan BJ, Engelberg RA, Downey L, Curtis JR. Using the medical record to evaluate the quality of end-of-life
14. Higgins TL, Teres D, Copes WS, et al. Assessing contemporary intensive care unit outcome: an updated Mortality Probability Admission Model (MPM0-III). Crit Care Med 2007;35:827e835. 15. Higgins TL, Kramer AA, Nathanson BH, et al. Prospective validation of the intensive care unit admission Mortality Probability Model (MPM0-III). Crit Care Med 2009;37: 1619e1623. 16. Kuzniewicz MW, Vasilevskis EE, Lane R, et al. Variation in ICU risk-adjusted mortality: impact of methods of assessment and potential confounders. Chest 2008;133: 1319e1327. 17. Kramer AA, Higgins TL, Zimmerman JE. Comparison of the mortality probability admission model III, National Quality Forum, and acute physiology and chronic health evaluation IV hospital mortality models: implications for national benchmarking. Crit Care Med 2014;42:544e553. 18. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today’s critically ill patients. Crit Care Med 2006;34:1297e1310.
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
9
19. Minne L, Ludikhuize J, De Jonge E, De Rooij S, AbuHanna A. Prognostic models for predicting mortality in elderly ICU patients: a systematic review. Intensive Care Med 2011;37:1258e1268.
25. Bowen DJ, Kreuter M, Spring B, et al. How we design feasibility studies. Am J Prev Med 2009;36:452e457.
20. Minne L, Abu-Hanna A, de Jonge E. Evaluation of SOFA-based models for predicting mortality in the ICU: a systematic review. Crit Care 2008;12:R161.
27. Waltz CF, Strickland OL, Lenz ER. Measurement in Nursing and Health Research, 2nd ed. New York: Springer Publishing Company, 2010.
21. Vincent JL, de Mendonc¸a A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on ‘‘sepsis-related problems’’ of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793e1800.
28. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care 2007;19:349e357.
22. Metnitz PGH, Moreno RP, Almeida E, et al. SAPS 3dFrom evaluation of the patient to evaluation of the intensive care unit. Part 1: objectives, methods and cohort description. Intensive Care Med 2005;31:1336e1344. 23. Moreno RP, Metnitz PGH, Almeida E, et al. SAPS 3dFrom evaluation of the patient to evaluation of the intensive care unit. Part 2: development of a prognostic model for hospital mortality at ICU admission. Intensive Care Med 2005;31:1345e1355. 24. Sekhon M, Cartwright M, Francis JJ. Acceptability of healthcare interventions: an overview of reviews and development of a theoretical framework. BMC Health Serv Res 2017;17:88.
26. Tickle-Degnen L. Nuts and bolts of conducting feasibility studies. Am J Occup Ther 2013;67:171e176.
29. Hsieh H-F, Shannon SE. Three approaches to qualitative content analysis. Qual Health Res 2005;15:1277e1288. 30. Sandelowski M. What’s in a name? Qualitative description revisited. Res Nurs Health 2010;33:77e84. 31. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801e810. 32. Lautrette A, Darmon M, Megarbane B, et al. A communication strategy and brochure for relatives of patients dying in the ICU. N Engl J Med 2007;356:469e478. 33. Ayala GX, Elder JP. Qualitative methods to ensure acceptability of behavioral and social interventions to the target population. J Public Health Dent 2011;71(Suppl 1): S69eS79.
-
No.
