Abstract
Background:
In 2018, the Centers for Medicare and Medicaid Services (CMS) mandated that patients considering implantation of an implantable cardioverter-defibrillator (ICD) for primary prevention of sudden cardiac death undergo shared decision-making (SDM) using a decision-aid.
Objective:
To observe the impact of the CMS’s mandate on core measures of SDM using a natural experiment.
Research Design, Subjects, and Measures:
Patients who underwent implantation of a primary prevention ICD within the Emory Healthcare system between 2017-2019 (pre and post SDM mandate) were surveyed. Survey domains included knowledge about the ICD, decisional conflict, values-choice concordance, and engagement in decision-making. Patients who had an ICD implant after the mandate were also asked about their views of the decision aid. Responses of patients who had ICD implanted prior to the mandate were compared to those after the mandate using either student’s t-test or chi-squared tests.
Results:
Of 101 patients who completed the survey, 45 had an ICD placed before the mandate and 56 had an ICD placed after. There were no major differences between knowledge, decisional conflict, values choice concordance, or patient engagement. Compared to patients with ICDs placed before the mandate, patients with ICDs after the mandate were more likely to subjectively feel more informed about the benefits of the procedure but were less likely to be able to correctly identify the frequency of complications.
Conclusions:
Policy effects to promote SDM that solely focus on a decision-aid may not substantively impact patient centered care.
Introduction
Implantable cardioverter-defibrillators (ICD) effectively prevent sudden cardiac death in high-risk populations. However, ICD implantation may not be right for some eligible patients. ICDs do not enhance quality-of-life and entail risks of procedural complications or inappropriate shocks. Many patients also have a poor understanding of important risks and benefits of ICD implantation and tend to overestimate benefits and underestimate risks.1–3 In February 2018, the Centers of Medicare and Medicaid Services (CMS) mandated that physicians conduct and document a shared decision-making interaction using a patient decision aid (DA) for individuals referred for ICD implantation for primary prevention of sudden cardiac death.4
This mandate has been controversial5 for several reasons. First, many argue that a decision with such a positive balance of benefits in terms of mortality balanced with minimal procedural risks should not even be considered in an SDM framework. ICD implantation improves mortality, is guideline-recommended, and is a relatively safe outpatient procedure. Although the rate of long-term complications may be higher,6 procedural complications are rare.7 Second, CMS did not clarify how a DA should be used or what the intended outcome of the mandate really was. Third, while SDM may be conceptually appropriate for enhancing patient-centered care, the data on the impact of formal SDM processes for this decision were minimal at the time of the mandate.
Three years after enactment of this mandate, there remain questions about its real-world impact on patients, clinicians, and decisions. We utilized the natural experiment created by the SDM mandate to better understand how mandated DA use has impacted relevant domains of patients’ experiences and decisions.
Methods
Using the natural experiment created by the CMS mandate in February 2018, we surveyed patients who underwent initial implantation of a single- or dual-chamber ICD for primary prevention of sudden cardiac death between January 1, 2017, and December 31, 2019 within the Emory Healthcare system. Importantly, during this period, there were no other institutional changes to patient education. Surveys were administered by a single individual (E.A) via telephone between September 2020 and February 2021. Completion of the survey was voluntary, patients were compensated for their time, and the study was approved by the Emory University Institutional Review Board.
The pre-mandate group consisted of patients who had an ICD implanted prior to February 2018; the post-mandate group were patients with ICDs implanted after February 2018. The SDM interaction was performed by the cardiac electrophysiologist responsible for performing the ICD implantation procedure. The same DA (Appendix 1) was used at all sites.8 For outpatients, which account for 80-90% of encounters, the DA was provided at a preoperative clinic visit. The way the DA was incorporated into SDM was at the discretion of the treating physician.
The survey instrument was adapted from instruments that have been developed and validated for use in a multicenter trial evaluating SDM for ICD implantation (ClinicalTrials.gov identifier: NCT03374891). Survey questions explored several key SDM domains: knowledge about ICDs, decision conflict9, values choice concordance, and patients’ decision engagement. Post-mandate patients were asked about their experience with the patient DA. Age at implantation, history of ischemic or non-ischemic cardiomyopathy, and pre-procedural left ventricular ejection fraction were obtained from medical records.
