Impact of the COVID-19 Pandemic on Providing Recommendations During Goals-of-Care Conversations: A Multisite Survey

Abstract

Background:

Goals-of-care conversations (GoCCs) are essential for individualized end-of-life care. Shared decision-making (SDM) that elicits patients' goals and values to collaboratively make life sustaining treatment (LST) decisions is best practice. However, it is unknown how the COVID-19 pandemic onset and associated changes to care delivery, stress on providers, and clinical uncertainty affected SDM and recommendation-making during GoCCs.

Aim:

To assess providers' attitudes and behaviors related to GoCCs during the COVID-19 pandemic and identify factors associated with provision of LST recommendations.

Design:

Survey of United States Veterans Health Administration (VA) health care providers.

Setting/Participants:

Health care providers from 20 VA facilities with high COVID-19 caseloads early in the pandemic who had authority to place LST orders and practiced in select specialties (n = 3398).

Results:

We had 323 respondents (9.5% adjusted response rate). Most were age ≥50 years (51%), female (63%), non-Hispanic white (64%), and had ≥1 GoCC per week during peak-COVID-19 (78%). Compared with pre-COVID-19, providers believed it was less appropriate and felt less comfortable giving an LST recommendation during peak-COVID-19 (p < 0.001). One-third (32%) reported either “never” or “rarely” giving an LST recommendation during GoCCs at peak-COVID-19. In adjusted regression models, being a physician and discussing patients' goals and values were positively associated with giving an LST recommendation (B = 0.380, p = 0.031 and B = 0.400, p < 0.001, respectively) at peak-COVID-19.

Conclusion:

Providers who discuss patients' preferences and values are more likely to report giving a recommendation; both behaviors are markers of SDM during GoCCs. Our findings suggest potential areas for training in conducting patient-centered GoCCs.

Introduction

Goals-of-care conversations (GoCCs) are discussions between patients (or their surrogates) and health care providers that elicit preferences and values for end-of-life care.¹ During GoCCs, guidelines recommend that providers discuss patients' goals and values for end-of-life care; explain life sustaining treatments (LST), such as cardiopulmonary resuscitation and mechanical ventilation; and create a care plan based on preferences, values, and clinical status.^1,2 When conducted well, GoCCs can improve quality of life, increase patient and family satisfaction, reduce family distress, and reduce end-of-life medical costs.^3,4

Shared decision-making (SDM) within health care ranges from patient-driven, where a patient retains complete decision-making autonomy, to provider-driven, where a patient defers to the provider's knowledge, expertise, and recommendations.^5,6 One component of SDM during GoCCs is provider provision of recommendations, which is promoted as an ethical practice and reflects patient-centered SDM.^7–10

A recommendation can be a critical component of patient-centered SDM during GoCCs as it typically stems from a deliberative process between providers and patients.¹¹ Recommendations guided by patients' preferences and values preserve patients' autonomy and allow for more fully informed end-of-life decisions.^12,13 Importantly, recommendations are wanted by patients and families; more than 90% of surrogate decision makers for critically ill patients wished to receive one.¹⁴ Yet, providing a recommendation, a routine practice in many aspects of health care, is uncommon during GoCCs, a time when clinician expertise may be particularly beneficial.^15–19

Nearly 50% of providers in an intensive care setting declined to give a recommendation during GoCCs when it was sought by surrogate decision makers.⁶ Similarly, in a survey assessing hypothetical behaviors during code status discussions, only 30% of physicians-in-training reported that they were likely to offer a recommendation on cardiopulmonary resuscitation.²⁰

At the onset of the COVID-19 pandemic, there was high mortality, with dramatic increases in hospital utilization.^21–23 Many providers conducted GoCCs under the stress of changing clinical workloads and settings, concerns about resources, and prognostic uncertainty.^22–25 Thus, we sought to understand how the COVID-19 pandemic affected SDM in GoCCs, especially with respect to the provision of recommendations about life-sustaining treatment (LST).

Methods

Study setting and population

From each of the four U.S. census regions, we selected the five VA facilities with the highest cumulative COVID-19 cases from March to October 2020. We expected these sites to be most impacted by the onset of the pandemic with respect to clinical context (e.g., providers deployed to settings different from their usual practice), workload, and burnout.

We subsequently identified eligible providers, defined as those who (1) had independent authority to place LST orders; (2) practiced in inpatient, outpatient, or long-term care facility settings; and (3) practiced in internal medicine or its subspecialties; emergency medicine; family medicine; neurology; or general surgery. We assumed that providers in these specialties would most likely be having GoCCs early in the pandemic. From March 2020 to March 2021, 3550 providers met our inclusion criteria.