- -
2019
Journal of Pain and Symptom Management
1
Brief Report
The Acceptability and Feasibility of Using Mortality Prediction Scores for Initiating End-of-Life Goals-of-Care Communication in the Adult Intensive Care Unit Shelly Orr, PhD, RN, CNE Virginia Commonwealth University School of Nursing, Richmond, Virginia, USA
Abstract Context. Uncertainties in prognosis remain a barrier to end-of-life (EOL) communication in the intensive care unit (ICU), thus strategies are needed for increasing the precision of prognosis and timeliness of EOL goals-of-care communication. Use of mortality prediction scores offers one approach to this issue. Objectives. This study evaluated the acceptability and feasibility of providers’ use of patient mortality prediction scores as part of routine practice to increase prognosis precision and timeliness of EOL communication as well as providers’ intentions to change practice related to EOL goals-of-care communication based on awareness of the scores. Methods. An explanatory mixed-methods approach was used to provide Sequential Organ Failure Assessment (SOFA) patient mortality prediction scores to ICU providers, who then completed an acceptability and feasibility questionnaire and participated in follow-up interviews conducted to further understand questionnaire responses and gain insight into their perceptions based on having SOFA scores. Results. Providers reported that using SOFA scores was acceptable and feasible, although there was some disagreement about effectiveness of SOFA scores for determining mortality risk. Providers with limited ICU experience were eager, and accepting of the scores while those with more experience found the scores to be an adjunct to their own intuition, although all acknowledged the benefit of looking at score trends. An important finding was the need to consider SOFA scores in relation to patient clinical context. Conclusion. Use of SOFA scores as a means to potentially increase EOL goals-of-care communication emerged as most beneficial and acceptable to providers with limited ICU experience. J Pain Symptom Manage 2019;-:-e-. Ó 2019 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved. Key Words Mortality risk, intensive care, end-of-life, goals of care, SOFA
Introduction Although health care teams recognize that profoundly ill patients in adult intensive care units (ICUs) may die, many families are caught by surprise when their loved one dies in a setting with the most advanced technology and intense care available. ICU deaths account for about 20% of patient deaths in U.S. hospitals, and this rate is increasing1 due in part to deficiencies in end-of-life (EOL) care
Address correspondence to: Shelly Orr, PhD, RN, CNE, Virginia Commonwealth University School of Nursing, 1100 East Leigh Street, Richmond, VA 23298-0567, USA. E-mail: [email protected] Ó 2019 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved.
communication that can compromise quality of EOL care2 and increase resource utilization.3,4 Previous studies suggest that EOL goals-of-care communication is infrequent among health care providers, patients, and families; often occurs late in the course of illness5,6; and relies on family members to act as patient surrogates in discussions.7 Prognostic uncertainty continues to be a source of distress for patients, their families, and health care providers.8 Increased severity of illness (SOI) scores
Accepted for publication: September 12, 2019.
0885-3924/$ - see front matter https://doi.org/10.1016/j.jpainsymman.2019.09.011
2
are associated with a significant increase in the relative risk of hospital death.9 Family meetings about EOL goals of care can improve family satisfaction with the EOL experience,10 but uncertainties in prognosis (e.g., SOI) are a barrier to EOL communication in the ICU.11 SOI mortality risk prediction scores are not routinely calculated, and there is little research examining their use for improving the timeliness of EOL goals-of-care communication in patients who are likely to die in the ICU. Multiple valid and reliable SOI scoring systems are available for predicting ICU mortality, but there is no consensus about how or when to use them in provider-patient/family communication about EOL. Moreover, evidence-based standards of care for EOL goals-of-care communication in adult ICUs do not exist, thus strategies for increasing the timeliness of discussions about EOL goals of care are needed.5,12 As an initial step to address this gap, the study reported here aimed to determine the acceptability and feasibility of using SOI mortality risk prediction scores for initiating EOL goals-of-care communication in the adult ICU.
Methods A mixed-methods design with multiple phases was used to address the aims of the study. The first phase was qualitative and consisted of a focus group for selection of an SOI scoring system. The second and third phases used a mixed-methods sequential explanatory design13 in which quantitative data were collected first followed by qualitative data. The rationale for using this mixed-methods approach was that it permitted the use of interviews to further explain and interpret findings from quantitative questionnaires. The study was completed in a medical ICU (MICU) at a large academic medical center. In this unit, two medical teams, each comprise an attending physician, a fellow physician, and a mix of interns, residents, acute care nurse practitioners (ACNPs), and physician assistants (PAs), provided patient care. The principal investigator (PI) had previously conducted research on this unit; therefore, the PI was familiar with the unit structure but had no direct oversight of staff. The study was approved by the relevant institutional review board. Written consent was obtained from all study participants for all phases of the study. In the first phase of the study, attending physicians, fellow physicians, ACNPs, and PAs working in the MICU were recruited for a focus group via electronic mail as they were deemed expert providers responsible for medical care of patients admitted to the unit. Based on an integrative review completed by the PI and feasibility of implementation in the study
Orr
Vol.
