Abstract
Background
Code status orders in hospitalized patients guide urgent medical decisions. Inconsistent terminology and treatment options contribute to varied interpretations.
Objective
To compare two code status order options, traditional (three option) and modified to include additional care options (four option).
Design
Prospective, randomized, cross-sectional survey conducted on February–March 2020. Participants were provided with six clinical scenarios and randomly assigned to the three or four option code status order. In three scenarios, participants determined the most appropriate code status. Three scenarios provided clinical details and code status and respondents were asked whether they would provide a particular intervention. This study was conducted at three urban, academic hospitals.
Participants
Clinicians who routinely utilize code status orders. Of 4006 participants eligible, 549 (14%) were included.
Main Measures
The primary objective was consensus (most commonly selected answer) based on provided code status options. Secondary objectives included variables associated with participant responses, participant code status model preference, and participant confidence about whether their selections would match their peers.
Key Results
In the three scenarios participants selected the appropriate code status, there was no difference in consensus for the control scenario, and higher consensus in the three option group (p-values < 0.05) for the remaining two scenarios. In the scenarios to determine if a clinical intervention was appropriate, two of the scenarios had higher consensus in the three option group (p-values 0.018 and < 0.05) and one had higher consensus in the four option group (p-value 0.001). Participants in the three option model were more confident that their peers selected the same code status (p-value 0.0014); however, most participants (72%) preferred the four option model.
Conclusions
Neither code status model led to consistent results. The three option model provided consistency more often; however, the majority of participants preferred the four option model.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11606-023-08146-y.
INTRODUCTION
Code status orders are designed to guide decisions about cardiopulmonary resuscitation and may have an impact on other urgent treatment decisions. Clinicians routinely enter code status orders while admitting patients to the hospital. The goal of documenting a patient’s code status is twofold: (1) provide uniform language to guide rapid action in the case of clinical deterioration and (2) ensure that cardiopulmonary resuscitation decisions align with a patient’s wishes. A basic code status order provides two choices: (1) Full Code (perform all interventions) or (2) Do Not Resuscitate (DNR) (allow a natural death).1 Additional code status choices are also available, including Do Not Intubate (DNI), to more accurately reflect a patient’s wishes.2
While many facilities use full code, DNR, and/or DNI code status orders, code status order options may vary in the number of options as well as the scope of interventions covered by code status.3 For example, some code status orders only direct care decisions about cardiopulmonary resuscitation in the event of cardiopulmonary arrest, whereas other code status orders provide direction on the use of interventions such as mechanical ventilation for respiratory failure and vasopressors for hypotension. The inconsistent language and options offered in code status orders across institutions has the potential to create unintended variations in care.
Although code status orders are intended to guide cardiopulmonary resuscitation, past studies have demonstrated that code status orders may influence clinician decision-making in circumstances outside the scope of the code status order.4–9 These studies show that a clinician’s inference about a patient’s goals, values, and preferences for treatment influences the treatments that are offered and/or believed to be acceptable to patients. Discordance between clinician code status orders and patient preferences frequently exists.10 Some institutions attempt to provide guidance about interventions beyond cardiopulmonary resuscitation in code status orders. When the scope of code status orders extends beyond cardiopulmonary resuscitation, the intention may be to reduce the impact of biases that clinicians may have about patients with DNR orders. It’s unclear whether such guidance improves concordance between patient preferences and care received or whether it adds clarity for clinicians in urgent scenarios.11
The lack of a clear definition of terms may result in a different interpretation across various patients, disciplines, and hospital cultures. While there have been studies of various types of terminology for code status orders, few attempts have been made to study the different interpretation of the terms.12–14 In one study, resuscitation plans designed to provide broader guidance about overall care were inconsistently applied and interpreted in a series of vignettes, revealing the potential for clinically significant differences in treatment choices.15 The State of Ohio has codified the use of two different “do not resuscitate order” options: (1) comfort care (provision of only interventions oriented towards comfort) and (2) comfort care arrest (comfort care only after respiratory or cardiac arrest).16 Patients with comfort care arrest orders received more life-prolonging interventions than the comfort care order, suggesting possible mitigation of bias from DNR orders. However, the study did not examine variations in how clinicians might convert a patient’s stated wishes into a code status or how they use each specific code status to make decisions in specific scenarios. In addition, no conclusions can be made about whether patients with either code status received care concordant with their goals or if factors outside the code status contributed to treatment decisions.