Comparison of continuous baseline variables was performed with the Students’ t-test; categorical variables of survey questions were compared using either chi squared or Fisher’s exact tests. All analysis was done using SAS 9.4 (Cary, NC).
Results
A total of 369 patients with primary prevention ICD implants were identified, 191 pre-mandate and 186 post-mandate (Appendix 2). Of these, 100 patients were excluded due to invalid contact information (86) and death or disability (14). Of the remaining 269, 28 declined to participate, 140 did not respond, and 101 completed the survey (response rate of 38%). Forty-five patients had an ICD implanted pre-mandate, and 56 underwent implantation post-mandate (Table 1). Non-response rates were similar between both groups. Comparison of non-responders to responders demonstrated similar age at implant, but non-responders were more likely to be male (62% vs 39.6%, p=0.0001).
Table 1.
Demographic information
| Overall (n=101) n (%) | Pre-SDM mandate (n =45) n (%) | Post-SDM mandate (n=56) n (%) | P-value | |
|---|---|---|---|---|
| Age | ||||
| < 65 years | 51 (50.50) | 21 (46.67) | 30 (53.57) | 0.4903 |
| ≥ 65 years | 50 (49.50) | 24 (53.33) | 26 (46.43) | |
| Gender | ||||
| Female | 61 (60.40) | 17 (37.78) | 23 (58.93) | 0.7366 |
| Male | 40 (39.60) | 28 (62.22) | 33 (41.07) | |
| Race | ||||
| Asian (East Asian or South Asian) | 2 (1.98) | 1 (2.22) | 1 (1.79) | 0.7549 |
| Black or African American | 46 (45.54) | 18 (40.00) | 28 (50.00) | |
| Hispanic/Latino(a) | 1 (0.99) | 0 (0) | 1 (1.79) | |
| White/Caucasian | 49 (48.51) | 25 (55.56) | 24 (42.86) | |
| Other | 3 (2.97) | 1 (2.22) | 2 (3.57) | |
| Education | ||||
| Some high school | 3 (2.97) | 3 (6.67) | 0 (0) | 0.4182 |
| Graduated from high school | 22 (21.78) | 8 (17.78) | 14 (25.00) | |
| Some college | 21 (20.79) | 8 (17.78) | 13 (23.21) | |
| Graduated from college | 32 (31.68) | 14 (31.11) | 18 (32.14) | |
| Some graduate school | 5 (4.95) | 3 (6.67) | 2 (3.57) | |
| Graduated from a graduate school | 18 (17.82) | 9 (20.00) | 9 (16.07) | |
| In general, would you say your health is: | ||||
| Excellent | 4 (3.96) | 1 (2.22) | 3 (5.36) | 0.1857 |
| Very good | 19 (18.81) | 10 (22.22) | 9 (16.07) | |
| Good | 37 (36.63) | 15 (33.33) | 22 (39.29) | |
| Fair | 25 (24.75) | 15 (33.33) | 10 (17.86) | |
| Poor | 16 (15.84) | 4 (8.89) | 12 (21.43) | |
| Left Ventricular Ejection Fraction | ||||
| ≤35% | 81 (80.20) | 34 (75.56) | 47 (83.93) | 0.2939 |
| >35% | 20 (19.80) | 11 (24.44) | 9 (16.07) | |
| Etiology of cardiomyopathy | ||||
| Ischemic | 37 (36.63) | 16 (35.56) | 21 (37.50) | 0.8402 |
| Non-ischemic | 64 (63.37) | 29 (64.44) | 35 (62.50) | |
| Has your ICD ever shocked you? | ||||
| Yes | 9 (8.91) | 3 (6.67) | 6 (10.71) | 0.4779 |
| No | 92 (91.09) | 42 (93.33) | 50 (89.29) |
There were no significant differences observed in composite knowledge about ICDs, decision conflict, values-choice concordance, or engagement in the decision-making process pre- and post-mandate. Overall, percentage of correctly answered questions regarding basic knowledge about ICDs (Table 2) was low in pre- and post- mandate participants (57.4% vs 53.0%, p=0.245). Participants implanted pre-mandate were more likely to correctly identify the frequency of minor complications (66.7% vs 37.5%, p=0.012). Mean decision conflict was low in both groups (11.89 vs 6.96, p=0.0877). Individual components of the decision conflict scale (Table 3) were similar, though pre-mandate patients were less likely to report an understanding of the benefits of ICD (88.2% vs 94.6%, p=0.0456). Answers to questions about values and engagement in decision-making were similar (Table 3).
