Abstract
Background:
Current guidelines recommend against placement of implantable cardioverter defibrillators in patients with a life expectancy less than one year. These patients may benefit from early palliative care services; however, identifying this population is challenging.
Objective:
Determine whether a validated prognostic tool, based on patient factors and health care utilization from electronic medical records, accurately predicts one-year mortality at the time of implantable cardioverter defibrillator placement.
Design:
We used the United States (U.S.) Veterans Administration's “Care Assessment Needs” one-Year Mortality Score to identify patients at high risk of mortality (score ≥95) before their procedure. Data were extracted from the Corporate Data Warehouse. Logistic regression was used to assess the odds of mortality at different score levels.
Setting/Subjects:
Patients undergoing a new implantable cardioverter defibrillator procedure between October 1, 2015 and September 30, 2017 in the U.S. Veterans Administration.
Results:
Of 3194 patients with a new implantable cardioverter defibrillator placed, 657 (21.8%) had a score ≥95. The mortality rate among these patients was 151/657 (22.9%) compared with 281/3194 (8.8%) for all patients undergoing a new implantable cardioverter defibrillator procedure. Patients with a score ≥95 had 14.0 (95% confidence interval 8.0–24.4) higher odds of death within one year of the procedure compared with those with a score ≤60.
Conclusions:
The “Care Assessment Needs” Score is a valid predictor of one-year mortality following implantable cardioverter defibrillator procedures. Integrating its use into the management of Veterans Administration (VA) patients considering implantable cardioverter defibrillators may improve shared decision making and engagement with palliative care.
Keywords: decision making, heart assist device, mortality, palliative care, risk assessment, Veterans Health Services
Introduction
Implantable cardioverter defibrillators (ICDs) greatly reduce mortality in patients with selected cardiac conditions.1–4 However, although life-saving in many patients, ICDs are costly, can be painful and traumatic when they are triggered,5,6 and appear to be less effective in patients with multiple comorbidities.7 Current evidence-based clinical guidelines recommend against placement of ICDs in patients with a life expectancy of less than one year, as risks of the procedure and side effects of the device outweigh benefits.4,8 Patients who have severe heart conditions or other comorbidities that limit their functional status and a high risk of near-term mortality may benefit from supportive and quality of life-driven cardiology, palliative care, and/or hospice care, rather than invasive procedures, including ICD placement9; however, integrating palliative care into heart disease management is still evolving compared with palliative care's more established role in some other serious advanced illnesses, particularly cancer.10–12
Despite guidelines suggesting limited benefit in patients with high near-term mortality, applying these recommendations in clinical practice is challenging due to limitations in clinicians’ ability to accurately predict mortality. Bedside or gestalt estimates of life expectancy have limited predictive value; for example, the so-called “surprise question”—would the physician be surprised if the patient died within the next 6, 12, or 18 months—is a poor predictor of one-year mortality in patients without a cancer diagnosis.13 Instead of relying on clinician impression alone, mortality risk prediction tools can assist with bedside medical decisions and counseling. Existing prognostic tools, including the Multicenter Automatic Defibrillator Implementation Trial II risk score, Kramer's one-year mortality score and Bilchick's risk factors reliably predicted mortality outcomes in ICD patients14–17; however, there is little evidence that these mortality risk prediction tools have been used in shared decision making regarding ICD use.18
Qualitative research with primary care physicians and cardiologists documents their confusion about currently available mortality risk prediction tools and suggests that potentially inappropriate factors, including age alone, currently guide implantable cardioverter defibrillator treatment referral or placement.19–21
The limited use of appropriate mortality risk assessment and shared decision making may be due, in part, to the lack of point-of-care access to valid mortality risk information. The objective of our study was to evaluate whether the electronic “Care Assessment Need” (CAN) One-Year Mortality Risk Score,22 which has been validated in the VA in other patient populations,23 accurately identifies at-risk patients with heart disease. As this prognostic tool automatically calculates a mortality risk score from existing electronic health data, it can be scaled to point-of-care for cardiology providers.22 Thus, the score could be used to both inform shared decision making about the benefits and harms of this invasive procedure, and trigger consideration of earlier palliative care involvement.
