The applicability of the “surprise question” as a prognostic tool in patients with severe chronic comorbidities and its relationship with the advance directives

doi:10.21203/rs.3.rs-2471759/v1

Download PDF

Research Article

The applicability of the “surprise question” as a prognostic tool in patients with severe chronic comorbidities and its relationship with the advance directives

https://doi.org/10.21203/rs.3.rs-2471759/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 12 Oct, 2023

Read the published version in BMC Medical Education →

You are reading this latest preprint version

Background: Life expectancy in recent decades has increased the prevalence of chronic diseases in the population, requiring an approach to new health topics, such as discussions on quality of life and expectations about death and dying. The concept of advance directives (ADs) gives individuals the opportunity to make known their decisions about the treatments they would like to receive at the end of life. Despite the recognition of relevance in clinical practice, the applicability of the concept presents challenges, including establishing the appropriate prognosis for each patient and the ideal time to approach the patient. Some prognostic tools were developed, such as the surprise question (SQ): “Would you be surprised if your patient died in 12 months?”, which is used in some clinical settings to predict patient deaths and to make decisions regarding advance directives (Ads).

Method: In our one-year observational study, from July 1, 2016, to February 28, 2017, second-year resident physicians in the medical residency program (MRPIM) applied SQ to all patients with multiple and varied chronic noncommunicable comorbidities, who were followed up at the general medicine outpatient clinic (GMOC) of a tertiary university hospital in São Paulo. The frequency of the outcome (death or nondeath within 12 months) was analyzed by correlating it with the clinical data (impact of the variables studied).

Results: Eight hundred forty patients participated in the study: 6% died within one year, and there was a 24.9% prediction of death (214 patients), of which 14.9% died within one year. The correct prognosis for the subgroup of 626 patients among the residents (NPV) was 96.8% (CI = 95.4%-98.2%) and PPV = 14.9% (CI 10.1%-19, 6%), with a positive correlation between the outcomes death and the answer No to SQ, according to the number of comorbidities.

Conclusion: The SQ, in addition to care, contributed to health education, communication and care planning shared by the doctor and patient.

The project was registered under nr: 69430117.3.000.0048, under Brazilian Federal Government Committee of Ethic in Research with humans

advanced directive

palliative care

surprise question

outpatient

noncommunicable chronical diseases

Life expectancy in Brazil has increased in recent decades. Data from the Brazilian National Institute of Geography and Statistics (IBGE) indicate that the average life expectancy is 76.8 years (IBGE-DOU 25/11/2021). As a consequence, the prevalence of incurable and noncommunicable chronic diseases has increased in the population, making discussions about the quality of life and expectations about the process of death and dying essential.

At the end of the 1960s, the “living will” guidelines for ADs appeared in the United States as a way for patients to make end-of-life decisions about life support treatments and may include the choice of a legal representative to guarantee that their directives may be carried out if they lose the capacity to make their desires known ^1,2.

In Brazil, the living will or AD is a concept that is still poorly understood, while the situation is quite different in the United States, with specific laws and nongovernmental movements such as “death with dignity”, which seek to expand the rights of patients to decide on their end-of-life care ³.

In Latin America, Puerto Rico ⁴ was the first country to legislate ADs, and later, Argentina and Uruguay ⁴ also did so. Brazil still lacks specific regulations, which is why on August 31, 2012, the Federal Council of Medicine (Conselho Federal de Medicina - CFM) approved Resolution CFM 1.995, which provides for ADs ⁵. This document defines them as the set of desires, previously and expressly manifested by patients, about the care and treatment they want to receive or not if they are unable to freely express their will. If the patient has designated a representative for this purpose, the representative’s decisions will be taken into account by the health team, but the patient's desires will prevail over any other nonmedical opinion, including the family's. Only decisions in disagreement with the code of medical ethics guarantee the professional the right to revoke them ⁵.

One of the challenges is to know the appropriate time to start the discussion about ADs with a patient ^6,7. The literature shows that palliative treatment, initiated at the appropriate time, results in improved care, with lower costs for the individual and the health care system ². Understanding the patient, their values, their life and their family should precede the discussion about future priorities and care preferences. Planning end-of-life directives is a dynamic process. Studies show that patient preferences vary over time and that the choices about the proposed treatments depend on the health and life context. Billings AJ and Bernacki R point out that, generally, the proposal of ADs can occur very early, very late or at the right time. Commonly, patients decide for more "aggressive" "treatments when they are approached at the beginning of the course of their diseases ^8. An approach that is too late tends to ignore the values, desires and preferences of the patient regarding end-of-life care, and often occurs in emergency rooms or intensive care units, initiated by young doctors, who often are still in professional training and in timely patient care ⁸.