Survey development and data collection procedures

We developed a survey, adapting two instruments that assessed providers' self-reported attitudes and behaviors with regards to GoCCs.^26,27 We pilot-tested the draft survey with three primary care and three intensive care physicians and made revisions based on feedback. The final survey (Supplementary Material S1) was disseminated to eligible providers in April 2021 using a modified Dillman approach, where they were e-mailed an electronic survey invitation, with up to two reminders sent at weekly intervals to non-respondents.²⁸ Except for two open-ended items, all survey items required a response. No incentive was offered.

Survey instrument

Defining pre- and peak-COVID-19

The survey instructed respondents to identify a period of “peak-COVID-19,” defined as a time of increased personal workload, work hours, COVID-19 cases at their facility, clinical demand, or something else that made it feel like the peak after the onset of the COVID-19 pandemic. We defined “pre-COVID-19” as a period before the respondents were aware of COVID-19. Thus, the “pre-COVID-19” and “peak-COVID-19” eras were personalized.

Demographics and practice characteristics

The survey inquired about participants' demographics, clinical effort, clinical setting, role (e.g., attending, fellow, resident, advanced nurse practitioner [ANP], or physician assistant [PA]), and date they obtained their terminal degree. In addition, the survey elicited the average number of GoCCs that providers engaged in per week during peak-COVID-19 (none, 1–2, 3–4, or ≥5).

Attitudes

These items were all asked twice: first with participants thinking about their pre-COVID-19 experiences and subsequently considering their peak-COVID-19 experiences. There were two items regarding general attitudes about LST recommendations during GoCCs, including comfort with and perceived appropriateness of giving LST recommendations.²⁶ Next, providers indicated their level of agreement with six statements about the general ethics of providing specific LST recommendations (e.g., it is a health care provider's duty, it unduly influences patients).²⁶

Providers indicated the ethical appropriateness of using nine specific dialogue techniques during GoCCs (e.g., use vivid imagery, recommend a time-limited therapy trial).²⁶ Providers indicated their comfort with prognosticating for patients with respiratory failure, for both those with and without COVID-19. All items were rated on a 4-point scale.

Self-reported behaviors

Providers who reported having at least one GoCC during peak-COVID-19 were queried about frequency of engaging in each of eight specific behaviors during peak-COVID-19 GoCCs (e.g., encourage the patient to articulate their values).^26,27 We selected one of the self-reported behaviors during peak-COVID-19—making a specific recommendation about LST decisions—as our primary outcome based on the a priori hypothesis that it might be omitted from SDM during early pandemic GoCCs given limited prognostic data.^7,26,29 Responses to this item serve as a proxy for providers' propensity for making a specific recommendation in GoCCs during the pandemic.

Analyses

Demographics of providers who did versus did not conduct GoCCs during peak-COVID-19

We compared demographics and practice characteristics of providers who reported having at least one GoCC during peak-COVID-19 with those who had no GoCCs using Chi-square tests for categorical variables and one-way analysis of variances for continuous variables. Specialty was reduced to: (1) internal medicine/primary care/family medicine, (2) palliative care, (3) geriatric medicine, (4) pulmonary and critical care, (5) emergency medicine, or (6) other specialties; role was reduced to medical doctors or ANPs/PAs; and frequency of GoCCs per week was reduced to: none, 1–2, and ≥3. Race and ethnicity were dichotomized to non-Hispanic white versus all others.

Attitudes during peak-COVID-19 compared to pre-COVID-19

We used Wilcoxon ranked sum tests to compare responses between peak-COVID-19 and pre-COVID-19, including items about general attitudes about LST recommendations, general ethics of providing a specific LST recommendation, and ethical appropriateness of using specific dialogue techniques during GoCCs. We also compared providers' self-reported comfort level with prognosticating for patients with respiratory failure for both those with and without COVID-19 infection.

The Bonferroni method was used to correct for multiple comparisons, with a Bonferroni-corrected alpha of 0.002.³⁰ We also assessed if mean differences between “peak-COVID-19” and “pre-COVID-19” responses were moderated by any demographic or practice characteristic.

LST recommendation during peak-COVID-19

We assessed for zero-order associations between provider demographics and practice characteristics and our primary outcome using bivariate linear regressions.

We next built four multivariable regression models to understand associations with the propensity to give LST recommendations during peak-COVID-19: (1) general attitudes about LST recommendations, (2) general ethics about providing LST recommendations, (3) ethical appropriateness of specific dialogue techniques, and (4) self-reported behaviors.

The dependent variable in all models was providers' self-reported frequency for making a specific LST recommendation during peak-COVID-19. All models control for gender, medical specialty, role, and frequency of GoCCs per week; these were included based on significant zero-order associations with an effect estimate >0.25 with the primary outcome. In all models, we excluded providers with zero GoCCs during peak-COVID-19 (n = 70).

Ethical approval

The VA Boston Healthcare System Institutional Review Board determined that this study was exempt from the requirement of 38 CFR 16 Protection of Human Subjects.