Methods
We conducted a retrospective cohort study of VA patients undergoing a new ICD procedure between October 1, 2015 and September 30, 2017 with one year of follow-up through September 30, 2018. The VA Boston Institutional Review Board approved this research with a waiver of informed consent before data collection and analysis.
Data sources
The data used in our study were obtained from the VA electronic medical record-derived Corporate Data Warehouse, a single national data repository for the VA. Data extracted from the Warehouse included CAN one-Year Mortality Risk Score, Current Procedure Terminology (CPT)-coded procedures, inpatient stays and outpatient visits, and mortality.
Study sample
We defined our index cohort as those who underwent a first eligible ICD cardiac device procedure, whether a new generator (CPT 33249) or a new generator attached to existing epicardially placed leads (CPTs 33240, 33231, 33230), performed between October 1, 2015 and September 30, 2017. We excluded patients with any cardiac device procedures in the prior 10 years to ensure that our sample were patients undergoing their first ICD implantation. Replacement or deactivation of an ICD were outside the scope of our study.
CAN Score
The VA developed four risk prediction algorithms, the CANS, to assess the probability of 90-day or one-year mortality or hospital admission.23 Each CAN Score is updated weekly to assess probabilities for an individual Veteran compared with the overall VA patient sample. CAN Scores are recorded in the Data Warehouse and currently available in a dedicated table for primary care providers.23 It is possible to also link the scores with the Computerized Patient Record System (the VA electronic medical record) for more immediate point-of-care use. The CAN Scores algorithms include: patient age, sex, marital status, disability level, comorbidities, visit history (including the type of visit), admission history, laboratory results, and pharmacy use (including drug type and refills).22,24 The CAN Scores are intended to inform primary care providers about patient frailty and clinical care needs and utilization and are not quality or performance measure tools; they have been used in prior work to identify high-risk patients for greater primary care engagement.25 Every patient in our sample had a CAN One-Year Mortality Risk Score available in the VA Corporate Data Warehouse within one week of the ICD procedure.
Outcome
Our outcome of interest was whether the patient died within one year of their ICD procedure. The VA records the date of death in the patient table in the Corporate Data Warehouse.26
Covariates
While we are concerned with the relationship between CAN One-Year Mortality Risk Score and mortality in our sample of ICD patients, we recognize there may be factors relevant to this outcome that are not included in the CAN Score algorithm. These include procedure characteristics: whether the procedure was done directly following an emergency room visit or used existing cardiac leads. We also considered whether patients had a palliative care consult in the VA in the 90 days before CIED placement using previously established CPT codes and clinic visit names.27 Lastly, we considered the region of the facility where the ICD was placed to control for geographic practice patterns.
Cohort characteristics
While we did not need to identify patient characteristics or comorbidities for our analysis as these are already calculated in the CAN Score algorithm, we did collect this information to describe our cohort. Patient characteristics from the VA Data Warehouse included demographics and diagnostic data from visits and admissions. We calculated patient comorbidities, measured with the U.S. Agency for Healthcare Research and Quality Comorbidity Software.28
Analysis
We compared the median CAN One-Year Mortality Risk Score, as well as patient and procedure characteristics, for patients undergoing new ICD placement who did and did not die within one year of device placement using chi-square tests or t-tests as appropriate. To determine the relationship between CAN Score and one-year mortality, we estimated a logistic regression model controlling for the procedure (related to an emergency room visit or using existing cardiac leads) and facility (region) characteristics as these factors may be relevant specifically in our sample of VA patients undergoing ICD procedures. Patients who have an ICD placed emergently may be systemically different than patients with a planned ICD, and we ran a sensitivity analysis, excluding these patients to assess whether our results changed.
Results
Cohort characteristics
Our cohort included 3194 patients undergoing a new ICD procedure in 69 VA facilities, 281 (8.8%) of whom died within one year (Table 1). Median time to death was 172 days (interquartile range 173 days). There was no difference between survivors and patients who died by race, ethnicity, gender, or procedure characteristics. Patients who died within a year of their procedure were significantly older (71.6 years [standard deviation (SD) 9.5] vs. 67.5 years [SD 9.5], p < 0.0001), were less likely to be married (40.9% vs. 48.5%, p = 0.01), and had significantly higher rates of various comorbid conditions than those who survived. Approximately 75% of the sample had chronic heart disease, but the rate was 84% in patients who died within a year of their procedure. Overall, the patient cohort was 98% male and 73% White race, reflective of the VA patient population with heart disease.29 Only 40/3,194 patients (1.3%) had a palliative care visit within the VA during the 90 day period prior to their ICD procedure.