The time to begin the palliative care approach depends on many factors, but an essential element is the evaluation of the prognosis of the disease. The patient needs to be aware of his or her diagnosis, prognosis and therapeutic options for maintaining life; the health team needs to know and understand the values, goals and preferences of the patient. The proposal of ADs needs to be made with appropriate time for the patient to be able to reflect and make the decisions, together with his or her family ⁹. Therefore, trying to identify those who benefit most from this type of approach is essential ^6,8. There are several ways to predict the progression of chronic patients; however, there is no method with 100% sensitivity and specificity ¹⁰. Christakis et al. ⁷ conducted a cohort study with 468 patients involving 343 physicians in reference “hospices”. The median survival of patients was 24 days. Approximately 20% of the prognoses were accurate, 63% were overly optimistic and 17% were pessimistic ⁷, which is why tools are used to assist in the difficult and imprecise task of prediction.

A useful method that can help identify patients at high risk of dying is the strategy of the surprise question (SQ), in which the doctor is asked to answer: “Would you be surprised if your patient died in 12 months? “The health professional answers “no” if the patient has a high chance of death in this period and “yes” in the opposite case. Thus, it is intended to reduce errors in the act of predicting ^6,10.

Cited for the first time in 2001, the surprise question was developed to direct patients to the program “Improving Care through the End of Life” in Washington. In this program, physicians responsible for primary care identified patients with severe and progressive diseases, and if they answered “no” to the surprise question, these patients were referred to palliative care specialists ¹¹. Since then, the surprise question has been used frequently ^8,11,12,13 in the USA. Studies have shown a sensitivity variation of 61.4% to 75% and specificity of 70% to 90.8% ^11,12,13. The method was initially studied to evaluate patients with cancer ⁶ or chronic kidney disease ¹¹ and was also used as screening in general clinical practice to identify patients with indications for palliative care ¹⁴. The surprise question favors the improvement of care because it allows the patient to be evaluated globally, rethink and plan treatment more cautiously, since physicians tend to overestimate a patient's prognosis ^6,7. On the other hand, a critical analysis of this instrument questions the subjectivity and difficulty of specifying the exact moment when the answer would change from “yes” to “no”.

In this sense, we identified that despite the limitations and having more well-established applications in the treatment of oncological and dialysis patients, its application could be expanded as an aid tool in the prognosis and planning of care for patients with any advanced-stage clinical disease.

In teaching hospitals, there is an increase in the prevalence of patients with more than one serious and incurable chronic disease. In Brazil, the provision of care is, in most cases, still centered on the doctor, with the use of polypharmacy, without active participation of patients and relatives, despite the poor prognosis. Thus, there is an urgent need to broaden the view and care of such patients and change the teaching-learning process for medical professionals, especially resident physicians.

The present prospective and observational study aimed to stimulate the application of ADs to improve care for patients with multiple comorbidities in a general medicine outpatient clinic located in a tertiary university hospital, in São Paulo, Brazil, where resident physicians acquire skills under supervision.

Objectives

Main:

Apply the “surprise question” to the physicians of patients enrolled in the ACMG outpatient clinic, where PRMCM residents work, at a tertiary university hospital in São Paulo, Brazil. In this outpatient clinic, the vast majority of patients have multiple severe chronic comorbidities.

Specific:

Determine the sensitivity and specificity of the “surprise question” in this environment;
Evaluate whether the surprise question applies to the prognosis of the studied population;
Verify the contribution of the surprise question in decision-making.

The tool, called the surprise question (SQ), was translated and validated for the Portuguese language, explained to resident physicians and fixed on computers in the care rooms: “Would you be surprised if your patient died in 12 months?”, with a simple, binary, yes or no response.

This was an observational study (cohort) in which the main outcome was death and not death at 12 months.