Results

There were 323 respondents; after removing 152 providers who were unable to be reached, the adjusted response rate was 9.5% (323/3398). Respondents were primarily 50 years or older (61%), female (63%), non-Hispanic white (64%), and physicians (58%). All respondent demographics are in Table 1.

Table 1.

Characteristics of Respondents Who Had ≥1 Goals of Care Conversation Compared With Those Who Had No Goals-of-Care Conversations During Perceived Peak COVID-19

	Total (n = 323) n (%)	≥1 GoCC (n = 253) n (%)	No GoCCs (n = 70) n (%)	p
Age, years
20–39	42 (13)	27 (11)	15 (21)	0.014
40–49	83 (26)	73 (29)	10 (14)
50–59	110 (34)	87 (34)	23 (33)
60 or older	88 (27)	66 (26)	22 (31)
Gender
Female	202 (63)	164 (65)	38 (54)	0.055
Male	113 (35)	82 (32)	31 (46)
Prefer not to answer	8 (2)	7 (3)	1 (1)
Race/ethnicity
White	208 (64)	158 (63)	50 (71)	0.180
Non-white	115 (36)	95 (38)	20 (28)
Black or African American	17 (5)	14 (5)	3 (4)
Asian	48 (15)	41 (16)	7 (10)
Hispanic	18 (6)	16 (6)	2 (3)
Prefer not to answer	32 (10)	24 (10)	8 (12)
Role
Attending (MD/DO)	188 (58)	143 (57)	45 (64)	0.387
ANPs/PAs	135 (42)	110 (43)	25 (36)	0.387
Specialties
Internal medicine/primary care/family medicine	89 (27)	70 (28)	19 (27)	<0.001
Palliative care	40 (12)	40 (16)	0 (0)
Geriatrics	46 (14)	41 (16)	5 (7)
Pulmonary and critical care medicine	34 (11)	32 (13)	2 (3)
Emergency medicine	15 (5)	11 (4)	4 (6)
Other subspecialties^a	99 (31)	59 (23)	40 (57)
Year of graduation, median year (IQR)	2000 (1989–2010)	2001 (1989–2009)	1998 (1986–2010)	0.443
Clinical effort
0^b	3 (1)	0 (0)	3 (4)	<0.001
>0–80%	145 (45)	103 (41)	42 (60)
81–100%	175 (54)	150 (59)	25 (36)
Clinical setting (%)
Inpatient, median (IQR)	5 (0–41)	5 (0–41)	5 (0–30)	0.853
Outpatient, median (IQR)	85 (30–100)	80 (20–100)	95 (70–100)	0.016
GoCCs during peak COVID-19
None^b	70 (22)	—	70 (100)	NA
At least one GoCCc	13 (4)	13 (5)	—
1–2 GoCCs per week	104 (32)	104 (41)	—
≥3 GoCCs per week	136 (42)	136 (53)	—

Reported specialties include anticoagulation services, allergy and immunology, addiction medicine, cardiology, dermatology, endocrinology, employee health services, hematology/oncology, infectious disease, nephrology, neurology/traumatic brain injury, occupational health services, psychiatry/mental health, pain medicine, rheumatology, radiologic services, surgery, wound care, and unspecified.

Providers were included in the analysis assessing attitudes but were not included in the multivariable analyses.

Providers who had at least one GoCC during peak COVID-19, but it did not average ≥1 per week.

ANP, advanced nurse practitioner; DO, doctor of osteopathic medicine; GoCC goals of care conversation; IQR, interquartile range; MD, doctor of medicine; PA, physician assistant.

Demographics of providers who did versus did not conduct peak-COVID-19 GoCCs

Most respondents stated they had at least one GoCC per week during peak-COVID-19 (78%). Providers with at least one peak-COVID-19 GoCC were slightly younger than those with no GoCCs during peak-COVID-19 (74% vs. 68% age <60 years, p = 0.014). Compared with providers who reported no peak-COVID-19 GoCCs, those with at least one peak-COVID-19 GoCC were more likely to have a specialty of pulmonary/critical care (13% vs. 3%, p < 0.001) or palliative care (16% vs. 0%, p < 0.001). All comparisons between those who did versus did not conduct peak-COVID-19 GoCCs are in Table 1.

Attitudes during peak-COVID-19 compared to pre-COVID-19

Fewer providers reported that it was very or somewhat appropriate to give LST recommendations during peak-COVID-19 compared with pre-COVID-19 (83% vs. 94%, p < 0.001). Providers less frequently felt very or somewhat comfortable giving LST recommendations during peak-COVID-19 compared with pre-COVID-19 (73% vs. 89%, p < 0.001). Providers more often somewhat or strongly agreed that giving recommendations was appropriate only if the patient wanted it during peak-COVID-19 compared with pre-COVID-19 (70% vs. 64%, p < 0.001).