Table 1.
Patient and Procedure Characteristics of Implantable Cardioverter Defibrillator Procedures by One-Year Mortality
| Variables | ICD patients surviving >one year | ICD patients dying within one year |
|---|---|---|
| Total | 2913 | 281 (8.8%) |
| CAN Score | ||
| CAN one-year mortality risk score, median (IQR) | 75 (30) | 95 (13)*** |
| Very low mortality risk score, CAN 0–59 | 703 (24.1%) | 14 (5%)*** |
| Low mortality risk score, CAN 60–74 | 543 (18.6%) | 22 (7.8%)*** |
| Medium mortality risk score, CAN 75–84 | 447 (15.4%) | 26 (9.3%)* |
| High mortality risk score, CAN 85–94 | 674 (23.1%) | 68 (24.2%) |
| Very high mortality risk score, CAN 95–100 | 546 (18.7%) | 151 (53.7%)*** |
| Patient characteristics | ||
| Age (years) | 67.5 (9.5) | 71.6 (9.5)*** |
| Female gender | 60 (2.1%) | 6 (2.1%) |
| Married vs. widowed, divorced, or single | 1414 (48.5%) | 115 (40.9%)* |
| White race | 2109 (72.4%) | 211 (75.1%) |
| African American race | 627 (21.5%) | 57 (20.3%) |
| Race not identified or other | 177 (6.1%) | 13 (4.6%) |
| Hispanic ethnicity | 153 (5.3%) | 15 (5.3%) |
| 1–99% service-connected disability | 525 (18%) | 57 (20.3%) |
| 100% service-connected disability | 854 (29.3%) | 79 (28.1%) |
| Procedure characteristics | ||
| Palliative care visit in prior 90 days | 28 (1.0%) | 12 (4.3%)** |
| ICD procedure with existing cardiac leads | 694 (23.8%) | 75 (26.7%) |
| Emergent problem | 267 (9.2%) | 27 (9.6%) |
| Comorbidities | ||
| AIDS | 19 (0.7%) | 3 (1.1%) |
| Alcohol abuse | 274 (9.4%) | 35 (12.5%) |
| Deficiency anemias | 536 (18.4%) | 94 (33.5%)*** |
| Rheumatoid arthritis/collagen vascular diseases | 62 (2.1%) | 9 (3.2%) |
| Chronic blood loss anemia | 34 (1.2%) | 5 (1.8%) |
| Congestive heart failure | 2174 (74.63%) | 237 (84.34%)** |
| Chronic pulmonary disease | 849 (29.2%) | 121 (43.1%)*** |
| Primary or secondary hypercoagulable state | 146 (5%) | 23 (8.2%)* |
| Depression | 553 (19%) | 67 (23.8%)* |
| Diabetes without chronic complications | 1233 (42.3%) | 137 (48.8%)* |
| Diabetes with chronic complications | 827 (28.4%) | 99 (35.2%)* |
| Drug abuse | 173 (5.9%) | 20 (7.1%) |
| Hypertension | 2446 (84%) | 250 (89%)* |
| Hypothyroidism | 283 (9.7%) | 37 (13.2%) |
| Liver disease | 177 (6.1%) | 27 (9.6%)* |
| Lymphoma | 29 (1%) | 6 (2.1%) |
| Fluid and electrolyte disorders | 491 (16.9%) | 91 (32.4%)*** |
| Metastatic cancer | 17 (0.6%) | 4 (1.4%) |
| Other neurological disorders | 200 (6.9%) | 40 (14.2%)*** |
| Obesity | 657 (22.6%) | 62 (22.1%) |
| Paralysis | 65 (2.2%) | 13 (4.6%)* |
| Peripheral vascular disease | 488 (16.8%) | 78 (27.8%)*** |
| Psychoses | 191 (6.6%) | 22 (7.8%) |
| Pulmonary circulation disorders | 127 (4.4%) | 22 (7.8%)* |
| Renal failure | 656 (22.5%) | 112 (39.9%)*** |
| Solid tumor without metastasis | 169 (5.8%) | 33 (11.7%)*** |
| Peptic ulcer disease (excl. bleeding) | 20 (0.7%) | 7 (2.5%)* |
| Valvular disease | 479 (16.4%) | 65 (23.1%)* |
| Weight loss | 122 (4.2%) | 29 (10.3%)*** |
| Geographic region of VA facility | ||
| Northeast | 331 (11.4%) | 38 (13.5%) |
| Midwest | 747 (25.6%) | 52 (18.5%)* |
| South | 1330 (45.7%) | 143 (50.9%) |
| West | 505 (17.3%) | 48 (17.1%) |
Data from 3194 new ICD procedures performed in VA surgical facilities between October 1, 2015 and September 30, 2017. Differences were assessed with chi-square tests for categorical variables or two-tailed t-tests for continuous variables between patients with or without one-year mortality.