The study participants were 60 second-year medical residents of the MRPIM, who initially attended a 60-minute seminar on Advanced Directives, held in March 2016 before the July 1, 2016, initiation of the study. Twenty days before the seminar, legislation from the Federal Council of Medicine, scientific articles and guiding questions were made available to residents as a stimulus to the introduction of the topic. The reading of the material was voluntary.

The main objectives of face-to-face training were to train residents:

1. to identify cases of patients with indications for palliative care, chronic disease progress curves and life expectancy below 10 years;

2. to recognize ethical and legal aspects of ADs;

3. to write ADs;

4. to develop communication skills on the subject with patients and relatives, including concepts of tolerable truth (truth telling in health care) and stages of grief.

5. to perform a spiritual approach with patients and family members.

After training on the subject, the residents were instructed and encouraged to use the surprise question with patients during their visits to the GMOC. For nine months, the residents followed-up, longitudinally, a group of patients with chronic comorbidities who were cared for in a tertiary service. The patients in the outpatient clinic had, on average, three or more chronic and severe comorbidities.

At the end of each day, the records of visits with answers to the “surprise question” for each patient were stored in a database, ensuring the anonymity of each patient.

The analyses were included in the study from 2016 February 28, to July 1, 2017. After 12 months of data collection, the outcomes of each patient (occurrence of death or current living conditions) were evaluated by means of hospital medical records or by phone.

All resident physicians who agreed to participate in the study signed an informed consent form.

Statistical analysis

Frequency analysis was used to evaluate the number of associated comorbidities for each patient with positive and negative outcomes.

The quantitative data are described as the means, standard deviations and medians, and the qualitative data are described as frequencies and percentages.

Continuous variables were defined as the mean ± standard deviation. Linear models were fitted to assess the impact of factors in determining outcomes.

The chi-square test was applied to analyze the relationship between death outcome and number of comorbidities; between the number of comorbidities and the response; between death outcome and answer to the surprise question; and between death versus response to the SQ and AD approach. As the number of comorbidities can be grouped into ordered categories, the relationship between the no response and the number of comorbidities was also evaluated using the Chi squared test for trend.

Logistic regression was used to analyze the relationship between each explanatory variable (age, sex, number of comorbidities) and the surprise question and the death and nondeath outcomes.

Sensitivity, specificity, accuracy and positive predictive values (PPV) and negative predictive values (NPV) were calculated, as well as confidence intervals (CIs) for these values. To compare means, the nonparametric Wilcoxon test was used, and Fisher’s exact test was used for proportions. The CI is an interval that contains the true value of the measure with a probability of 95%.

The graphs are presented in histograms and boxplots constructed from the arrangement of the values, facilitating the visualization of how the data are distributed in four groups (quartiles) with the same number of samples. The second quartile is the value that divides the data into two groups with the same number of values. The box is drawn beginning in the 1st quartile and ending in the 3rd quartile and contains 50% of the most central data. The dash in the middle of the box is the median. Points marked with a circle or asterisk are values that can be considered extreme.

A level of significance of 5% was considered, which is equivalent to a CI of 95%. The statistical analyses were performed using JMP® Pro version 13 - SAS Institute Inc., Cary, NC, USA, 1989-2019 and the freely available statistical language R (version 4.2.1, The Comprehensive R Archive Network).

A total of 840 patients were included. The resident physicians answered “no” to the surprise question (expected death within one year) to 214 patients in the total sample (25.5%).

Of the 840 patients included, 6% died within 1 year. Considering the expectation of dying in one year, the correct prognosis by the resident physicians was 14.9% for a subgroup of 214 patients to whom they answered “No” to the SQ. Additionally, the survival rate for the subgroup of 626 patients was 96.8%. Among the patients who would have a life expectancy above one year, only 3.2% died, as shown in Table 1, Figures 1 and 2

Table 1. Data descriptive data of the 840 patients studied

PARAMETER	RESULTS		OBSERVATIONS
SEX (%)	68.1% female
AGE (AVERAGE ± STANDARD DEVIATION)	60.9 years (DP ±14.9 years)		Minimum = 16 years, Maximum = 96 years
NUMBER OF COMORBITIES (AVERAGE ± STANDARD DEVIATION)	4.6 comorbidities (± 2.2)
DEATH IN 1 YEAR	14.9% of patients for whom doctors would not be surprised 3.2% of patients for whom doctors would be surprised
ANSWER TO SURPRISE QUESTION APPROACH TO MANAGEMENT OF ADVANCED DIRECTIVES	NO	10.8% of patients	p-valor < 0.001
	YES	0.1% of patients