Providers less frequently reported always or often asking patients if they wanted recommendations about LST decisions during peak-COVID-19 compared with pre-COVID-19 (27% vs 34%, p = 0.001). Lastly, providers were more often very or somewhat uncomfortable prognosticating outcomes for patients with respiratory failure from COVID-19 compared to another etiology (49% vs. 30%, p < 0.001). No other comparisons of attitudes between time periods were statistically significant (data not shown). Additional analyses did not find any associations between any demographic or practice characteristic with a shift in response between time periods (Supplementary Table S1).

LST recommendation during peak-COVID-19

There was a wide distribution in the self-reported frequency with which providers made a specific recommendation about LST decisions during peak-COVID-19, with 35% reporting either “always” or “often” and 32% reporting either “never” or “rarely.” An increased propensity for giving LST recommendations during peak-COVID-19 was associated with being a physician (B = 0.559, p < 0.001), male (B = 0.332, p = 0.04), and specializing in palliative care (B = 0.940, p < 0.001).

We also found that increased propensity for giving an LST recommendation during peak-COVID-19 was associated with having very few (i.e., <1 GoCC per week; B = 0.875, p = 0.01) or very frequent GoCCs (i.e., ≥3 GoCCs per week; B = 0.337, p = 0.03). We controlled for these variables in our multivariable regression models (Table 2). Other demographic or practice characteristics were either not statistically significantly associated with propensity for giving an LST recommendation or had very small effect estimates (Supplementary Table S2).

Table 2.

Multivariable Regression Models Predicting Propensity to Give a Life Sustaining Treatment Recommendation During Peak COVID-19

	Model 1 (p)	Model 2 (p)	Model 3 (p)	Model 4 (p)
Gender—male (ref: female)	0.08 (0.682)	0.19 (0.298)	0.12 (0.495)	−0.06 (0.706)
Specialty (ref: internal medicine/primary care/family medicine)
Palliative care	0.56 (0.021)	0.66 (0.004)	0.63 (0.008)	0.45 (0.031)
Geriatrics	0.04 (0.847)	0.09 (0.681)	−0.09 (0.677)	−0.12 (0.512)
Pulmonary and critical care medicine	0.01 (0.978)	0.05 (0.829)	−0.01 (0.959)	0.14 (0.511)
Emergency medicine	−0.07 (0.840)	0.06 (0.877)	−0.23 (0.835)	0.38 (0.235)
Other	−0.03 (0.878)	−0.10 (0.620)	0.08 (0.674)	0.20 (0.239)
Role—MD (ref: PA/ANP)	0.38 (0.031)	0.38 (0.029)	−0.01 (0.971)	0.22 (0.145)
GoCC frequency per week (ref: 1–2)
<1	1.20 (<0.001)	0.98 (0.003)	0.737 (0.024)	0.92 (0.001)
≥3	0.08 (0.618)	0.17 (0.248)	0.17 (0.260)	0.18 (0.169)
How appropriate is it to provide an LST recommendation? (increasing appropriateness)	0.06 (0.494)
How comfortable are you with providing an LST recommendation? (increasing comfort)	0.36 (<0.001)
Providing a patient with a specific recommendation about LST… (increasing agreement)
Is a health care provider's duty		0.45 (<0.001)
Is appropriate only if the patient wants the recommendation		−0.03 (0.744)
Makes it easier for the patient to make LST decisions		0.19 (0.146)
Further burdens the patient		−0.23 (0.082)
Unduly influences the patient		0.09 (0.441)
Places too great a burden on the provider		0.01 (0.931)
How ethically appropriate would it have been for a provider to… (increasing appropriateness)
Discourage interventions by using vivid imagery (e.g., “CPR would break your ribs”)?			0.02 (0.787)
Recommend a time-limited trial of therapy?			0.32 (0.004)
Focus the discussion on the medical facts that support the choice the provider thinks is best for the patient?			0.19 (0.058)
Divide the decision about limiting life support into smaller decisions about individual interventions?			−0.13 (0.197)
Discuss the patient's small chance of recovery?			0.07 (0.565)
Explain what the provider would do if the patient were the provider's family member?			−0.00 (0.985)
Emphasize those patient values that support the choice the provider thinks is best for the patient?			0.31 (0.003)
Discuss the provider's prognostic uncertainty with the patient or family?			0.03 (0.787)
Make an independent treatment decision and inform the patient of that decision?			0.16 (0.065)
How often, if at all, do you… (increasing frequency)
Identify patient's health care proxy?				−0.02 (0.845)
Provide medical information to the patient?				−0.02 (0.877)
Engage in deliberations with the patient about the decision?				0.26 (<0.001)
Encourage patient to articulate their values?				−0.04 (0.678)
Ask COVID-19 patients if they want your recommendation?				0.10 (0.006)
Suggest that a particular decision is most consistent with the patient's values?				0.40 (<0.001)
Make an independent decision and inform the patient of that decision?				0.21 (<0.001)
Tell the patient that the decisions regarding code status can be changed at any time?				0.03 (0.691)

Models 1–4 are adjusted for statistically significant demographic and practice characteristics (gender, specialty, role, and GoCC frequency). Model 1: General attitude about LST recommendations. Model 2: General ethics of providing a specific LST recommendation. Model 3: Ethical appropriateness of using specific dialogue techniques. Model 4: Self-reported behavior. Bolded covariates are statistically significant associations.