p-Values signified by the following symbols: *<0.05; **<0.001; ***<0.0001.
CAN Score incorporates age, sex, comorbidities, and prior utilization.
Comorbidity categories from U.S. AHRQ comorbidity software.38
AHRQ, Agency for Healthcare Research and Quality; CAN, care assessment need; ICDs, implantable cardioverter defibrillators; IQR, interquartile range; VA, Veterans Administration.
Proportion of cases at risk of mortality and one-year mortality rates
We found ICD placement procedures performed in 657 (21.8%) patients in our sample with a high likelihood of one-year mortality risk (CAN Score ≥95). The mortality rate in this group was 151/657 (22.9%) compared with patients with the lowest mortality risk prediction (CAN Score ≤60) who had a mortality rate of 14/717 (1.9%). Among the 40 patients with a VA palliative care visit, the mean CAN Score was 87.8 (SD 14.0) compared to a CAN Score of 73.3 (SD 22.1, p = 0.0007) in patients who did not see palliative care; 19/40 (47.5%) had CAN Score ≥95.
We found no significant relationship between covariates for procedure characteristics in our adjusted model predicting mortality given the CAN One-Year Mortality Risk Score quintile; however, the odds ratio of mortality in one year was lower for facilities in the Midwest compared with the South (Table 2). The odds ratio of mortality was 14.0 (95% confidence interval 8.0–24.4) for patients with a CAN One-Year Mortality Risk Score ≥95 compared with patients with a score ≤60. The addition of procedure and facility characteristics did not alter the odds ratios of the CAN Score quintiles and the c-statistic in both adjusted and unadjusted models was 0.75. Excluding patients with emergent procedures (sensitivity analysis sample n = 2900) produced similar results (Appendix Table A1).
Table 2.
Results of Logistic Regression Predicting One-Year Mortality After VA Implantable Cardioverter Defibrillator Procedures
| Variables | Unadjusted model |
Adjusted model |
|---|---|---|
| OR (95% CI) | OR (95% CI) | |
| One-year CAN mortality risk score | ||
| Very low mortality risk score, CAN 0–59 | Ref | Ref |
| Low mortality risk score, CAN 60–74 | 2.0 (1.0–4.0) | 2.1 (1–4.1) |
| Medium mortality risk score, CAN 75–84 | 2.9 (1.5–5.7) | 3 (1.5–5.8) |
| High mortality risk score, CAN 85–94 | 5.1 (2.8–9.1) | 5.2 (2.9–9.3) |
| Very high mortality risk score, CAN 95–100 | 13.9 (7.9–24.3) | 14 (8–24.4) |
| Procedure characteristics | ||
| ICD procedure with existing cardiac leads | 1.2 (0.9–1.6) | |
| Emergent problem | 0.9 (0.6–1.4) | |
| Facility characteristics | ||
| South | Ref | |
| Northeast | 0.9 (0.6–1.4) | |
| Midwest | 0.6 (0.5–0.9) | |
| West | 0.8 (0.6–1.2) | |
| Model performance | ||
| Intercept MLE (SE) | −3.9 (0.3)* | −3.8 (0.3)* |
| C-statistic | 0.735 | 0.745 |
Of the 3194 VA ICD procedures from October 1, 2015 to September 30, 2017 in the sample, 281 (8.8%) died within one year.
p < 0.0001.