Table 2 summarizes the response to the surprise question and the outcomes that occurred for the set of patients in the sample:

Table 2. Resident Physicians’ Responses to the Surprise versus Outcomes Question

		Participants %
		Absolute number	% of total sample
Response of resident physicians to the surprise question	No	214	25.6
	Yes	626	74.4
Were Advance Directives addressed?	No	813	96.8
Were Advance Directives addressed?	Yes	27	3.2
Outcomes	Non -death	788	93.8
Outcomes	Death	52	6.2

Table 2. Absolute numbers and percentages of resident physicians' responses to the surprise question; approach to advance directives and outcomes.

Table 3 shows the sensitivity and specificity values for the residents’ responses to the surprise question in regard to the outcome of death

Measure	Value	IC95%
Sensitivity	61.5%	48.3%-74.8%
Specificity	76.8%	73.9%-79.8%
Accuracy	75.9%	73.0%-78.8%
NPV	96.8%	95.4%-98.2%
PPV	14.9%	10.1%-19.6%

Table 3. Sensitivity, specificity, accuracy, negative predictive value (NPV) and positive predictive value (PPV) value and their respective confidence intervals, considering the outcome death or nondeath in one year.

Table 4 shows the correlation between the number of comorbidities and the physicians’ response to the surprise question and whether, if positive, the ADs were addressed.

Table 4. Mean comorbidities/patient, their association with physicians’ responses to the surprise question, ADs approach or not, outcome (death or not) with their respective standard deviations and p values

		Mean comorbidities/patient	Standard deviation (±)	p value
Response of Resident Physicians to the Surprise Question	No	5.4	2.3	<0.0001
	Yes	4.3	2.0	<0.0001
Were ADs addressed?	No	4.5	2.2	<0.0001
Were ADs addressed?	Yes	6.1	1.5	<0.0001
Outcome	Non -death	4.5	2.2	0.0799
Outcome	Death	5.1	2.3	0.0799

ADs = advanced directives

In order to refine the statistical analysis, the data were reanalyzed using logistic regression and Chi-square test as a tool. The results are presented below:

1. Number of comorbidities (up to 4 comorbidities X 5 or more comorbidities) and outcomes:

a. Progression to death was associated with the number of morbidities (Chi square = 5.1638,df = 1, p value = 0.02306).

b. The answer to the surprise question was associated with the number of morbidities of the patients (Chi square = 42.593, df = 1, p value = 6.741 x 10^-11).

c. The AD approach was associated with the number of morbidities of the patients (Chi square = 16.21, df = 1, p value = 5.669 x 10^-05).

2. Analysis of the proportions (outcome-ordered categories of number of morbidities)

a. The evolution to death did not show a significant association with the increase in the number of morbidities (X-squared = 6.6104, df = 5, p value = 0.2513).

b. The answer to the surprise question showed a statistically significant association with the number of morbidities in the trend analysis (Chi square = 55.263, df = 5, p value = 1.153e-10).

c. The AD approach showed a statistically significant association with the number of morbidities in the trend analysis (X-squared = 19.144, df = 5, p value = 0.001807).

3- Relationship between progression to death and response to the surprise question and AD approach

a. The answer to the surprise question was associated with the evolution of the patient's death (X-squared = 35.706, df = 1, p value = 2.295e-09).

b. The approach to AD was not associated with the evolution of the patient's death (X-squared = 2.232, df = 1, p value = 0.1352).

4. Logistic regression

a. The probability of answering “yes” to the surprise question is not associated with the patient’s gender.

b. "The probability of answering" "yes" to the surprise question has an inverse association with the age of the patient (the older the patient is, the lower the probability of answering "yes").

c. The probability of answering “yes” to the surprise question is inversely and linearly associated with morbidity categories (the lower the number of morbidities is, the greater the probability of answering “yes”).

d. The association of the probability of answering “yes” to the surprise question with the number of morbidities does not change after adjusting for the age of the patients; and the probability of addressing ADs is not associated with the patient’s gender.

f. The probability of addressing ADs is associated with the patient’s age.

g. The probability of addressing ADs is associated with the fact that the patient has up to 4 (does not address) or 5 or more morbidities (addresses), but there is no linear relationship with the number of morbidities.