The goal of this study is to compare clinician use of two code status order sets by examining variations in how clinicians may use code status in specific clinical scenarios. Using a series of vignettes, variations in clinician selection of code are explored as well as care decisions in urgent scenarios besides cardiopulmonary arrest. It is hypothesized that the inclusion of modifiers intended to align interventions with patient’s goals may not reduce variation in code status selection by clinicians or care decisions compared to a code status order set limited to CPR and intubation.
METHODS
A randomized, cross-sectional survey was performed on February–March of 2020. This study was conducted at three urban, academic hospitals within two different hospital systems, including a health system with over 1400 beds in multiple hospitals and a separate safety-net hospital with 315 inpatient beds. Participants included nurse practitioners (NPs), physician assistants (PAs), and physicians (residents, fellows, and attendings) who routinely enter and utilize code status orders during in-hospital patient care. Exclusion criteria included providers who do not provide inpatient care. The study was approved by the Institutional Review Board of both hospital systems.
A list of email addresses for all physicians, NPs, and PAs who work inpatient was obtained from each hospital system. Demographics, including participate role, was not available in the email lists and was self-identified by participants. All participants received a series of clinical vignettes with questions about either code status selection or care decisions based on a code status. Participants were sent an email requesting that they participate in the survey via REDCap. Three waves of survey requests were sent to participants via email in 2-week intervals. Emails that were undeliverable were excluded. Participants who replied that they were not practicing clinically were also excluded. Any participants that did not complete the survey in its entirety were considered incomplete and excluded.
A survey was created using six hypothetical clinical scenarios which are common in clinical practice (Fig. 2a and b). Surveys underwent validity testing. First, feedback on content and clarity was obtained from physician faculty and fellows. Next, surveys were sent to three experts in medical decision-making and modified based on feedback. In the first three scenarios, clinicians were asked to select the most appropriate code status after reviewing a brief vignette which detailed the clinical scenario and patient wishes (Fig. 2a). Scenario 1 was designed to serve as an internal control and had the least ambiguity. Three additional vignettes (Fig. 2b) provided clinical details along with the patient’s code status. Respondents were asked to decide whether they would provide a particular care intervention or indicate if they did not have enough information to proceed. The goal of this was to examine how the respondents interpreted code status and provided modifiers, as applicable, in the context of non-cardiopulmonary arrest scenarios.
All participants were randomly assigned to complete the survey with one of two different code status order sets via a random number generator in Excel prior to the survey being distributed. Participants in group A were asked to answer using a three option code status model (full code, DNR, DNR/DNI) versus group B participants who were given a four option code status model (full code, DNR/comprehensive care, DNR/DNI/comprehensive care, and DNR/comfort care). The four option model utilized in this study is identical to the options provided at one of the hospital systems utilized in the study. Comprehensive care is defined as appropriate medical interventions up to and including life-support measures in the ICU. DNR/comprehensive care is defined as comprehensive treatment (including intubation) to be initiated or continued until the event of cardiac arrest. DNR/DNI/comprehensive care is defined as comprehensive treatment to be initiated or continued until the event of cardiac and/or respiratory arrest. Although these terms are defined in hospital policy, they are not included in the order set when placing the order, and therefore were not provided to participants in the survey. The three option model differs slightly from the other study site — the hospital utilizes “partial code” rather than DNR; however, DNR was utilized based on review of prior studies3 as a more generalizable comparator.
The primary outcomes included consensus in vignette-based code status selection and variations in care choices based on the provided code status options. Secondary outcomes included variables associated with participant responses, participant preference for a three versus four option code status model, and participant confidence about whether their selections would match the selections of their peers.
Data was analyzed using STATA. Demographics were analyzed using Fisher’s exact or t-test as appropriate. Responses for scenarios 1 through 3 were categorized as consensus, which was defined as the most selected answer for that scenario, versus non-consensus, which was a collection of all of the other answer choices. Scenarios 4 through 6, which examined if proposed intervention aligned with the code status order, were answered as yes/no/not enough information. All scenario comparisons were analyzed with Fisher’s exact test. A bivariate logistical regression model was utilized to examine a priori selected demographic variables with preferred code status. For variables included in the model, any category “other” with a cell count of five or less was not included.