-
No.
- -
2019
setting, the PI presented four well-validated and reliable SOI scoring systems with corresponding references (Mortality Probability Model III,14e17 Acute Physiology and Chronic Health Evaluation IV,16,18 Sequential Organ Failure Assessment [SOFA],19e21 Simplified Acute Physiology Score III22,23) to the focus group participants in table format. After discussion among the participants about the perceived feasibility of using each SOI system, the PI requested that they reach consensus on a single system. They were also asked to complete a demographic form embedded in Research Electronic Data Capture (REDCap). The second phase of the study used a quantitative questionnaire to determine if participants could use SOI mortality prediction scores as part of their routine practice in the ICU and to evaluate their perceptions of acceptability and feasibility of using the scores. Questions examining acceptability were based on the knowledge that acceptability is a multifaceted construct that examines how well participants consider an intervention to be appropriate.24 For this study specifically, the PI assessed how well the scores were received and understood by the participants, how well they trusted the scores, and how appropriate they believed it was to use the scores in their daily practice. Questions examining feasibility were based on the knowledge that assessment of feasibility is helpful in determining whether an intervention is appropriate for further testing.25 For this study specifically, the PI assessed whether participants had enough time to incorporate the scores into their daily practice and initial information regarding whether receiving the scores was beneficial. All MICU fellows, residents, and intern physicians, as well as all ACNPs and PAs, were recruited on an ongoing basis via electronic mail and face to face for this portion of the study. Attending physicians were excluded because of the short length of their rotations in the ICU. The PI or research assistant (RA) calculated daily mortality risk percentages for MICU patients admitted under the care of study participants for 10 consecutive days, using the free web-based calculator available to the public. To ensure consistency and congruency with the chosen system’s published protocol, the PI developed a user’s manual for the PI and RA. Calculated scores were limited to three days per patient for study feasibility purposes. Laminated reference cards were given to providers during study enrollment for assistance with interpreting the scores. Mortality risk percentages were shared with participating providers on a card each morning before team rounding. After the 10-day period, participants received a link to a REDCap questionnaire asking about acceptability and feasibility of using mortality risk prediction scores as part of their workflow and practice.26 Results of the questionnaire were
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
3
Table 1 Topics for Focused Interviews 1. Tell me about any experience you have had with patients during their end of life. 2. What data, if any, do you currently use to guide your end-of-life care practice? 3. A few weeks ago, you were provided with Sequential Organ Failure Assessment (SOFA) mortality prediction scores for your patients. How did you feel about that? Specifically tell me about: a. what made you trust or distrust the score b. why you thought being provided with the score was a good thing or a bad thing c. what you did with the information d. how you think the scores could be incorporated into your daily routine as a provider e. any ways in which it may have changed your thinking about your patient’s prognosis 4. Tell me about any ways your practice may have changed regarding communication with patients and/or families about end-of-life goals-ofcare since your experience with using SOFA mortality prediction scores. If your practice has not changed, tell me about any intentions you have for changing (or not) based on the experience with the SOFA scores. 5. When caring for an ICU patient with a high risk for mortality, tell me about your perceived ability to impact their end-of-life experience?
retrieved from REDCap in the form of descriptive statistics. Participants for this portion of the study were also asked to complete a demographic form embedded in REDCap. Finally, to further explain the results of the quantitative portion of the study and to learn more about the potential impact of using SOI mortality risk prediction scores, all participants who received daily mortality risk percentages and completed the acceptability and feasibility REDCap questionnaire were asked to participate in a follow-up interview as the third phase of the study. Those agreeing to participate were scheduled for one-hour face-to-face interviews with the PI. Although specific topics were covered during the interview (Table 1), the PI permitted the interview to move freely from topic to topic by allowing the participant’s cues to determine the flow.27 Each interview was digitally voice recorded and transcribed verbatim. To ensure that a comprehensive analysis framework was followed, the Consolidated Criteria for Reporting Qualitative Research checklist for interviews was used.28 A qualitative descriptive approach was used to analyze the interview data29,30 with a fluid process wherein the PI reviewed transcripts after every two to three interviews, determined emerging themes, then confirmed them with subsequent participants. The resulting themes across all transcripts were identified along with exemplars until theoretical saturation was obtained. To ensure that the resulting themes were credible and finalize them, the PI discussed the findings with experts who were familiar with the subject under study. Finally, the final themes and exemplars were examined to help explain the results of the acceptability and feasibility questionnaire.