CPR, cardiopulmonary resuscitation; LST, life sustaining treatment.

Model 1

Using two items about providers' general attitudes about LST recommendations as predictors, more comfort with giving an LST recommendation was associated with greater propensity for giving an LST recommendation during peak-COVID-19 (B = 0.360, p < 0.001).

Model 2

Using six items about the general ethics of providing a specific LST recommendation as predictors, only the belief that giving an LST recommendation is a health care practitioner's duty was associated with greater propensity for giving an LST recommendation during peak-COVID-19 (B = 0.450, p < 0.001).

Model 3

Using nine items about the ethical appropriateness of using specific dialogue techniques as predictors, the propensity for giving an LST recommendation during peak-COVID-19 was positively associated with offering a time-limited therapy trial (B = 0.320, p = 0.004) and emphasizing patient values aligned with provider preferences (B = 0.310, p = 0.003).

Model 4

Using eight items about self-reported behaviors during GoCCs, the propensity for giving an LST recommendation during peak-COVID-19 was positively associated with asking whether an LST recommendation is welcomed (B = 0.100, p = 0.006), engaging in the deliberative process during GoCCs (B = 0.260, p < 0.001), aligning patient values with a specific decision (B = 0.400, p < 0.001), and making an independent decision regardless of patient wishes (B = 0.210, p < 0.001).

Discussion

In our multisite survey assessing attitudes toward and behaviors related to SDM in GoCCs, only 35% of providers reported “always” or “often” making LST recommendations during peak-COVID-19, whereas 32% of providers reported “never” or “rarely” making LST recommendations. Providers also felt less comfortable and found it less appropriate to make LST recommendations during peak-COVID-19. This shift was not moderated by demographic variables, but this may have been due to the novelty of the disease itself (i.e., all providers had limited prognostic knowledge about the disease) and the anecdotal reports that more experienced providers (who are also older) reduced face-to-face clinical care given their higher personal risk of adverse outcomes if they become infected with COVID-19.

Provider characteristics and behaviors associated with a propensity for making LST recommendations during peak-COVID-19 included being a physician, asking if an LST recommendation is welcomed, discussing patients' preferences and values, and offering a time-limited therapy trial. Given the stress on providers and patients early in the pandemic, identifying behaviors associated with patient-centered SDM during GoCCs can highlight areas for improvement.

Reluctance to give LST recommendations during GoCCs has been demonstrated in other settings and provider populations.^{6,15,20,31,32} The hesitancy to provide LST recommendations found in our study could result from unique challenges of the COVID-19 pandemic, including prognostic uncertainty, increased workload and pressure, and pandemic-specific communication difficulties.³³

Even though most providers felt it was their duty to give LST recommendations and that recommendations did not unduly influence patients, there was a significant decline in the perceived appropriateness of providing a recommendation from pre- to peak-COVID-19. This decline might reflect other changes to SDM during GoCCs at the beginning of the pandemic, such as limited discussions about patients' goals and values.^7,12,13,34 Our study highlights the need to reinforce patient-centeredness during GoCCs, including eliciting patient's goals and values and, if appropriate, providing a patient-informed recommendation, even during times of uncertainty.

Recommendations during GoCCs shape patients' goals and values into treatment decisions.^12,13 Thus, a recommendation should only come after eliciting and understanding the patient's goals and values.³⁵ We found a strong association between providers who discuss patients' goals and values and propensity to give LST recommendations. A multicenter study found that only 8% of recorded family meetings had a recommendation based on patients' preferences and values.¹⁵

Interestingly, family meetings that had a recommendation contained more deliberative dialogue, suggesting a connection between discussing patients' goals and values and recommendations, both markers of patient-centered SDM within GoCCs. Our study aligns with this finding, suggesting that discussing patients' goals and values may facilitate recommendation provision, an important component of SDM during GoCC.

We found an association between the tendency to suggest a short trial of therapy and the propensity for giving LST recommendations. Similarly, in another multicenter study, offering a time-limited therapy trial during critical illness was associated with improved quality of family GoCCs and reduced intensity and duration of intensive care unit treatment.³⁶

A qualitative study assessing providers' approach to conflicts during GoCCs found that offering time-limited therapy trials helped family members accept a likely terminal prognosis and an unlikely benefit of aggressive care.³⁷ The association in our study between the tendency to suggest a short therapy trial and propensity for giving LST recommendations likely reflects a deliberative negotiation process integral to balanced SDM in GoCCs.