RESULTS
There were 4056 participants identified. After excluding participants who did not meet inclusion criteria, 4006 clinicians were eligible for enrollment (see Appendix, Fig. 1). One-hundred and forty-five participants started but did not complete the survey and 3312 did not respond. A total of 549 (14%) participants completed the survey and were included in the analysis.
There were more female respondents than males (54% versus 45%) (Table 1). Almost half of the participants were attending physicians (48.1%), followed by residents (23.0%) and advanced practice providers (22.4%). There was a statistically significant difference in the clinical practice role of the participants between the groups (p-value 0.025). There was no difference in where participants spent their clinical time, hospital setting, or specialty (p-values > 0.05). More than half of clinicians have been practicing clinically for 1–10 years (57.5% of the entire cohort). There was no difference between what code status model was utilized at the respondent’s primary clinical site between groups (p-value 0.14).
Table 1.
Demographics
| Variable | Entire cohort N = 549, N (%) |
Three option model N = 281, N (%) |
Four option model N = 268, N (%) |
p-value |
|---|---|---|---|---|
|
Gender Male Female Prefer not to say |
247 (45.3) 296 (54.3) 2 (0.4) |
132 (47.3) 145 (52.0) 2 (0.7) |
115 (43.3) 151 (56.8) 0 (0.0) |
0.232 |
|
Role Advance practice provider Resident Fellow Attending Other |
123 (22.4) 126 (23.0) 33 (6.0) 264 (48.1) 3 (0.6) |
51 (18.2) 73 (26.0) 22 (7.8) 133 (47.3) 2 (0.7) |
72 (26.9) 53 (19.8) 11 (4.1) 131 (48.9) 1 (0.4) |
0.025 |
|
Clinical time Outpatient clinic Ward PCU ICU OR ER Other |
103 (18.8) 174 (31.8) 21 (3.8) 89 (16.2) 39 (7.1) 81 (14.8) 41 (7.5) |
53 (18.9) 89 (31.8) 8 (2.9) 46 (16.4) 24 (8.6) 46 (16.4) 14 (5.0) |
50 (18.7) 85 (31.7) 13 (4.9) 43 (16.0) 15 (5.6) 35 (13.1) 27 (10.1) |
0.182 |
|
Hospital setting Urban academic Urban academic safety net Community Suburban Pediatric Veterans Other |
240 (43.7) 107 (19.5) 62 (11.3) 98 (17.9) 18 (3.3) 24 (4.4) |
135 (48.0) 49 (17.4) 23 (8.2) 54 (19.2) 9 (3.2) 11 (3.9) |
105 (39.2) 58 (21.6) 39 (14.6) 44 (16.4) 9 (3.4) 13 (4.9) |
0.088 |
|
Specialty Medicine and subspecialties Pediatrics Surgery Emergency/anesthesia Other |
276 (50.3) 81 (14.8) 75 (13.7) 97 (17.7) 20 (3.6) |
143 (50.9) 40 (14.2) 36 (12.8) 54 (19.2) 8 (2.9) |
133 (49.6) 41 (15.3) 39 (14.6) 43 (16.0) 12 (4.5) |
0.699 |
|
Years of practice 1–10 11–20 21 + |
315 (57.5) 128 (23.4) 105 (19.2) |
165 (58.9) 61 (21.8) 54 (19.3) |
150 (56.0) 67 (25.0) 51 (19.0) |
0.663 |
|
Model used at clinical setting Three option model Four option model Other |
107 (19.7) 399 (73.4) 38 (7.0) |
49 (17.8) 212 (76.8) 15 (5.4) |
58 (21.6) 187 (69.8) 23 (8.6) |
0.141 |
For scenario 1 (Fig. 2a), which was the control scenario, there was no significant difference in consensus about an appropriate code status between the three option and the four option groups (p-value 0.288) (Table 2). In contrast, for scenarios 2 and 3, the three option group selected the consensus code status 99.3% and 97.8%, respectively, whereas the four option group selection of the consensus option was 87.3% for scenario 2 and 70.2% for scenario 3 (p-values < 0.05).
Table 2.