Results Participant Characteristics A total of 15 providers participated in the study (Table 2). The age of participants ranged from 26 to 51, with most being female (60%) and white (93%).
The sample had five ACNP/PAs and 10 physicians, who had variable years of practice experience both in the ICU setting and in EOL. For the purpose of analysis, ACNPs and PAs were grouped together to protect the confidentiality of the one participating PA, and although it is acknowledged that their disciplines are not interchangeable, they shared the same clinical responsibilities in the study site.
Phase 1dFocus Group Five providers (two ACNP/PAs, two attending physicians, and one fellow physician) participated in the focus group to select an SOI instrument. After discussion, the providers selected the SOFA as the most feasible for use given its free and easy online access, limited number of variables, ability to provide admission and daily scores, and increasing the presence in the setting (and therefore, recognition among providers) because of its role in the updated sepsis guidelines.31 SOFA assigns one to four points to the following organ systems depending on the level of organ dysfunction: circulatory, respiratory, renal, hematologic, hepatic, and central nervous system, and then the total score is translated into an estimated mortality percentage.19 Data required for the calculation are typically collected on ICU admission and daily thereafter throughout the ICU stay, with the most abnormal values for each day being used for scoring.21
Phase 2dAcceptability and Feasibility Questionnaire Two of the ACNP/PAs who participated in the focus group also participated in the second phase of the study, in addition to a new group of providers, consisting of three additional ACNP/PAs, four interns, two residents, and one fellow physician (total N ¼ 12). This group received SOFA scores for 10 days for patients admitted under their care (total score N ¼ 145). These scores came from a total of 70 patients and were calculated using the online calculator (http://clincalc.com/icumortality/sofa.aspx). There was an average of 2.1 daily scores provided per patient because some patients were transferred out of the unit
4
Orr
Table 2 Demographic Characteristics of Study Participants Across All Study Phases (n ¼ 15) Variable Gender Female Male Race American Indian or Alaska Native Asian Black or African American Native Hawaiian or Other Pacific Islander White Other Health profession ACNP or PA Physician Intern Resident Fellow Attending Years of practice in an ICU setting Less than one year One year Two years Three years Four years Greater than four years Previous experience with EOL None Personal (i.e., loss of someone close to participant) Professional Coursework on EOL care Hands-on experience with patients during their EOL
Frequency/Percent 9 (60) 6 (40) 0 1 (6.7) 0 0 14 (93.3) 0 5 10 4 2 2 2
(33.3) (66.7) (40) (20) (20) (20)
4 (26.7) 0 1 (6.7) 4 (26.7) 2 (13.3) 4 (26.7) 1 (6.7) 9 (60) 11 (73.3) 5 (45.5) 9 (81.8)
ACNP ¼ acute care nurse practitioner; PA ¼ physician assistant; ICU ¼ intensive care unit; EOL ¼ end of life.
or died and thus did not have a full three days of calculations. In addition, some providers who were simultaneously enrolled in the study cared for the same patients, which generated some concurrent score calculations. On average, participants were provided with scores for 2.6 patients per day. When reviewing a 25% random sample of each other’s daily calculations for inter-rater reliability purposes, there were two SOFA scores in which the PI and RA had conflicting calculations; these were resolved by a simultaneous review of the various SOFA variables. Questionnaire data (Fig. 1) revealed that overall, the participants found the use of SOFA scores for their patients to be acceptable and feasible as part of routine workflow and practice. However, there was some disagreement among participants, particularly related to the use of the scores as an effective way for determining patient mortality risk, which warranted further exploration.
Phase 3dFace-to-Face Interviews Seven of 12 participants who received daily SOFA scores participated in a follow-up interview. A summary of their responses is presented in Table 3,
Vol.
-
No.