Our study shows a strong association between being a physician and propensity for providing LST recommendations. This aligns with findings that advanced practitioners are more likely to feel uncomfortable making end-of-life care recommendations compared with physicians.³⁸ This potential barrier to providing an informed recommendation could have been further exacerbated during the pandemic due to increased frequency of GoCCs and recruitment of advanced practitioners to conduct some of these conversations to alleviate personnel burden.^39,40

Another study assessing physicians' views on advance care planning found that only 29% of providers received formal training for GoCCs; however, younger physicians and physicians caring for a more racially and ethnically diverse population were more likely to have had such training.⁴¹ Taken together, these findings support the importance of training for all providers, including advanced practitioners, in conducting patient-centered GoCCs. Specifically, targeted behavioral areas for improvement conducting GoCCs include discussing patients' preferences and values and offering time-limited therapy, as these practices appear to facilitate provision of an informed recommendation.

Our findings should be interpreted in the context of the following limitations. A response rate of 9.5% is lower than historical studies surveying medical professionals and might reflect additional pandemic-related demands on providers resulting in limited bandwidth to participate in surveys.^42,43 However, our respondents' demographics and practice characteristics are comparable to published VA-based survey studies, which could mitigate concern from non-response bias.^44–48

We pilot tested the survey using a limited number of physicians and, although we could have conducted a broader evaluation, this process needed to be balanced with pragmatic concerns about providers' limited time. Our survey elicited responses about attitudes and behaviors over the year before survey completion; to minimize recall error, we avoided asking respondents to reflect on specific GoCCs and instead to provide their general attitudes and behaviors related to GoCCs held during each timeframe.

In addition, use of subjectively defined timeframes “pre-COVID-19” and “peak-COVID-19” allowed participants to anchor their attitudes and behaviors to when they personally felt most overwhelmed; this likely increased their recall yet may yield variability for the specific dates referenced by each respondent. Findings may not generalize to non-VA settings; however, given the global nature of the pandemic, there is applicability to other settings.

As with all surveys, findings rely on self-reported behavior, which may not reflect actual behavior. Although we found that providers were less comfortable making LST recommendations, it is important to investigate how communication challenges and the evolving therapies for COVID-19 affected providers' comfort making LST recommendations over the course of the pandemic. Future studies should also evaluate downstream effects of pandemic GoCCs by assessing trends in end-of-life decisions to determine how changes in SDM affected these decisions.

Conclusion

Our survey of providers' attitudes and behaviors regarding SDM in GoCCs during the pandemic found that providers were less comfortable and felt it was less appropriate to give LST recommendations during GoCCs during the first peak of the COVID-19 pandemic compared with the pre-pandemic. This was especially relevant during discussions with patients with COVID-19 and GoCCs conducted by advanced practitioners. However, we also identified several provider characteristics and behaviors associated with propensity for providing such recommendations, which can be leveraged to enhance the patient-centeredness of GoCCs.

Footnotes

Disclaimer

The contents of this manuscript do not represent the views of VA or the United States Government.

Funding Information

This project is funded by a VA HSR&D SWIFT award (PI: A.M.L.) and supported in part by resources from the VA Boston Healthcare System. N.M. is funded through a National Institutes of Health Institutional Training Grant (T32 HL007035).

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

References

Bernacki

, Block

. Communication about serious illness care goals: A review and synthesis of best practices. JAMA Intern Med, 2014; 174:1994–2003; doi: 10.1001/jamainternmed.2014.5271

Jain

, Bernacki

. Goals of care conversations in serious illness: A practical guide. Med Clin North Am, 2020; 104:375–389; doi: 10.1016/j.mcna.2019.12.001

Zhang

, Wright

, Huskamp

, et al. Health care costs in the last week of life: Associations with end-of-life conversations. Arch Intern Med, 2009; 169:480–488; doi: 10.1001/archinternmed.2008.587

Miller

, Scott

, Ersek

, et al. Honoring veterans' preferences: The association between comfort care goals and care received at the end of life. J Pain Symptom Manage, 2021; 61:743–754.e741; doi: 10.1016/j.jpainsymman.2020.08.039

Kon

. The shared decision-making continuum. JAMA, 2010; 304:903–904; doi: 10.1001/jama.2010.1208

White

, Malvar

, Karr

, et al. Expanding the paradigm of the physician's role in surrogate decision-making: An empirically derived framework. Crit Care Med, 2010; 38:743–750; doi: 10.1097/CCM.0b013e3181c58842

Jacobsen

, Blinderman

, Alexander Cole

, et al. “I'd recommend …” how to incorporate your recommendation into shared decision making for patients with serious illness. J Pain Symptom Manage, 2018; 55:1224–1230; doi: 10.1016/j.jpainsymman.2017.12.488

Epstein

, Alper

, Quill

. Communicating evidence for participatory decision making. JAMA, 2004; 291:2359–2366; doi: 10.1001/jama.291.19.2359

Shaw

, Elger

. Evidence-based persuasion: An ethical imperative. JAMA, 2013; 309:1689–1690; doi: 10.1001/jama.2013.2179

10.