Clinical Scenarios to Select the Appropriate Code Status
| Scenario | Answer choices | All participants N (%) |
Consensus N (%) |
Non consensus N (%) |
p-value |
|---|---|---|---|---|---|
| 1 |
Three option group Full code DNR, intubate in event of respiratory arrest DNR/DNI |
7 (2.5) 7 (2.5) 263 (95.0) |
263 (95.0) | 14 (5.1) | 0.288 |
|
Four option group Full code DNR/comprehensive care DNR/DNI/comprehensive care DNR/comfort care |
3 (1.1) 2 (0.8) 16 (6.1) 243 (92.1) |
244 (92.1) | 20 (7.6) | ||
| 2 |
Three option group Full code DNR, intubate in event of respiratory arrest DNR/DNI |
0 (0.0) 2 (0.7) 278 (99.3) |
278 (99.3) | 2 (0.7) | < 0.05 |
|
Four option group Full code DNR/comprehensive care DNR/DNI/comprehensive care DNR/comfort care |
0 (0.0) 6 (2.3) 233 (87.3) 28 (10.5) |
233 (87.3) | 34 (12.7) | ||
| 3 |
Three option group Full code DNR, intubate in event of respiratory arrest DNR/DNI |
1 (0.4) 5 (1.8) 272 (97.8) |
272 (97.8) | 6 (2.2) | < 0.05 |
|
Four option group Full code DNR/comprehensive care DNR/DNI/comprehensive care DNR/comfort care |
0 (0.0) 9 (3.4) 188 (70.2) 71 (26.5) |
188 (70.2) | 80 (29.9) |
In scenario 4 (Fig. 2b), there was less variation in deciding if the next clinical step (intubation) is appropriate based on code status in the three option group than the four option group (p-value 0.018) (Table 3). In the three option group, 95% of respondents selected “yes” and 2.1% and 2.9% selected “no” and “not enough information” respectively. In contrast, the four option group had 89.9%, 7.1%, and 3.0% selecting “yes”, “no”, and “not enough information” respectively.
Table 3.
Clinical Scenarios to Determine Next Steps in Care
| Scenario | Three option group N (%) |
Four option group N (%) |
p-value |
|---|---|---|---|
|
Scenario 4 Yes No Not enough information |
267 (95.0) 6 (2.1) 8 (2.9) |
241 (89.9) 19 (7.1) 8 (3.0) |
0.018 |
|
Scenario 5, Part 1 Yes No Not enough information |
203 (72.5) 20 (7.1) 57 (20.4) |
94 (35.3) 108 (40.6) 64 (24.1) |
< 0.05 |
|
Scenario 5, Part 2* Yes No |
36 (64.3) 20 (35.7) |
35 (55.6) 28 (44.4) |
0.355 |
|
Scenario 6, Part 1 Yes No Not enough information |
135 (48.4) 85 (30.5) 59 (21.2) |
163 (60.8) 78 (29.1) 27 (10.1) |
0.001 |
|
Scenario 6, Part 2 Yes No |
31 (16.0) 163 (84.0) |
40 (21.2) 149 (78.8) |
0.236 |
*Question only populated for those who selected “not enough information” in scenario 5, part 1
For scenario 5, part 1 (Fig. 2b), the three option group respondents had almost three-quarters who selected the answer “yes” for the next steps in the patient’s care (ordering antibiotics), whereas the four option group had 35.3% select “yes”, 40.6% select “no”, and 24.1% select not enough information (p-value < 0.05). Scenario 5, part 2, which provided additional history in order to decide if ordering antibiotics was appropriate, had similar results between both groups (p-value 0.355). Later, a typo in the code status was identified for this scenario, where “DNI” was included in the “DNR/comfort care” code status.
In scenario 6 (Fig. 2b), applying next clinical steps (electrical cardioversion) was significantly different (p-value 0.001) between both groups with 48.4% of participants in the three option group selecting “yes”, 30.5% selecting “no”, and 21.2% selecting “not enough information”, compared to the four option group who had 60.8% “yes”, 29.1% “no”, and 10.1% “not enough information”. Scenario 6, part 2, which additional context to decide regarding decision about electrical cardioversion, had similar responses between both groups (p-value 0.236).
There was a preference in both groups for the four option model (75% in the three option model and 68% in the four option model, p-value 0.070). The logistic regression model assessed variables associated with the participants preferred code status model, which is shown in Table 4. Survey group, gender, role, years of experience, specialty, and model utilized at the primary clinical site were all not associated with the preferred code status model.