- -
2019
organized according to main themes that corresponded with the purpose of this study, then divided into subthemes that were identified throughout the transcriptions. Resulting themes and corresponding exemplars helped to better understand results from the acceptability and feasibility questionnaire and to learn more about the potential impact of using SOI mortality risk prediction scores for timelier EOL goals-of-care communication.
Discussion Although the PI acknowledges her own perceptions regarding EOL communication and prognostication, the purposeful involvement of three additional nonEOL researchers as mentors throughout the study was done with the intent to minimize any bias in study design and analysis and support reflexivity. Regardless, it remains important to address all study limitations. The single ICU used for this study may limit the generalizability to other types of critical care units, including similar ICUs. In addition, the small nature of this study provides acceptability and feasibility information only and further limits generalizability. The PI also acknowledges that the single coder used to examine the interview data is an additional limitation. However, despite these limitations, this study has provided useful information and a foundation for future work. Increasing the timeliness of EOL goals-of-care communication in the adult ICU is warranted to ensure that care is in alignment with the wishes of the patient. In addition to the benefits for the patient, proactive communication reduces anxiety, depression, and post-traumatic stress disorder for family members whose loved one dies in the ICU.32 Thus, the use of mortality risk scores as a facilitator to such communication has the potential to improve care. However, the idea is not that a specific SOFA mortality risk prediction score would be used as a cut-off point for mandating goals-of-care communication; nor would that specific score likely be shared with the family. Instead, when used in context, can it heighten providers’ awareness, and therefore, prompt earlier EOL goals-of-care conversations? Implementation of future research using mortality risk scores in the ICU may fail if providers’ perceptions are not considered first.33 This small-scale study addressed this need, revealing that providers found SOFA feasible and acceptable for calculating mortality risk prediction scores, easy to use as part of the daily workflow, and widely accepted and trusted. Nonetheless, interview data indicated that context must be considered and SOFA scores alone should not be used to initiate EOL goals-of-care communication
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
5
I clearly understood what the SOFA mortality prediction score meant. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
8
I trust that the SOFA mortality prediction scores that were provided to me were accurate. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
I had enough time to incorporate the mortality prediction score into the plan of care for my patients. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
It is appropriate for me to know an accurate prediction of my patient’s mortality risk. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
7
Knowing my patient’s mortality prediction score made me think about the prognosis more. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
Using SOFA mortality prediction scores is an effective way for determining my patient’s mortality risk. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
I believe it would be beneficial to use the results of SOFA mortality prediction calculations on a daily basis. Strongly Agree Agree Neither Agree or Disagree Disagree Strongly Disagree 0
1
2
3
4
5
6
Fig. 1. Summary of acceptability and feasibility questionnaire findings. SOFA ¼ Sequential Organ Failure Assessment.
6
Table 3 Summary of Focused Interview Findings Main Theme Effects on clinical decision making
Exemplars
Context: The most substantial of all themes identified, participants reported the need to consider SOFA scores in relation to patient context. Respondents suggested that a number alone should not determine mortality risk and whether a goals-of-care conversation should occur, as there could be contextual issues (i.e., intubation simply for a procedure, not patient condition) related to the score being elevated Level of experience: Participants with limited ICU experience (less than one year), either personally or professionally, were eager and accepting of SOFA mortality risk prediction scores provided to them. According to their reports, this was due to the ability of SOFA scores to detect nuances that they did not always see in the clinical picture alone and because they experienced instances in which the SOFA scores they received were more accurate than their subjective assessment. In addition, those with less experience indicated that high mortality risk scores pushed them to have earlier goals-of-care conversations than they would have had if they had not been provided with the score. In contrast, participants with more ICU experience indicated that scores confirmed their own judgment or intuition. Because the scores matched their subjective assessments (i.e., trends in laboratories and vitals and previous experiences with similar diagnoses), they trusted the scores. However, some participants said they might distrust the scores if they differed drastically from their own assessments. Those with more experience voiced concerns that although those with less experience could benefit from having the scores, they might not consider the whole picture and would prematurely change their level of care or have goals-of-care conversations Trends: All participants spoke of the value of looking at trends in clinical data to examine the big picture because in their view, an admission SOFA score alone would not be as helpful because it was only a snapshot. Participants followed trends in daily scores as a way for determining whether treatment interventions were useful or not More than mortality prediction: Most participants also noted that there were benefits of using SOFA scores beyond just risk of mortality. For example, they used individual organ system scores within the SOFA calculation to identify specific areas in which additional intervention might or might not be beneficial. This theme is related to the theme of context in that the individual components of SOFA can provide information on which body system(s) may be causing an increase in mortality risk. In addition, two of the participants suggested using SOFA scores on the general inpatient units as a means to potentially prevent transfer to an ICU if it was not in alignment with the patient’s goals of care Approval of SOFA: All participants indicated they were at least somewhat familiar with SOFA. They indicated that SOFA could be