h. After adjusting for age, the probability of addressing ADs is no longer significantly associated with the fact that the patient belongs to the group with up to 4 morbidities or to the group that has 5 or more morbidities, i.e., the patient’s age is more important for the resident physician when addressing ADs than the number of comorbidities.

i. The probability of progression to death was not associated with the patient’s sex.

j. The probability of progression to death is significantly associated with the age of the patient.

k. The probability of progression to death is significantly associated with the fact that the patient has up to 4 (lower mortality) or more than 5 morbidities (higher mortality).

l. The association of the probability of progression to death with the number of morbidities of the patient disappeared after adjusting for the patient's age (i.e., the patient's age was a more important factor in determining death than the number of comorbidities).

Our study evaluated the performance of the surprise question in 840 patients monitored by resident physicians under supervision in a high-complexity outpatient clinic of a tertiary care teaching hospital. The sample population of patients included in the present study had a wide variety of diagnoses and chronic, degenerative clinical conditions, introducing potential complex variables related to resident physicians (clinical inexperience) and patients (clinical complexity) for the performance analysis of the tool used.

Despite physicians with little clinical experience who had recent training in the subject, who were caring for complex patients, with an average age of 60 years or more and multiple comorbidities, the calculation of sensitivity, specificity and accuracy for the surprise question was similar to the result of the literature for more specific patients (chronic kidney and oncology). ^11,12,13

In our sample, the number of comorbidities per patient was associated with a higher probability of answering “no” to the surprise question.

The higher number of comorbidities was significantly correlated with the lack of surprise on the part of residents regarding the death of their patients in one year, as well as providing a greater approach to ADs, although the patient's age was the main reason for this approach. There was no significant correlation between the number of comorbidities and the outcome of death (Table 4).

Our study also showed that resident physicians tend to choose survival more frequently than the death of their patients, as shown in Table 3. In a systematic review, Glare P et al.¹⁵ argued that although physicians consistently overestimate survival, their predictions are highly correlated with actual survival.

White N et al., in a study with 20 vignettes (of which 14 dealt with cases of uncertain prognosis), concluded that, even when a patient had a more than 90% probability of death in 72 hours, doctors made the correct estimate in 75% of the situations. The population in White’s study was composed of volunteers with an average of more than 15 years of training and more than 10 years of practice in palliative care. The authors concluded that prognostication plays an important role in clinical practice; however, physicians should make clear to patients and family members the professional limitations of this practice ¹⁶.

In our study, resident physicians, predominantly male (data not shown), had less than three years of training and little experience in palliative care. Nevertheless, applying the surprise question, the NPV was higher than that obtained in the study by White (18), where the participants had greater professional experience, except for the differences in the tools applied and the estimated time of death.

According to data from the literature from 2003 to 2019, the PPV and NPV values of our study are consistent with the observation that physicians are better predictors of life than death.

Another important fact in our study showed that, although we were not surprised by the possibility of death in one year, the proposal of advance directives was addressed with only 10.8% of these patients. According to our results, resident physicians tend to approach ADs with patients who have more comorbidities (more than 5) and are older ages. In our analyses, it was not possible to determine whether there was an age that precipitates the decision to approach ADs, but it was possible to distinguish the influence of the patient's age and the number of comorbidities above 5 as determinants for the residents to address ADs. When refining the statistical analysis, we found greater weight for age than for the number of comorbidities. The resident physician decided to approach ADs for polymorbid and older patients.

Unpublished data from a study conducted in a waiting room of the same outpatient clinic in the same period in a population other than this study (N = 93), using a semistructured interview, found that among patients, death is still a subject that generates little reflection, especially in the male population; consequently, issues related to palliative care (PC) and the ADs are also little known by patients.

The present study demonstrates the need for greater effort in the training of physicians on approaching patients about ADs. In line with the literature, there is a need to develop a prognostic score for most patients. In addition, there is an urgent need to involve our society in the subject.

In the literature review, during the preparation of this manuscript, we found that the introduction to ADs applied to residents who participated in the present study occurred coincidentally according to the proposal of training physicians in conversations related to clinical conditions with poor prognosis, published after our study was completed ¹⁵.