Table 4.
Logistic Regression of Preferred Code Status Model
| Variable | Crude OR (95% CI) |
p-value | Adjusted OR (95% CI) |
p-value |
|---|---|---|---|---|
|
Survey group (reference: 3 option model) Four option model |
1.4 (0.98, 2.1) | 0.066 | 1.5 (0.98, 2.2) | 0.061 |
|
Gender (ref: male) Female |
1.4 (0.94, 2.0) | 0.104 | 1.5 (0.94, 2.2) | 0.092 |
|
Role (ref: attending) Advance practice provider Resident Fellow |
0.98 (0.61, 1.6) 0.68 (0.42, 1.1) 1.16 (0.54, 2.5) |
0.943 0.135 0.706 |
0.74 (0.43, 1.3) 0.85 (0.47, 1.5) 1.5 (0.67, 3.6) |
0.266 0.581 0.309 |
|
Years of practice (ref: 1–10) 11–20 21 + |
1.2 (0.76, 1.9) 1.5 (0.96, 2.5) |
0.409 0.063 |
1.2 (0.71, 2.1) 1.7 (0.97, 3.0) |
0.491 0.065 |
|
Specialty (ref: medicine and subspecialties) Pediatrics Surgery Emergency/anesthesia Other |
1.1 (0.65, 1.9) 1.6 (0.9, 2.7) 0.87 (0.51, 1.5) 0.76 (0.24, 2.4) |
0.696 0.113 0.611 0.633 |
1.0 (0.56, 1.8) 1.6 (0.90, 2.8) 0.88 (0.51, 1.5) 0.96 (0.30, 3.1) |
0.972 0.107 0.666 0.953 |
|
Model used at clinical setting (ref: 3 option) Four option model Other |
1.0 (0.62, 1.6) 1.1 (0.46, 2.4) |
0.972 0.901 |
0.98 (0.59, 1.6) 1.1 (0.47, 2.6) |
0.932 0.819 |
Participants in the three option model were more confident that their peers selected the same code status than those in the four option model (p-value 0.0014).
DISCUSSION
In this study examining the impact of code status options on clinician decision-making, the use of a four option code status with goals of care modifiers led to greater variation in code status selection compared to a three option. Vignettes were used as a model for what a clinician may face in a rapid response or code situation when they may only have code status to drive initial decision-making. With either code status menu, clinicians had significant disagreement about clinical decisions besides cardiopulmonary resuscitation. The presence of goals of care modifiers appeared to lead to greater disagreement in some cases, although with the possibility of greater certainty about some decisions.
Greater consensus in clinician code status selections existed when clinicians were given three code status options compared to four options. Information about patient goals and preferences given in the vignettes may have been insufficient for the group given four code status options to make a code status determination, yet this may reflect the reality of clinical practice where information about goals of care may be incomplete at the time the code status is entered. This finding may also be due to “choice overload,” whereby having more choices degrades the decision-making process.17, 18 There are limited data regarding the impact of choice overload in clinical decision-making.
The goals of care modifiers in the four option code status order are designed to guide care decisions outside of cardiopulmonary resuscitation. The wide variation in responses in the four option group suggests the possibility that clinicians do not share a uniform understanding of how such modifiers should be applied. One concern frequently cited about DNR orders is the possibility of withholding of potentially beneficial and desired care interventions outside of cardiopulmonary resuscitation4–9, which has also been shown in prior vignette-based studies.8, 19, 20 A previous study demonstrated significant variation in the definitions by providers for various resuscitation terms, including phrases like “comfort care” and “full ward measures,” which in turn led to differences in application of said terms in clinical scenarios.15 Moreover, some terms such as “comfort care” can contribute to provider bias that can impact clinical decision-making. Clinicians should remain cautious not to make assumptions about patient wishes based on the presence of a modifier such as comfort or comprehensive.