There was a patient, for example, that was kind of middle of the road, so probably around a 50% mortality risk, but they had a procedure done and they were intubated and were put on 100% on the ventilator. They were bronched or something
One time when I should have trusted it and did not, we had one patient who we had jump one day in his score and the only thing that really changed was that his bilirubin had gone up. I was like I don’t know, he was still pretty well and then of course that ended up not going very well It allowed me to take a moment and be like I think we should have that talk instead of waiting a day or two to see what happened kind of
Orr
There is so much noise from day to day and it’s like when you weigh yourself. You know, like I gained six pounds in a day. When I look at the [SOFA] trend, there are similar things
Maybe use SOFA even prior to ICU to see if escalation is actually warranted or not? You hang on to that person on the floor to the last moment when then they are like crashing and it’s like, we have got to get them down there right now or they are going to die in this moment. Whereas, maybe if you had someone that said look, they have been getting worse, maybe you could have coordinated that better
Vol. - -
If your clinician doesn’t put it into context with the whole clinical picture, or they don’t know the clinical picture, it could be very dangerous. Over-
No.
Feasibility and acceptability
Subthemes
2019
-
resuscitation, under-resuscitation, premature goals of care conversations, et cetera
Vol. No. - -
Being provided the score was a great thing because I did not feel like I had the time to calculate myself on my patients
2019
EOL care planning and practice
It was awesome
Alright, because we fix things. We don’t help them break slower or something, we try to fix everything and it’s just a sensitive subject. People come to us to be fixed, you know, I guess so, it’s just, it can be very tough
So, I don’t always like to use numbers because it gives them [family] that either false sense of security or e but if someone’s asking for a hard and fast number, at least then I can say, based on these scores, it changes from day to day, but today there’s a really high mortality
I think it was generally well received. The people [team] seemed open to it
Mortality Prediction for EOL Communication
used in calculating SOI scores throughout hospitalization rather than just at admission. However, they highlighted the importance of educating users on SOFA scores because they could fall into the hands of a less-experienced provider. They suggested that SOFA scores should be in the EHR in a format that required users to click on it to learn the score’s conversion to a mortality risk prediction percentage and individual system scores Time as a consideration: Many participants acknowledged the time required for someone to calculate SOFA scores and indicated because of this issue, they would not be able to perform the calculations daily. They suggested that SOFA scores should be autogenerated in the EHR Promising opportunity: Most participants revealed they were excited about their participation in the study, appreciated getting the scores each day, and looked forward to seeing how the results of the study might change current practice Fixing everyone: Most of the physician participants spoke of their desire to fix everyone, indicating that knowing that death was coming, they found it extremely difficult to move away from a curative approach; one participant said this was facing reality. In contrast, two ACNPs highlighted their years of bedside nursing as the likely key to their ability to move from curing to caring with those very sick patients early in the course of their admission. In general, participants with more experience indicated that they were already having conversations with families about poor prognosis on a consistent basis despite the difficulties. Although the scores gave them confidence in their assessment of the patient’s risk, they did not change the timeliness of goals-of-care conversations Communication with patient/family: Many participants verbalized the importance of sensitivity when discussing mortality risk with patients and/or their families. A few participants emphasized that numbers (i.e., mortality risk percentages) should not be shared with families as they can distract them and provide false reassurance. In addition, they indicated that communication with family about patient prognosis could positively affect the EOL care provided because it often provided the opportunity to discuss whether the patient’s wishes were being honored Collaboration with team: Many participants indicated that they discussed SOFA scores with the interdisciplinary team during daily rounds, which was well received and provided an opportunity to talk with peers about different treatment or care options that could influence the patient’s trajectory. In addition, many felt that having the scores should be a formal part of daily rounds Looking at the big picture: A couple of the participants suggested that providers’ focus should go beyond the ICU admission to include the patient’s quality of life even after the ICU stay. They highlighted the need to have goals-of-care conversations with all patients with life-limiting disease regardless of whether their risk for mortality during hospitalization was low
There was another one, this was a couple of weeks ago, not cancer but similar. Lots of co-morbidities coming into this. The score was pretty low. I think it was like 10 or 15 percent. I was like, I don’t know. I mean, yes, this hospitalization, sure, fine, but the next one and the next one and the next one and the six before this, she’s on a bad path
SOFA ¼ Sequential Organ Failure Assessment; ICU ¼ intensive care unit; SOI ¼ severity of illness; EHR ¼ electronic health record; ACNPs ¼ acute care nurse practitioners; EOL ¼ end of life.