In a systematic review with 16 studies, Glare et al. ¹⁵ sought to evaluate the prognostic properties of the SQ that were better in studies involving cancer patients than those with other diseases. In studies involving patients with noncancerous diseases, the likelihood ratios were less useful (LR setpositive = 2.7 [95% CI 2.1–3.6] and negative LR= 0.53 [95% CI 0.46–0.61]). Heterogeneity was low or absent for the odds ratio and likelihood ratios but generally high for sensitivity, specificity, PPV and NPV for both cancer patients and other conditions. In contrast to our study, in this systematic review, there were 2/3 false-positives for patients with diseases other than cancer.

According to Glare et al., SQ was not originally conceived as a prognostic tool but rather as a screening test for patients who could benefit from a palliative approach. Since then, it has been used for palliative or end-of-life interventions and has even been incorporated as a screening and palliative needs test (NECPAL). According to the present study, clinical experience contributes to the performance of the surprise question, with moderate interobserver harmony.

According to the authors, further studies are needed to determine whether SP combined with other clinical indicators improves the identification of patients with palliative care needs.

As far as we know, our study contributes strongly to strengthening the surprise question as a screening test, especially for the introduction of talk about advance directives, especially in countries, such as Brazil, where such strategies are not yet widely employed and the most paternalistic care prevails.

As in the present case, the SQ tool assists in the development of health care and education methods, improving professional training through prognostic communication strategies and care planning shared by the doctor and patient, in a clear expansion of the autonomy of the latter.

Study limitations:

The residents monitored the patients for a little less than a year. The answer to the surprise question was based on data reported by the resident who previously cared for the patient, with no record of his or her impression of the patient's prognosis.
Despite the multimorbid profile of the patients, the number of deaths was low, which may have influenced the reduced value of PPV.

The sensitivity, specificity and accuracy of the results obtained in this study were in agreement with the findings in the literature for chronic kidney and cancer patients (11,12,13), even if we consider that the patients in our study were elderly with multiple morbidities and that residents had not yet accumulated clinical experience.
The more morbidities a patient has, the more likely he or she will produce a “No” answer to the surprise question.
The physicians correctly matched the prognosis of death in the NO subgroup five times for the surprise question.
The resident physicians predicted better survival in the SIM subgroup for the surprise question, according to the NPV.
Physicians addressed advance directives for only 10.8% of patients who would not be surprised in case of death in one year, representing the need for training of physicians on this topic.
The calculation of NPV and PPV shows that the physicians in the present study are good predictors of life but not of death.

Competing Interest

The author(s) declare(s) that they have no competing interests.

Ethics approval and consent to participate

The authors state that this project has been approved by Brazilian Federal Government Committee of Ethics Research registered under nr: 69430117.3.000.0048, at Plataforma Brasil and all the participants signed a consent form to participate. All methods were performed in accordance with the relevant guidelines and regulations

Consent for Publication

Not Applicable.

Availability of data and materials

The authors declare that the raw data are available and uploaded as supplementary file.

Funding

Patricia Pires Pucetti, Pucetti, PP, and Leonardo Freitas, Freitas LV both received a scholarship grant from Conselho Regional de Medicina do Estado de São Paulo in the years 2016 and 2017, during the execution of the project.

Authors' contributions

Chin An Lin and Maria Nunes oriented, supervised the execution of the project and wrote the manuscript.

Patricia Pires, Leonardo Freitas, Paola Reis, Felipe Silva and Livia Herbst executed the project.

Patricia Pires conducted the organization of raw data.

Rui Nunes and Chin Jia Lin reviewed and corrected the manuscript.

Chin Jia Lin performed new statistic analysis.

Aknowledgement

The Authors would like to thank Thiago de Moura Neres for his valuable help in constructing the data tables and Marcus Vinicius do Carmo de Castro for performing first statistical analysis.