Table 2.
Patients’ responses to questionnaire regarding their knowledge about ICDs
| Overall (n=101) n (%) | Pre-SDM mandate (n =45) n (%) | Post-SDM mandate (n=56) n (%) | P-value | |
|---|---|---|---|---|
| What is the primary purpose of an ICD | ||||
| To prevent sudden cardiac death* | 73 (72.28) | 34 (75.56) | 39 (69.64) | 0.3402 |
| To prevent heart failure | 17 (16.83) | 5 (11.11) | 12 (21.43) | |
| To improve heart failure symptoms, such as shortness of breath and leg swelling | 11 (10.89) | 6 (13.33) | 5 (8.93) | |
| How often should an ICD be checked by the doctor? | ||||
| Several times a year* | 87 (86.14) | 39 (86.67) | 48 (85.71) | 1.0000 |
| Every 5 years | 13 (12.87) | 6 (13.33) | 7 (12.50) | |
| The ICD never needs to be checked | 1 (0.99) | 0 (0) | 1 (1.79) | |
| If someone decides that they no longer want their ICD on and they have it turned off, what happens next? | ||||
| They have to enroll in hospice. | 2 (1.98) | 0 (0) | 2 (3.57) | 0.5928 |
| They will die shortly. | 11 (10.89) | 4 (8.89) | 7 (12.50) | |
| Nothing needs to be done, they can continue to receive medications and other treatments for their heart failure. * | 88 (87.13) | 41 (91.11) | 47 (83.93) | |
| When would it be recommended that an ICD be turned off? | ||||
| While you are asleep | 4 (3.96) | 3 (6.67) | 1 (1.79) | 0.5195 |
| Near the end of life* | 25 (24.75) | 11 (24.44) | 14 (25.00) | |
| It should never be turned off | 72 (71.29) | 31 (68.89) | 41 (73.21) | |
| If 100 people have an ICD, about how many will have their lives saved by the ICD over the next 5 years? Select the answer that is closest to correct. | ||||
| 0 | 0 (0) | 0 (0) | 0 (0) | 0.4865 |
| 1 | 5 (4.95) | 3 (6.67) | 2 (3.57) | |
| 7* | 13 (12.87) | 4 (8.89) | 9 (16.07) | |
| 28 | 24 (23.76) | 13 (28.89) | 11 (19.64) | |
| 53 | 59 (58.42) | 25 (55.56) | 34 (60.71) | |
| Minor complications can happen with an ICD such as bleeding or the device becoming disconnected. If 100 people get an ICD, about how many of them will experience a minor complication as a result of their ICD surgery? | ||||
| 0-1 | 22 (21.78) | 6 (13.33) | 16 (28.57) | 0.0116 |
| 2-5* | 51 (50.50) | 30 (66.67) | 21 (37.50) | |
| 6-10 | 23 (22.77) | 6 (13.33) | 17 (30.36) | |
| More than 10 | 5 (4.95) | 3 (6.67) | 2 (3.57) | |
| Serious complications can happen with an ICD such as infections, lung collapse, or death. If 100 people get an ICD, how many of them will experience a serious complication as a result of their ICD surgery? | ||||
| 0-1* | 51 (50.50) | 25 (55.56) | 26 (46.43) | 0.6658 |
| 2-5 | 33 (32.67) | 14 (31.11) | 19 (33.93) | |
| 6-10 | 12 (11.88) | 5 (11.11) | 7 (12.50) | |
| More than 10 | 5 (4.95) | 1 (2.22) | 4 (7.14) | |
| The shocks from an ICD cause some patients to develop emotional problems. | ||||
| Yes* | 45 (44.55) | 17 (37.78) | 28 (50.00) | 0.2193 |
| No | 56 (55.45) | 28 (62.22) | 28 (50.00) | |
| People with an ICD will not feel it if their ICD shocks them. | ||||
| Yes | 12 (11.88) | 7 (15.56) | 5 (8.93) | 0.3063 |
| No* | 89 (88.12) | 38 (84.44) | 51 (91.07) | |
| Sometimes an ICD will not shock a person when a shock is needed. | ||||
| Yes* | 41 (40.59) | 22 (48.89) | 19 (33.93) | 0.1281 |
| No | 60 (59.41) | 23 (51.11) | 37 (66.07) | |
| Sometimes an ICD will shock a person when a shock is not needed. | ||||
| Yes* | 47 (46.53) | 23 (51.11) | 24 (42.86) | 0.4085 |
| No | 54 (53.47) | 22 (48.89) | 32 (57.14) | |
| Percent of questions correctly answered | ||||
| Mean (SD) | 55.0 (19.1) | 57.4 (19.5) | 53.0 (18.7) | 0.2452 |
Table 3.