Bold indicates significance (p > 0.05).
CI, confidence interval; MLE, maximum likelihood estimate; OR, odds ratio; SE, standard error.
Discussion
In our large, U.S. national sample from the VA, we found that one in five patients undergoing new ICD procedures had a high predicted one-year mortality risk using the CAN prediction tool. Furthermore, the odds of dying within a year after the procedure was nearly 14 times higher among patients with a high CAN Score than those with a low CAN Score. We found evidence that the CAN Score accurately predicts one-year mortality and that many patients with a high risk of near-term mortality still underwent ICD placement procedures; VA palliative care engagement in these patients was limited (<1% of cases), including among patients in the highest CANS Score category (15/697, 2.2%). Our findings suggest that there may be opportunities to use prognostic tools, like the CAN Score, to inform shared decision making about this invasive procedure and trigger consideration of earlier palliative care involvement.
Clinical guidelines suggest that patients with high near-term mortality risk do not benefit from invasive cardiac device procedures.4,8 However, validated ICD mortality prediction scores are rarely used by providers to assess patient risk in a shared decision-making conversation.18 Using a validated electronic mortality risk prediction tool calculated weekly for VA patients,22 we identified high rates of ICD placements in heart disease patients who may not be good candidates for these procedures given their mortality risk. A separate observational study of VA patients undergoing new and replacement ICD procedures between 2007 and 2015 found a one-year mortality rate of 13%; the authors concluded this rate is higher than clinical trials where patients are carefully screened for mortality risk.30 Our findings suggest that clinicians in the VA, as in non-VA settings, may not use established ICD mortality prediction scores in treatment decision making, potentially due to the challenges of calculating a patient-specific mortality risk score with the existing options.
An accurate prognostic tool has the potential to inform patient-centered shared decision making about ICD placement. A shared decision-making framework has two key components: (1) clinicians share personalized medical information with a patient (and their family), which would include personalized discussion of prognosis and (2) clinicians elicit a patient's “big picture” life goals and treatment preferences.31 Ideally at the conclusion of a shared decision-making conversation, the clinician makes a treatment recommendation that is directly informed by the patient's goals and preferences.32 A literature review of shared decision making in ICD procedures highlighted the importance of accurate risk stratification.33 Our results demonstrate that a CAN One-Year Mortality Risk Score ≥95 accurately identifies patients at risk of near-term mortality following ICD placement.
A separate validation found a CAN One-Year Mortality Risk Score ≥95, which was indicative of patient frailty, highlights the potential for CAN Scores to play a larger role in patient care.34 Operationally, if the CAN One-Year Mortality Risk Score was available to specialists at the point of care, it could inform personalized discussions of prognosis as part of a patient-centered shared decision-making process about whether to pursue placement of an ICD.
Furthermore, given our findings about one-year mortality, the CAN One-Year Mortality Risk Score could also identify patients who might benefit from a palliative care referral, which may improve quality of life in this population. Patients with heart disease and a poor prognosis are often not identified, and thus rarely receive palliative care referrals before their last month of life.35 Only 1% of patients in our sample had a palliative care appointment before their procedure, highlighting this important gap in palliative care engagement. Patients with ICD devices also have low rates of advance directives.36 This represents an opportunity to improve quality-of-life outcomes, and to initiate discussions about whether invasive procedures, such as ICD placements, are within the patient's goals of care.11,37 A systematic review found evidence that palliative care was beneficial for patients with advanced heart failure, but the authors found inconsistencies in how these patients were identified.35 A reliable, automated, point-of-care prognostic tool such as the CAN One-Year Mortality Risk Score may provide significant value for patients considering ICD placement.
Several limitations should be considered when interpreting these results. The study was confined to the U.S. patients treated at the VA and included mostly White men, which could limit the generalizability of the results outside the veteran population. In addition, this was a retrospective, observational cohort study, and we were unable to assess the clinician's decision-making process that led to the device procedures. We could not determine whether risk assessment methods may have been used in the decision to refer the patient to cardiology or whether patient preferences may have driven device placements, although we observed only 1% of patients saw palliative care. There is a need for future qualitative research about how mortality prognostic calculators, such as the CAN One-Year Mortality Risk Score, influence shared decisions about undergoing new ICD placement procedures.