7
8
without determining what contributed to the score. Therefore, careful consideration must be given to ensure providers using the scores are educated on how it is calculated and interpreted and its limitations. Integration of SOFA scoring into the electronic health record would enhance the feasibility of its use in clinical practice as otherwise providers would have limited time to calculate the scores. The use of SOFA mortality risk scores for potentially increasing EOL goals-of-care communication emerged as most beneficial for providers with limited ICU experience, possibly as a result of their lack of clinical intuition. Because more experienced providers indicated that the scores only provided confirmation of their subjective assessments of patient mortality risk, the use of scores may not affect the probability of earlier goals-of-care communication. Nonetheless, experienced providers were initiating early conversations, perhaps due in part to the acuity of illness and frequent EOL occurrences in the MICU, and should be commended. The discrepancy in how inexperienced vs. experienced providers viewed the use of SOFA scores, and the limitations of this study, warrants further research to systematically investigate the effects of incorporating mortality risk scores into everyday clinical evaluation on subsequent EOL outcomes. Because using mortality risk prediction scores to prompt timelier EOL goals-of-care communication is only a single action in a complex clinical picture, other strategies that assist providers in effectively communicating with patients and/or their families during these difficult times are of utmost importance and should be included in future research as well.
Orr
Vol.
-
No.
- -
2019
care in the intensive care unit. Crit Care Med 2008;36: 1138e1146. 3. Quenot JP, Rigaud JP, Prin S, et al. Impact of an intensive communication strategy on end-of-life practices in the intensive care unit. Intensive Care Med 2012;38:145e152. 4. Curtis JR, Rubenfeld GD. Improving palliative care for patients in the intensive care unit. J Palliat Med 2005;8: 840e854. 5. Curtis JR, Engelberg RA, Bensink ME, Ramsey SD. Endof-life care in the intensive care unit: can we simultaneously increase quality and reduce costs? Am J Respir Crit Care Med 2012;186:587e592. 6. McNeely PD, H ebert PC, Dales RE, et al. Deciding about mechanical ventilation in end-stage chronic obstructive pulmonary disease: how respirologists perceive their role. CMAJ 1997;156:177e183. 7. Curtis JR, White DB. Practical guidance for evidencebased ICU family conferences. Chest 2008;134:835e843. 8. Etkind SN, Koffman J. Approaches to managing uncertainty in people with life-limiting conditions: role of communication and palliative care. Postgrad Med J 2016;92: 412e417. 9. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100: 1619e1636. 10. Kamel G, Paniagua M, Uppalapati A. Palliative care in the intensive care unit: are residents well trained to provide optimal care to critically ill patients? Am J Hosp Palliat Care 2015;32:758e762. 11. Billings JA. The end-of-life family meeting in intensive care. Part I: indications, outcomes, and family needs. J Palliat Med 2011;14:1042e1050. 12. Curtis JR. Palliative and end-of-life care for patients with severe COPD. Eur Respir J 2008;32:796e803. 13. Creswell JW, Plano Clark VL. Designing and conducting mixed methods research, 2nd ed. Thousand Oaks: Sage Publications, 2011.