Childers

, Back

, Tulsky

, et al. REMAP: A framework for goals of care conversations. J Oncol Pract, 2017; 13:e844–e850; doi: 10.1200/jop.2016.018796

11.

Crawford

, Petrie

, Harvey

. Shared decision-making and the implementation of treatment recommendations for depression. Patient Educ Couns, 2021; 104:2119–2121; doi: 10.1016/j.pec.2021.01.025

12.

Goldberg

, Meier

. A swinging pendulum: Comment on “on patient autonomy and physician responsibility in end-of-life care”. Arch Intern Med, 2011; 171:854; doi: 10.1001/archinternmed.2011.173

13.

Billings

, Krakauer

. On patient autonomy and physician responsibility in end-of-life care. Arch Intern Med, 2011; 171:849–853; doi: 10.1001/archinternmed.2011.180

14.

Johnson

, Bautista

, Hong

, et al. An empirical study of surrogates' preferred level of control over value-laden life support decisions in intensive care units. Am J Respir Crit Care Med, 2011; 183:915–921; doi: 10.1164/rccm.201008-1214OC

15.

Scheunemann

, Ernecoff

, Buddadhumaruk

, et al. Clinician-family communication about patients' values and preferences in intensive care units. JAMA Intern Med, 2019; 179:676–684; doi: 10.1001/jamainternmed.2019.0027

16.

, Xu

, Aladesanmi

. Provider recommendation for colorectal cancer screening: Examining the role of patients' socioeconomic status and health insurance. Cancer Epidemiol, 2009; 33:207–211; doi: 10.1016/j.canep.2009.07.011

17.

, Srivastav

, Amaya

, et al. Association of provider recommendation and offer and influenza vaccination among adults aged ≥18 years—United States. Vaccine, 2018; 36:890–898; doi: 10.1016/j.vaccine.2017.12.016

18.

Ali-Ahmed

, Pieper

, North

, et al. Shared decision-making in atrial fibrillation: Patient-reported involvement in treatment decisions. Eur Heart J Qual Care Clin Outcomes, 2020; 6:263–272; doi: 10.1093/ehjqcco/qcaa040

19.

Hawley

, Lantz

, Janz

, et al. Factors associated with patient involvement in surgical treatment decision making for breast cancer. Patient Educ Couns, 2007; 65:387–395; doi: 10.1016/j.pec.2006.09.010

20.

Einstein

, Einstein

, Mathew

. Dying for advice: Code status discussions between resident physicians and patients with advanced cancer—A national survey. J Palliat Med, 2015; 18:535–541; doi: 10.1089/jpm.2014.0373

21.

Yang

, Yu

, Xu

, et al. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: A single-centered, retrospective, observational study. Lancet Respir Med, 2020; 8:475–481; doi: 10.1016/S2213–S2600(20)30079-5

22.

Richardson

, Hirsch

, Narasimhan

, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA, 2020; 323:2052–2059; doi: 10.1001/jama.2020.6775

23.

Grasselli

, Greco

, Zanella

, et al. Risk factors associated with mortality among patients with COVID-19 in intensive care units in Lombardy, Italy. JAMA Intern Med, 2020; 180:1345–1355; doi: 10.1001/jamainternmed.2020.3539

24.

Elbay

, Kurtulmuş

, Arpacıoğlu

, et al. Depression, anxiety, stress levels of physicians and associated factors in Covid-19 pandemics. Psychiatry Res, 2020; 290:113130; doi: 10.1016/j.psychres.2020.113130

25.

Emanuel

, Persad

, Upshur

, et al. Fair allocation of scarce medical resources in the time of Covid-19. N Engl J Med, 2020; 382:2049–2055; doi: 10.1056/NEJMsb2005114

26.

Brush

, Rasinski

, Hall

, et al. Recommendations to limit life support: A national survey of critical care physicians. Am J Respir Crit Care Med, 2012; 186:633–639; doi: 10.1164/rccm.201202-0354OC

27.

Binder

, Huang

, Buss

. Uninformed consent: Do medicine residents lack the proper framework for code status discussions?. J Hosp Med, 2016; 11:111–116; doi: 10.1002/jhm.2497

28.

Hoddinott

, Bass

. The dillman total design survey method. Can Fam Physician, 1986; 32:2366–2368;

29.