Conclusion
Our research makes an important contribution to patient-centered shared decision making. A CAN One-Year Mortality Risk Score ≥95 is a strong predictor of mortality following ICD procedures and, as an automated point-of-care score, could increase the use of appropriate mortality risk assessment data in shared decision making for VA patients considering a cardiac device.
Appendix Table A1.
Results of Logistic Regression Predicting One-Year Mortality After VA Implantable Cardioverter Defibrillator Procedures—Excluding Emergent Procedures
| Variables | Unadjusted model |
Adjusted model |
|---|---|---|
| OR (95% CI) | OR (95% CI) | |
| One-year CAN mortality risk score | ||
| Very low mortality risk score, CAN 0–59 | Ref | Ref |
| Low mortality risk score, CAN 60–74 | 1.8 (0.9–3.8) | 1.8 (0.9–3.8) |
| Medium mortality risk score, CAN 75–84 | 3.3 (1.7–6.7) | 3.4 (1.7–6.8) |
| High mortality risk score, CAN 85–94 | 5.5 (2.9–10.2) | 5.5 (3–10.4) |
| Very high mortality risk score, CAN 95–100 | 15.4 (8.5–28.2) | 15.4 (8.4–28.2) |
| Procedure characteristics | ||
| ICD procedure with existing cardiac leads | 1.2 (0.9–1.6) | |
| Facility characteristics | ||
| South | ||
| Northeast | 1 (0.7–1.5) | |
| Midwest | 0.7 (0.5–1) | |
| West | 0.8 (0.6–1.2) | |
| Model performance | ||
| Intercept MLE (SE) | −4 (0.3)* | −3.9 (0.3)* |
| C-statistic | 0.744 | 0.753 |
Of the 2900 nonemergent VA ICD procedures from October 1, 2015 to September 30, 2017 in the sample, 254 (8.8%) died within one year.
p < 0.0001.
Bold indicates statistical significance (p < 0.05).
CAN, care assessment need; CI, confidence interval; ICD, implantable cardioverter defibrillator; MLE, maximum likelihood estimate; OR, odds ratio; SE, standard error; VA, Veterans Administration.
Authorship Confirmation Statement
E.K. and H.J.M. affirm that the article is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained. She attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. H.J.M. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Data Management and Sharing
The data that support the findings of this study are not publicly available due to U.S. federal data policies; however, the authors can provide limited deidentified data upon reasonable request and with permission of the VA Boston Healthcare System Data Security and Privacy office.
Funding Information
This work was support by the VA Health Services Research and Development Service (PO 18-031). W.B.E. is supported by NHLBI 1K12HL138049-01.
Author Disclosure Statement
W.B.E. is an expert witness for DLA Piper, LLC. All other authors have no conflicts of interest to report. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the U.S. Department of Veterans Affairs, Montclair State University, the University of Iowa, Dana Farber Cancer Institute, Harvard Medical School, or Boston University.