JÚNIOR, A.B.F; SATLER, L.R. Considerações sobre direito de morrer. In Portuguese : https://ambitojuridico.com.br/edicoes/revista-110/consideracoes-sobre-direito-de-morrer/. Accessed in 08^th August 2022.
LANGA, M.K. et al, Advance Directives and Outcomes of Surrogate Decision Making before Death. N Engl J Med. 2010; 362:1211-1218. April 1, 2010. N Engl J Med 2010; 362:1211-1218 DOI: 10.1056/NEJMsa0907901
SANDEEN, P. Be powered to control how you die. Available at: https://www.deathwithdignity.org/ Accessed in 2^nd, January, 2023.
FONTES, RS; SILVA Jr, AGS. Advance Directive: Historical Course in Latin America. Rev. Bioét. Vol.27. No1. Brasília, Jan/Mar, 2019. Doi: 10.1590/1983-80422019271290.
CONSELHO FEDERAL DE MEDICINA. Dispõe sobre as diretivas antecipadas de vontade dos pacientes. Resolução n. 1995, 31^st. August, 2012. In Portuguese. Available at: https://sistemas.cfm.org.br/normas/visualizar/resolucoes/BR/2012/1995 Accessed in 2^nd, January, 2023.
MOSS, A. H. et al. Prognostic Significance of the ‘‘Surprise’’ Question in Cancer Patients. Palliative Medicine, [S.l.], v. 13, n. 7, p. 837 – 840, ago. 2010.
Christakis NA, Lamont EB. Extent and determinants of error in doctors' prognoses in terminally ill patients: prospective cohort study. BMJ. 2000 Feb 19;320(7233):469-72. doi: 10.1136/bmj.320.7233.469. PMID: 10678857; PMCID: PMC27288.
BILLINGS, J. A.; BERNACKI, R. Strategic targeting of advance care planning interventions: the Goldilocks phenomenon. JAMA Intern Med, [SI], p. 620-624, 2014.
BURLÁ,C; REGO, G and NUNES, Rui. Alzheimer, dementia and the living will: a proposal. Med Health Care and Philos. . DOI: 10.1007/s11019-014-9559-8
Hui, D. Prognostication of Survival in Patients With Advanced Cancer: Predicting the Unpredictable? Cancer Control. 2015 Oct;22(4):489-97.
MOSS, A. H. et al. Utility of the "surprise" question to identify dialysis patients with high mortality. Clin J Am Soc Nephrol, [S.l.], p. 1379-84. September 2008, 3 (5) 1379-1384; DOI: https://doi.org/10.2215/CJN.00940208.
MORONI, M. et al. The “surprise question” in advanced cancer patients: a prospective study among general practitioners. Palliative Medicine, [SI], v. 28 (7), 959 – 964, 2014.
PANG, W. el al. Predicting 12-month mortality for peritoneal dialysis patients using the “surprise question”. Peritoneal Dialysis International, Hong Kong, v. 33, 60 – 66, Ago. 2012.
Yen Y, Lee Y, Hu H, et al. Early palliative care: the surprise question and the palliative care screening tool—better together. BMJ Supportive & Palliative Care 2022;12:211-217.
Glare P, Virik K, Jones M, Hudson M, Eychmuller S, Simes J et al. A systematic review of physicians' survival predictions in terminally ill cancer patients BMJ 2003; 327 :195 doi:10.1136/bmj.327.7408.195.
White N, Reid F, Vickerstaff V, et al. BMJ Supportive & Palliative Care Epub ahead of print: [please include Day Month Year]. doi:10.1136/ bmjspcare-2018-001761.
Paladino J, Koritsanszky L, Nisotel L, Neville BA, Miller K, Sanders J, Benjamin E, Fromme E, Block S, Bernacki R. Patient and clinician experience of a serious illness conversation guide in oncology: A descriptive analysis. Cancer Med. 2020 Jul;9(13):4550-4560. doi: 10.1002/cam4.3102. Epub 2020 May 4. PMID: 32363775; PMCID: PMC7333843.

No competing interests reported.

TablewithRawDataofAllPatientsIncludedintheStudy.docx

Download PDF

Journal Publication

published 12 Oct, 2023

Read the published version in BMC Medical Education →

Editorial decision: Major revision
05 Jul, 2023
Reviews received at journal
02 Jul, 2023
Reviewers agreed at journal
22 Jun, 2023
Reviewers invited by journal
21 Jun, 2023
Editor assigned by journal
21 Jun, 2023
Editor invited by journal
24 Jan, 2023
Submission checks completed at journal
24 Jan, 2023
First submitted to journal
12 Jan, 2023

You are reading this latest preprint version

The applicability of the “surprise question” as a prognostic tool in patients with severe chronic comorbidities and its relationship with the advance directives

Status:

Journal Publication

Version 1

Abstract

Figures

Background

Objectives

Methods

Results

Discussion

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1