Responses to 10-item decision conflict scale and decision participation questions
| Overall (n=101) n(%) | Pre-SDM mandate (n=45) n(%) | Post-SDM mandate (n=56) n(%) | P-value | |
|---|---|---|---|---|
| If you were able to choose how to live the rest of your life, what number on the scale from 1 to 10 would represent where you would want to be (1 represents living as long as possible, even if symptoms from heart failure or other illnesses worsen over time and 10 represents dying quickly from any cause– for example, dying in your sleep). | ||||
| Mean (SD) | 4.61 (3.52) | 4.35 (3.11) | 4.82 (3.82) | 0.5108 |
| Did you know which options were available to you? | ||||
| Yes | 76 (75.25) | 34 (75.56) | 42 (75.00) | 0.0706 |
| Unsure | 14 (13.86) | 9 (20.00) | 5 (8.93) | |
| No | 11 (10.89) | 2 (4.44) | 9 (16.07) | |
| Did you know the benefits of each option? | ||||
| Yes | 90 (89.11) | 37 (82.22) | 53 (94.64) | 0.0456 |
| Unsure | 4 (3.96) | 4 (8.89) | 0 (0) | |
| No | 7 (6.93) | 4 (8.89) | 3 (5.36) | |
| Did you know the risks and side effects of each option? | ||||
| Yes | 81 (80.20) | 32 (71.11) | 49 (87.50) | 0.0554 |
| Unsure | 7 (6.93) | 6 (13.33) | 1 (1.79) | |
| No | 13 (12.87) | 7 (15.56) | 6 (10.71) | |
| Were you clear about which benefits mattered most to you? | ||||
| Yes | 90 (89.11) | 39 (86.67) | 51 (91.07) | 0.4821 |
| Unsure | 4 (3.96) | 3 (6.67) | 1 (1.79) | |
| No | 7 (6.93) | 3 (6.67) | 4 (7.14) | |
| Were you clear about which risks and side effects mattered most to you? | ||||
| Yes | 80 (79.21) | 35 (77.78) | 45 (80.36) | 0.3204 |
| Unsure | 9 (8.91) | 6 (13.33) | 3 (5.36) | |
| No | 12 (11.88) | 4 (8.89) | 8 (14.29) | |
| Did you have enough support from others to make the choice? | ||||
| Yes | 88 (87.13) | 38 (84.44) | 50 (89.29) | 0.7315 |
| Unsure | 4 (3.96) | 2 (4.44) | 2 (3.57) | |
| No | 9 (8.91) | 5 (11.11) | 4 (7.14) | |
| Did you make the choice without pressure from others? | ||||
| Yes | 87 (86.14) | 40 (88.89) | 47 (83.93) | 0.7706 |
| Unsure | 2 (1.98) | 1 (2.22) | 1 (1.79) | |
| No | 12 (11.88) | 4 (8.89) | 8 (14.29) | |
| Did you have enough advice to make the choice? | ||||
| Yes | 96 (95.05 ) | 43 (95.56) | 53 (94.64) | 0.0777 |
| Unsure | 2 (1.98) | 2 (4.44) | 0 (0) | |
| No | 3 (2.97) | 0 (0) | 3 (5.36) | |
| Were you clear about the best choice for you? | ||||
| Yes | 94 (93.07) | 41 (91.11) | 53 (94.64) | 0.4911 |
| Unsure | 4 (3.96) | 3 (6.67) | 1 (1.79) | |
| No | 3 (2.97) | 1 (2.22) | 2 (3.57) | |
| Did you feel sure about what to choose? | ||||
| Yes | 95 (94.06) | 42 (93.33) | 53 (94.64) | 0.2866 |
| Unsure | 2 (1.98) | 2 (4.44) | 0 (0) | |
| No | 4 (3.96) | 3 (2.22) | 1 (5.36) | |
| Composite Decision Conflict Score | ||||
| Mean (SD) | 9.16 (13.82) | 11.89 (16.25) | 6.96 (11.19) | 0.0877 |
| How much do you feel that you and your doctor shared in the decision to have an ICD implanted? | ||||
| The final decision was entirely mine | 8 (7.92) | 3 (6.67) | 5 (8.93) | 0.7010 |
| Mostly my decision, with some input from the doctor | 22 (21.78) | 11 (24.44) | 11 (19.64) | |