References
- 1. Cleland JG, Daubert J-C, Erdmann E, et al. : The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med 2005;352:1539–1549. [DOI] [PubMed] [Google Scholar]
- 2. Connolly SJ, Hallstrom AP, Cappato R, et al. : Meta-analysis of the implantable cardioverter defibrillator secondary prevention trials. AVID, CASH and CIDS studies. Antiarrhythmics vs Implantable Defibrillator study. Cardiac Arrest Study Hamburg. Canadian Implantable Defibrillator Study. Eur Heart J 2000;21:2071–2078. [DOI] [PubMed] [Google Scholar]
- 3. Wever EF, Hauer RN, van Capelle FL, et al. : Randomized study of implantable defibrillator as first-choice therapy versus conventional strategy in postinfarct sudden death survivors. Circulation 1995;912195–2203. [DOI] [PubMed] [Google Scholar]
- 4. Al-Khatib SM, Stevenson WG, Ackerman MJ, et al. : 2017 AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: A report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2018;72:e91–e220. [DOI] [PubMed] [Google Scholar]
- 5. Sweeney MO, Wathen MS, Volosin K, et al. : Appropriate and inappropriate ventricular therapies, quality of life, and mortality among primary and secondary prevention implantable cardioverter defibrillator patients. Circulation 2005;111:2898–2905. [DOI] [PubMed] [Google Scholar]
- 6. Daubert JP, Zareba W, Cannom DS, et al. : Inappropriate implantable cardioverter-defibrillator shocks in MADIT II: Frequency, mechanisms, predictors, and survival impact. J Am Coll Cardiol 2008;51:1357–1365. [DOI] [PubMed] [Google Scholar]
- 7. Chan PS, Nallamothu BK, Spertus JA, et al. : Impact of age and medical comorbidity on the effectiveness of implantable cardioverter-defibrillators for primary prevention. Circ Cardiovasc Qual Outcomes 2009;2:16–24. [DOI] [PubMed] [Google Scholar]
- 8. Ferrero P, Iacovoni A, D'Elia E, et al. : Prognostic scores in heart failure—Critical appraisal and practical use. Int J Cardiol 2015;188:1–9. [DOI] [PubMed] [Google Scholar]
- 9. Goodlin SJ, Hauptman PJ, Arnold R, et al. : Consensus statement: Palliative and supportive care in advanced heart failure. J Card Fail 2004;10:200–209. [DOI] [PubMed] [Google Scholar]
- 10. Bakitas M, Macmartin M, Trzepkowski K, et al. : Palliative care consultations for heart failure patients: How many, when, and why? J Card Fail 2013;19:193–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Liu AY, O'Riordan DL, Marks AK, et al. : A comparison of hospitalized patients with heart failure and cancer referred to palliative care. JAMA Netw Open 2020;3:e200020–e200020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Gelfman LP, Bakitas M, Warner Stevenson L, et al. : The state of the science on integrating palliative care in heart failure. J Palliat Med 2017;20:592–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Downar J, Goldman R, Pinto R, et al. : The “surprise question” for predicting death in seriously ill patients: A systematic review and meta-analysis. Canadian Med Assoc J 2017;189:E484–E493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Goldenberg I, Vyas AK, Hall WJ, et al. : Risk stratification for primary implantation of a cardioverter-defibrillator in patients with ischemic left ventricular dysfunction. J Am Coll Cardiol 2008;51:288–296. [DOI] [PubMed] [Google Scholar]
- 15. Barsheshet A, Moss AJ, Huang DT, et al. : Applicability of a risk score for prediction of the long-term (8-year) benefit of the implantable cardioverter-defibrillator. J Am Coll Cardiol 2012;59:2075–2079. [DOI] [PubMed] [Google Scholar]
- 16. Kramer DB, Friedman PA, Kallinen LM, et al. : Development and validation of a risk score to predict early mortality in recipients of implantable cardioverter-defibrillators. Heart Rhythm 2012;9:42–46. [DOI] [PubMed] [Google Scholar]
- 17. Bilchick KC, Stukenborg GJ, Kamath S, et al. : Prediction of mortality in clinical practice for medicare patients undergoing defibrillator implantation for primary prevention of sudden cardiac death. J Am Coll Cardiol 2012;60:1647–1655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Hess PL, Matlock DD, and Al-Khatib SM: Decision-making regarding primary prevention implantable cardioverter-defibrillators among older adults. Clin Cardiol 2020;43:187–195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Liew SM, Blacklock C, Hislop J, et al. : Cardiovascular risk scores: Qualitative study of how primary care practitioners understand and use them. Br J Gen Pract 2013;63:e401–e407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sherazi S, Zareba W, Daubert JP, et al. : Physicians’ knowledge and attitudes regarding implantable cardioverter-defibrillators. Cardiol J 2010;17:267–273. [PMC free article] [PubMed] [Google Scholar]
- 21. Sadarmin PP, Wong KC, Rajappan K, et al. : Questionnaire survey of cardiologists’ knowledge, attitudes, and guideline application of implantable cardioverter defibrillator therapy. Pacing Clin Electrophysiol 2012;35:672–680. [DOI] [PubMed] [Google Scholar]
- 22. Wang L, Porter B, Maynard C, et al. : Predicting risk of hospitalization or death among patients receiving primary care in the Veterans Health Administration. Med Care 2013;51:368–373. [DOI] [PubMed] [Google Scholar]
- 23. Fihn SD and Box TL: Care Assessment Need (CAN) Score and the Patient Care Assessment System (PCAS): Tools for Care Management. https://www.hsrd.research.va.gov/for_researchers/cyber_seminars/archives/713-notes.pdf (Last accessed January 12, 2020).