The study results suggest that modifiers may alter but not eliminate the impact of anchoring and framing biases on treatment decisions when uncertainty is present.21 The results of scenarios 5 and 6 suggest a possible tradeoff of having the modifiers. The “comprehensive” modifier in scenario 6 led more clinicians to initially commit to cardioversion. However, for participants in both study groups who initially thought cardioversion was appropriate or were uncertain, being presented with additional information from a surrogate led a similar majority in each group indicating they would not proceed with cardioversion. Modifiers may lead clinicians to commit to a plan without seeking important context, such as additional surrogate input, chart review for documentation of goals of care conversations, or clinical observation over time. On the other hand, the “comfort” modifier in scenario 5 led more clinicians to an initial decision to withhold antibiotics, a decision which can be appropriate for a dying patient if antibiotics are deemed to be unnecessary for comfort and this is the patient’s primary goal.22 Modifiers might increase goal concordance of care in some circumstances, although may be insufficient or even misleading in situations when goals of care and preferences are nuanced.
The participants in this study were from diverse backgrounds, level, and years of training and experiences. There was some variability in the groups, namely in the role of the participants, which was likely due to participants who chose to respond to the survey rather than randomization. The majority of participants in both groups indicated a preference for the four option code status model. The survey was not designed to assess the reasons for this preference. Participants who already work with the four option model, which is the majority of participants, may have greater comfort or positive experiences with this model. The results of this study and further research clarifying the benefits and downsides of modifiers may modulate clinician preferences.16
There are other limitations to this study. The study was conducted at hospitals affiliated with a single academic health center in the Midwest, which may not be generalizable to a larger population. The vignette approach does not account for other information that clinicians may consider or have access to in real life, such as progress note documentation and the perspective of other team members such as bedside nurses. Many clinicians work across the hospital systems, which is not adequately captured in this survey. While the sample size was large, the response rate was low. There was a difference between both groups with the clinical role of participants, with more advanced practice providers in the four option group, which was likely due to variable response in both groups as randomization was done without knowledge of role and prior to surveys being emailed. Given the paucity of literature on this topic, it is unclear what confounding effect that might have on the results. Lastly, while scenarios 4, 5, and 6 were designed to examine how code status modifiers may influence decisions in non-arrest circumstances, the option of “not enough information” may have confounded analysis of initial participant responses to these questions.
Our study overall demonstrates that neither code status model led to consistent results in all scenarios. The four option model was preferred, although the three option model provided more consistency in code status application and greater participant confidence that answers would match those of others. Additional clinical context allowed for more consistent next steps with both models. Our findings imply that the studied four option code status model likely does not result in improved consistency of application. More studies are needed to determine the real-life impact of code status models on goal concordant care and whether this impacts uniformity of clinical decisions.
Supplementary Information
Below is the link to the electronic supplementary material.
Declarations
Conflict of Interest
The authors declare that they do not have a conflict of interest.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Rabkin M, Gillerman G, Rice N. Orders Not to Resuscitate. N Engl J Med. 1976;295:364–366. doi: 10.1056/NEJM197608122950705. [DOI] [PubMed] [Google Scholar]
- 2.Rubins JB. Use of Combined Do-Not-Resuscitate/Do-Not Intubate Orders Without Documentation of Intubation Preferences: A Retrospective Observational Study at an Academic Level 1 Trauma Center Code Status and Intubation Preferences. Chest. 2020;158(1):292–297. doi: 10.1016/j.chest.2020.02.020. [DOI] [PubMed] [Google Scholar]