| My doctor and I jointly made the final decision | 48 (47.52) | 23 (51.11) | 25 (44.64) | |
| Mostly my doctor’s decision, with some input from me | 16 (15.84) | 6 (13.33) | 10 (17.86) | |
| The final decision was made entirely by my doctor | 3 (2.97) | 0 (0) | 3 (5.36) | |
| I don’t remember | 4 (3.96) | 2 (4.44) | 2 (3.57) | |
| How much did you and your doctor talk about the reasons to get an ICD? | ||||
| A lot | 75 (74.26) | 29 (64.44) | 46 (82.14) | 0.1025 |
| A little bit | 22 (21.78) | 14 (31.11) | 8 (14.29) | |
| Not at all | 0 (0) | 0 (0) | 0 (0) | |
| I don’t remember | 4 (3.96) | 2 (4.44) | 2 (3.57) | |
| How much did you and your doctor talk about the reasons not to get an ICD? | ||||
| A lot | 25 (24.75) | 13 (28.89) | 12 (21.43) | 0.6443 |
| A little bit | 29 (28.71) | 11 (24.44) | 18 (32.14) | |
| Not at all | 32 (31.68) | 13 (28.89) | 19 (33.93) | |
| I don’t remember | 15 (14.85) | 8 (17.78) | 7 (12.50) | |
| Did any of your doctors ask you if you wanted an ICD? | ||||
| Yes | 69 (68.32) | 31 (68.89) | 38 (67.86) | 0.9118 |
| No | 32 (31.68) | 14 (31.11) | 18 (32.14) |
Of the 56 post-mandate patients, 39 (69.6%) remembered receiving the decision-aid. Of those who remembered the decision-aid, 36 (92.3%) reported reading the decision-aid prior to the procedure, 35 (97.2%) felt it helped them feel more comfortable with their decision, and 34 (94.4%) felt it helped improve discussions with their doctor.
Discussion
These observational data capitalize on a natural history experiment to gain insights into the impact of CMS-mandated SDM with a DA for primary prevention ICDs. This mandate does not seem to have substantially impacted patients’ knowledge about ICDs, decision conflict, values-choice concordance, or engagement in decision-making, but patients liked the DA and may feel more informed.
SDM can advance patient centered care, but this study raises concerns about policies focusing on DAs alone. First, these data demonstrate no obvious impact DA on key SDM domains. Simply mandating DA use may not substantially change the encounter. Without understanding how SDM should be conducted and how best to utilize the DA, the CMS mandate risks incentivizing perfunctory interactions. Importantly, a pilot trial of this decision-aid did suggest potential improvements, but multiple types of materials were provided, all in advance of a clinical encounter.10 Second, SDM is context specific; effective strategies likely vary based on the nature of the decision. For chronically ill patients considering an ICD, the salient issue is whether they have a quality of life that they want to preserve. Although our participants liked the DA, it is unclear whether it really helped them to make this assessment.
These findings suggest that it is critically important to study different implementation strategies for SDM in this context and evaluate various constructions of DAs. These studies could clarify which processes are most impactful, and future recommendations could incorporate implementation guidance.