- 24. Zulman DM, Chang ET, Wong A, et al. : Effects of intensive primary care on high-need patient experiences: Survey findings from a veterans affairs randomized quality improvement trial. J Gen Intern Med 2019;34:75–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Chang ET, Asch SM, Eng J, et al. : What is the return on investment of caring for complex high-need, high-cost patients? J Gen Intern Med 2021;36:3541–3544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Office of Informatics and Analytics (OIA), Health Information Governance, Veterans Health Administration (VHA). CDW Date of Death Analysis and Vital Status File Comparison. www.vhadataportal.med.va.gov/Portals/0/DataQualityProgram/Reports/CDW_Date_of_Death_Analysis_and_VSF_Comparison_Update.pdf (Last accessed April 2, 2020).
- 27. Carpenter JG, McDarby M, Smith D, et al. : Associations between timing of palliative care consults and family evaluation of care for veterans who die in a hospice/palliative care unit. J Palliat Med 2017;20:745–751. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Healthcare Cost and Utilization Project (HCUP): AHRQ Comorbidity Software, Version 3.7. www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp (Last accessed May 13, 2013).
- 29. Holder KA: United States Census Bureau, Social, Economic and Housing Statistics Division: Who Are Veterans? https://www.census.gov/newsroom/blogs/random-samplings/2016/11/who_are_veterans.html (Last accessed December 1, 2021).
- 30. Fudim M, Carlisle MA, Devaraj S, et al. : One-year mortality after implantable cardioverter-defibrillator placement within the Veterans Affairs Health System. Eur J Heart Fail 2020;22:859–867. [DOI] [PubMed] [Google Scholar]
- 31. Charles C, Gafni A, and Whelan T: Shared decision-making in the medical encounter: What does it mean? (or it takes at least two to tango). Soc Sci Med 1997;44:681–692. [DOI] [PubMed] [Google Scholar]
- 32. Murray E, Charles C, and Gafni A: Shared decision-making in primary care: Tailoring the Charles et al.: Model to fit the context of general practice. Patient Educ Couns 2006;62:205–211. [DOI] [PubMed] [Google Scholar]
- 33. Lewis KB, Carroll SL, Birnie D, et al. : Incorporating patients’ preference diagnosis in implantable cardioverter defibrillator decision-making: A review of recent literature. Curr Opin Cardiol 2018;33:42–49. [DOI] [PubMed] [Google Scholar]
- 34. Ruiz JG, Priyadarshni S, Rahaman Z, et al. : Validation of an automatically generated screening score for frailty: The care assessment need (CAN) score. BMC Geriatr 2018;18:106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Datla S, Verberkt CA, Hoye A, et al. : Multi-disciplinary palliative care is effective in people with symptomatic heart failure: A systematic review and narrative synthesis. Palliat Med 2019;33:1003–1016. [DOI] [PubMed] [Google Scholar]
- 36. Merchant FM, Binney Z, Patel A, et al. : Prevalence, predictors, and outcomes of advance directives in implantable cardioverter-defibrillator recipients. Heart Rhythm 2017;14:830–836. [DOI] [PubMed] [Google Scholar]
- 37. von Schwarz ER, He M, and Bharadwaj P: Palliative care issues for patients with heart failure. JAMA Netw Open 2020;3:e200011–e200011. [DOI] [PubMed] [Google Scholar]
- 38. Healthcare Cost and Utilization Project (HCUP): Beta Elixhauser Comorbidity Software for ICD-10-CM. https://www.hcup-us.ahrq.gov/toolssoftware/comorbidityicd10/comorbidity_icd10.jsp (Last accessed November 25, 2017).