Disclosures and Acknowledgments The author acknowledges the input of Elaine Amella, PhD, RN, FGSA, FAAN, Martina Mueller, PhD, Susan Newman, PhD, RN, CRRN, and Clareen Wiencek, PhD, RN, CNP, ACHPN, into the design and conduct of the study. The author also thanks Deborah B. McGuire, PhD, RN, FAAN, for her helpful review and editorial input. This work was supported by Sigma Theta Tau InternationaleeGamma Omega Chapter (Nursing Science and Practice grant). The author declares no conflicts of interest.
References 1. Zomorodi M, Bowen G. Value-behavior congruency when providing end-of-life care in the intensive care unit. J Hosp Palliat Nurs 2010;12:295e302. 2. Glavan BJ, Engelberg RA, Downey L, Curtis JR. Using the medical record to evaluate the quality of end-of-life
14. Higgins TL, Teres D, Copes WS, et al. Assessing contemporary intensive care unit outcome: an updated Mortality Probability Admission Model (MPM0-III). Crit Care Med 2007;35:827e835. 15. Higgins TL, Kramer AA, Nathanson BH, et al. Prospective validation of the intensive care unit admission Mortality Probability Model (MPM0-III). Crit Care Med 2009;37: 1619e1623. 16. Kuzniewicz MW, Vasilevskis EE, Lane R, et al. Variation in ICU risk-adjusted mortality: impact of methods of assessment and potential confounders. Chest 2008;133: 1319e1327. 17. Kramer AA, Higgins TL, Zimmerman JE. Comparison of the mortality probability admission model III, National Quality Forum, and acute physiology and chronic health evaluation IV hospital mortality models: implications for national benchmarking. Crit Care Med 2014;42:544e553. 18. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today’s critically ill patients. Crit Care Med 2006;34:1297e1310.
Vol.
-
No.
- -
2019
Mortality Prediction for EOL Communication
9
19. Minne L, Ludikhuize J, De Jonge E, De Rooij S, AbuHanna A. Prognostic models for predicting mortality in elderly ICU patients: a systematic review. Intensive Care Med 2011;37:1258e1268.
25. Bowen DJ, Kreuter M, Spring B, et al. How we design feasibility studies. Am J Prev Med 2009;36:452e457.
20. Minne L, Abu-Hanna A, de Jonge E. Evaluation of SOFA-based models for predicting mortality in the ICU: a systematic review. Crit Care 2008;12:R161.
27. Waltz CF, Strickland OL, Lenz ER. Measurement in Nursing and Health Research, 2nd ed. New York: Springer Publishing Company, 2010.
21. Vincent JL, de Mendonc¸a A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on ‘‘sepsis-related problems’’ of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793e1800.
28. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care 2007;19:349e357.
22. Metnitz PGH, Moreno RP, Almeida E, et al. SAPS 3dFrom evaluation of the patient to evaluation of the intensive care unit. Part 1: objectives, methods and cohort description. Intensive Care Med 2005;31:1336e1344. 23. Moreno RP, Metnitz PGH, Almeida E, et al. SAPS 3dFrom evaluation of the patient to evaluation of the intensive care unit. Part 2: development of a prognostic model for hospital mortality at ICU admission. Intensive Care Med 2005;31:1345e1355. 24. Sekhon M, Cartwright M, Francis JJ. Acceptability of healthcare interventions: an overview of reviews and development of a theoretical framework. BMC Health Serv Res 2017;17:88.
26. Tickle-Degnen L. Nuts and bolts of conducting feasibility studies. Am J Occup Ther 2013;67:171e176.
29. Hsieh H-F, Shannon SE. Three approaches to qualitative content analysis. Qual Health Res 2005;15:1277e1288. 30. Sandelowski M. What’s in a name? Qualitative description revisited. Res Nurs Health 2010;33:77e84. 31. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801e810. 32. Lautrette A, Darmon M, Megarbane B, et al. A communication strategy and brochure for relatives of patients dying in the ICU. N Engl J Med 2007;356:469e478. 33. Ayala GX, Elder JP. Qualitative methods to ensure acceptability of behavioral and social interventions to the target population. J Public Health Dent 2011;71(Suppl 1): S69eS79.