Gibbon

, GrayBuck

, Buck

, et al. Development and implementation of a clinician-facing prognostic communication tool for patients with COVID-19 and critical illness. J Pain Symptom Manage, 2020; 60:e1–e6; doi: 10.1016/j.jpainsymman.2020.05.005

30.

Bland

, Altman

. Multiple significance tests: The Bonferroni method. BMJ, 1995; 310:170; doi: 10.1136/bmj.310.6973.170

31.

Deep

, Griffith

, Wilson

. Communication and decision making about life-sustaining treatment: Examining the experiences of resident physicians and seriously-ill hospitalized patients. J Gen Intern Med, 2008; 23:1877–1882; doi: 10.1007/s11606-008-0779-6

32.

Johnston

, Pfeifer

. Patient and physician roles in end-of-life decision making. J Gen Intern Med, 1998; 13:43–45; doi: 10.1046/j.1525-1497.1998.00008.x

33.

Bradshaw

, Dunleavy

, Walshe

, et al. Understanding and addressing challenges for advance care planning in the COVID-19 pandemic: An analysis of the UK CovPall survey data from specialist palliative care services. Palliat Med, 2021; 35:1225–1237; doi: 10.1177/02692163211017387

34.

Quill

, Brody

. Physician recommendations and patient autonomy: Finding a balance between physician power and patient choice. Ann Intern Med, 1996; 125:763–769; doi: 10.7326/0003-4819-125-9-199611010-00010

35.

White

, Evans

, Bautista

, et al. Are physicians' recommendations to limit life support beneficial or burdensome? Bringing empirical data to the debate. Am J Respir Crit Care Med, 2009; 180:320–325; doi: 10.1164/rccm.200811-1776OC

36.

Chang

, Neville

, Parrish

, et al. Evaluation of time-limited trials among critically ill patients with advanced medical illnesses and reduction of nonbeneficial ICU treatments. JAMA Intern Med, 2021; doi: 10.1001/jamainternmed.2021.1000

37.

Mehter

, McCannon

, Clark

, et al. Physician approaches to conflict with families surrounding end-of-life decision-making in the intensive care unit. A qualitative study. Ann Am Thorac Soc, 2018; 15:241–249; doi: 10.1513/AnnalsATS.201702-105OC

38.

Ganguli

, Chittenden

, Jackson

, et al. Survey on clinician perceptions and practices regarding goals of care conversations. J Palliat Med, 2016; 19:1215–1217; doi: 10.1089/jpm.2015.0424

39.

Wiener

, Kim

SYI

, Fried

, et al. Early changes in rates of documented goals-of-care conversations in the veterans health administration during the COVID-19 pandemic. J Gen Intern Med, 2021; 36:1466–1469; doi: 10.1007/s11606-021-06652-5

40.

McGilton

, Krassikova

, Boscart

, et al. Nurse practitioners rising to the challenge during the coronavirus disease 2019 pandemic in long-term care homes. Gerontologist, 2021; 61:615–623; doi: 10.1093/geront/gnab030

41.

Fulmer

, Escobedo

, Berman

, et al. Physicians' views on advance care planning and end-of-life care conversations. J Am Geriatr Soc, 2018; 66:1201–1205; doi: 10.1111/jgs.15374

42.

Burt

, Woodwell

. Tests of methods to improve response to physician surveys. 2006.

43.

de Koning

, Egiz

, Kotecha

, et al. Survey fatigue during the COVID-19 pandemic: An analysis of neurosurgery survey response rates. Front Surg, 2021; 8:690680; doi: 10.3389/fsurg.2021.690680

44.

Davern

. Nonresponse rates are a problematic indicator of nonresponse bias in survey research. Health Serv Res, 2013; 48:905–912; doi: 10.1111/1475-6773.12070

45.

Bidassie

, Kovach

, Vallette

, et al. Breast cancer risk assessment and chemoprevention use among veterans affairs primary care providers: A national online survey. Mil Med, 2020; 185:512–518; doi: 10.1093/milmed/usz291

46.

Arigo

, Hooker

, Funderburk

, et al. Provider and staff perceptions of veterans' attrition from a national primary care weight management program. Prim Health Care Res Dev, 2015; 16:147–156; doi: 10.1017/s1463423614000139

47.

Martinez

, Gordon

, Tully

, et al. A mixed-methods study of veterans affairs providers' experiences communicating with patients about the affordable care act. Mil Med, 2017; 182:e1715–e1723; doi: 10.7205/milmed-d-16-00354

48.

Apaydin

, Rose

, McClean

, et al. Association between care coordination tasks with non-VA community care and VA PCP burnout: An analysis of a national, cross-sectional survey. BMC Health Serv Res, 2021; 21:809; doi: 10.1186/s12913-021-06769-7

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB

0.02 MB