This study has limitations. This is an observational study within a single health system, but patients were seen at 3 separate hospitals. Additionally, the structure of the SDM interaction was not standardized, though this likely reflects real heterogeneity in practice and the broad nature of the mandate. Other components of clinical encounters likely impact the extent to which effective SDM occurs. Also, there is potential for nonresponse bias, though non-responders revealed a similar distribution of age and timing of ICD implants. Finally, recall bias is possible though not suggested by these data; knowledge regarding short-term procedural risks, for example, was higher in the pre-mandate group.
SDM is appropriate for patients undergoing primary prevention ICD implantation, but a mandate focusing only on DA use may be insufficient to impact patient-centered care.
Supplementary Material
Funding Sources:
Ten Broeke Family Foundation
Conflict of Interests:
Dr. Rao reports receiving research funding from NIH grant # UL1TR002378, #TL1TR002382, AHRQ grant # 1F32HS028558-01 and from the Byron Williams Jr, MD Fellowship Fund.
Dr. Dickert reports receiving research funding from AHRQ, NIH, PCORI, and the Greenwall Foundation.
Footnotes
Publisher's Disclaimer: This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi:10.1111/pace.14414.
Ethics statement: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee
Patient consent statement: All patients provided informed consent prior to participation in this study.
Data availability:
Data can be made available if requested
References
- 1.Lewis KB, Stacey D and Matlock DD. Making decisions about implantable cardioverter-defibrillators from implantation to end of life: an integrative review of patients’ perspectives. Patient. 2014;7:243–60. [DOI] [PubMed] [Google Scholar]
- 2.Groarke J, Beirne A, Buckley U, O’Dwyer E, Sugrue D, Keelan T, O’Neill J, Galvin J and Mahon N. Deficiencies in patients’ comprehension of implantable cardioverter defibrillator therapy. Pacing Clin Electrophysiol. 2012;35:1097–102. [DOI] [PubMed] [Google Scholar]
- 3.Stewart GC, Weintraub JR, Pratibhu PP, Semigran MJ, Camuso JM, Brooks K, Tsang SW, Anello MS, Nguyen VT, Lewis EF, Nohria A, Desai AS, Givertz MM and Stevenson LW. Patient expectations from implantable defibrillators to prevent death in heart failure. J Card Fail. 2010;16:106–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jensen TSCJ, Ashby L, Dolan D, Canos D, Hutter J. National coverage determination for implantable cardioverter defibrillators. [Google Scholar]
- 5.Merchant FM, Dickert NW Jr. and Howard DH. Mandatory Shared Decision Making by the Centers for Medicare & Medicaid Services for Cardiovascular Procedures and Other Tests. JAMA. 2018;320:641–642. [DOI] [PubMed] [Google Scholar]
- 6.Ranasinghe I, Parzynski CS, Freeman JV, Dreyer RP, Ross JS, Akar JG, Krumholz HM and Curtis JP. Long-Term Risk for Device-Related Complications and Reoperations After Implantable Cardioverter-Defibrillator Implantation: An Observational Cohort Study. Ann Intern Med. 2016;165:20–29. [DOI] [PubMed] [Google Scholar]
- 7.Freeman JV, Wang Y, Curtis JP, Heidenreich PA and Hlatky MA. Physician procedure volume and complications of cardioverter-defibrillator implantation. Circulation. 2012;125:57–64. [DOI] [PubMed] [Google Scholar]
- 8.Matlock DD VD, Jenkins A, Mellis K, Vaorsy P, Masoudi F, Brega A, Magid D. A Decision Aid for Implantable Cardioverter-Debrillators (ICD) For patients with heart failure considering an ICD who are at risk for sudden cardiac death (primary prevention). . [Google Scholar]
- 9.O’Connor AM. User Manual - Decision Conflict Scare (10 item question format). Ottawa Hospital Research Institute; 1993:16. [Google Scholar]
- 10.Wallace BC, Jones J, Masoudi FA, Nowels CT, Varosy P, Thomson R, Elwyn G, Brega AG, Vermilye T, Knoepke CE, Sandhu A, Allen LA and Matlock DD. Development and piloting of four decision aids for implantable cardioverter-defibrillators in different media formats. Pacing Clin Electrophysiol. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data can be made available if requested