- 3.Batten JN, Blythe JA, Wieten S, et al. Variation in the design of Do Not Resuscitate orders and other code status options: A multi-institutional qualitative study. BMJ Qual Saf. 2021;30(8):668–677. doi: 10.1136/bmjqs-2020-011222. [DOI] [PubMed] [Google Scholar]
- 4.Stevenson EK, Mehter HM, Walkey AJ WR. Association between do not resuscitate/do not intubate status and resident physician decision-making. A national survey. Ann Am Thorac Soc. 2017;14(4):536–542. [DOI] [PMC free article] [PubMed]
- 5.Wilcox SR, Richards JB SE. Association between do not resuscitate/do not intubate orders and emergency medicine residents’ decision making. J Emerg Med. 2019;S0736–4679(30816–9) [DOI] [PubMed]
- 6.Beach MCMR. The effect of do-not-resuscitate orders on physician decision-making. J Am Geriatr Soc. 2002;50(12):2057–2061. doi: 10.1046/j.1532-5415.2002.50620.x. [DOI] [PubMed] [Google Scholar]
- 7.Hiraoka E, Homma Y, Norisue Y, Naito T, Kataoka Y, Hamada O, et al. What is the true definition of a “Do-Not-Resuscitate” order? A Japanese perspective. Int J Gen Med. 2016;9:213–220. doi: 10.2147/IJGM.S105302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Neufeld MY, Sarkar B, Wiener RS, Stevenson EKNC. The effect of patient code status on surgical resident decision making: A national survey of general surgery residents. Surgery. 2020;167(2):292–297. doi: 10.1016/j.surg.2019.07.002. [DOI] [PubMed] [Google Scholar]
- 9.Cohen R, Lisker G, Eichorn A, Multz A, Silver A. The impact of do-not-resuscitate order on triage decisions to a medical intensive care unit. J Crit Care. 2009;24(2):311–315. doi: 10.1016/j.jcrc.2008.01.007. [DOI] [PubMed] [Google Scholar]
- 10.Gehlbach TG, Shinkunas LA, Forman-Hoffman VL, Thomas KW, Schmidt GAKL. Code status orders and goals of care in the medical ICU. Chest. 2011;139(4):802–809. doi: 10.1378/chest.10-1798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sanders JJ, Curtis JR, Tulsky JA. Achieving Goal-Concordant Care: A Conceptual Model and Approach to Measuring Serious Illness Communication and Its Impact. J Palliat Med. 2018;21(S2):S17–S27. doi: 10.1089/jpm.2017.0459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Venneman SS, Narnor-Harris P, Perish MHM. Allow natural death" versus “do not resuscitate”: three words that can change a life. J Med Ethics. 2008;34(1):2–6. doi: 10.1136/jme.2006.018317. [DOI] [PubMed] [Google Scholar]
- 13.Breault JL. DNR, DNAR, or AND ? Is Language Important ? Ochsner J. 2011;11:302–306. [PMC free article] [PubMed] [Google Scholar]
- 14.Fan SY, Wang YW LI. Allow natural death versus do-not-resuscitate: titles, information contents, outcomes, and the considerations related to do-not-resuscitate decision. BMC Palliat Care. 2018;17(1). [DOI] [PMC free article] [PubMed]
- 15.Dignam C, Thomas J, Brown M TC. The impact of language on the interpretation of resuscitation clinical care plans by doctors. A mixed methods study. PLoS One. 2019;14(11). [DOI] [PMC free article] [PubMed]
- 16.Chen YY, Gordon NH, Connors AF, Garland A, Chang SC YS. Two distinct Do-Not-Resuscitate protocols leaving less to the imagination: an observational study using propensity score matching. BMC Med. 2014;12(146). [DOI] [PMC free article] [PubMed]
- 17.Peters E, Klein W, Kaufman A, DA Meilleur L. More Is Not Always Better: Intuitions About Effective Public Policy Can Lead to Unintended Consequences. Soc Issues Policy Rev. 2013;7(1):114–148. doi: 10.1111/j.1751-2409.2012.01045.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Iyengar SSLM. When choice is demotivating: can one desire too much of a good thing? J Pers Soc Psychol. 2000;79(6):995–1006. doi: 10.1037/0022-3514.79.6.995. [DOI] [PubMed] [Google Scholar]
- 19.Wilcox SR, Richards JB, Stevenson EK. Association between Do not Resuscitate/Do not intubate orders and emergency medicine residents’ decision making. J Emerg Med. 2019:1–7. 10.1016/j.jemermed.2019.09.033 [DOI] [PubMed]
- 20.Stevenson EK, Mehter HM, Walkey AJ, Soylemez Wiener R. Association between Do Not Resuscitate / Do Not Intubate Status and Resident Physician Decision-making. Ann Am Thorac Soc. 2017;14(4):536–542. doi: 10.1513/AnnalsATS.201610-798OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wellbery C. Flaws in Clinical Reasoning: A Common Cause of Diagnostic Error. Am Fam Physician. 2011;84(9):1042–1044. [PubMed] [Google Scholar]
- 22.Juthani-Mehta M, Malani PN, Mitchell SL. Antimicrobials at the End of Life: An Opportunity to Improve Palliative Care and Infection Management Manisha. Physiol Behav. 2019;176(3):139–148. doi: 10.1001/jama.2015.13080.Antimicrobials. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.