The insertion of an implantable cardioverter defibrillator (ICD) is an established therapy for the prevention of sudden cardiac death (Epstein, 2008; Jackson and Murphy 2017). The first ICDs were implanted in the 1980s and since then the number of ICD implantations has risen steadily and will continue to rise (Fluur et al, 2013). An ICD has lifesaving capacity and provides physical and psychological ‘backup’ for a person who has sustained life-threatening cardiac dysrhythmias (Hill et al, 2016). Because of the immediacy of treatment provided by ICDs and the consequent prolongation of life, people who would have been susceptible to premature death now live longer. For many of these individuals, their natural end of life comes from a deterioration in their health or from illness (Resuscitation Council (UK) et al, 2015). The provision of a dignified death is important for all individuals, but specific considerations are warranted for those with ICDs.
Towards the end of life the number and frequency of life-threatening dysrhythmias can naturally increase. The most recent literature suggests that up to one-third of patients receive an ICD shock in their final 24 hours of life (Stoevelaar et al, 2018). A shock from an ICD causes pain, distress and anxiety, in addition to the potential prolongation of the dying experience (Lampert et al, 2010; Thylen et al, 2016). Consequently, patients may prefer it if their ICDs were deactivated at the end of life. However, not everyone is aware that this option is available to them. Therefore, it is incumbent on health professionals to talk to patients about their options, and it is preferable to have these discussions in the early stages of treatment (Lampert et al, 2010; Fluur et al, 2013). Recent research suggests that this is not always translated into clinical practice (Kirkpatrick et al, 2012).
Context
The recommended procedure for deactivating an ICD involves consent from the patient and written instruction from the physician (Lampert et al, 2010). The Heart Rhythm Society (HRS) has endorsed proactive communication by clinicians, with patients, about ICD management (Boriani et al, 2019). The adoption of Advance Directives (AD) for patients with ICDs is encouraged, thereby eliminating the dilemmas that surrogate decision-makers face if a patient is unable to communicate their own preferences (Boriani et al, 2019). Although these recommendations provide a guide to assist practitioners in clinical practice, it may be difficult to determine a one-size fits-all approach to such a sensitive issue. Brady (2016) suggested that ICD deactivation be discussed prior to implantation and that triggers for discussing ICD deactivation be identified. Triggers refer to opportunities for opening conversations about deactivation. They include situations such as frequent hospital readmissions, a notable functional decline in the patient, the aftermath of ICD shocks, admission to the hospice and decisions not to resuscitate. In addition, patients should be screened for the presence of ICDs on admission to hospital and education and support of these patients should involve the multidisciplinary team (Brady, 2016).
The Working to Improve Discussions about Defibrillator Management trial (Goldstein et al, 2014) concluded that challenges exist in relation to communication and device deactivation. Research on this topic has been conducted internationally and irrespective of nation, sample size or research methods, the results and recommendations from studies are broadly similar. Collectively, the studies suggest that patients are regularly unaware of the practical issues to be considered when dying with an ICD in situ and there is a need for further emphasis and research into this important phenomenon (Fluur et al, 2013; Lewis et al, 2014; Strömberg et al, 2014). The research also revealed that discussions about ICD deactivation at the time of implant occur as little as 4% of the time (Marinskis et al, 2010) and that patients' knowledge about ICD-related decisions was poor (Lewis et al, 2014). Furthermore, there is limited research available concerning the factors that may be associated with patient knowledge (McEvedy et al, 2018). The aim of this study was therefore to investigate patients' knowledge and opinions about their ICDs during life, illness and at the time of death and the factors that might influence these.
Methods
Study design and setting
This was a cross-sectional, correlational, non-experimental study. Data were collected between October 2016 and February 2017 in an ICD clinic in a large tertiary hospital in Ireland.
Sample
To be eligible for inclusion in this study, patients must have had an ICD implanted at least 12 months previously, be over 18 years of age and be able to read and write in English. Those with a cognitive impairment and patients awaiting an urgent heart transplant were considered ineligible, in addition to those who had participated in drug trial studies in the previous 12 weeks.
Procedures
A staff member from the research site posted batches of letters of invitation and participant information leaflets to eligible patients whose ICD clinic appointments were scheduled for the forthcoming weeks. Those interested in participating contacted the researcher by telephone.
Data were collected using a previously devised and validated 39-item questionnaire, the EOL-ICDQ (Thylén et al, 2014). There were 11 questions on knowledge. A correctly answered question was allocated a score of one and an answer that was incorrect or inconclusive (don't know) was allocated a score of zero. Scores ranged from 0 to 10 and were categorised into ≤4 = insufficient knowledge and ≥5 as sufficient knowledge. The research nurse completed the questionnaire with the patient in a face-to-face interview in a private room at the ICD clinic. The study was approved by the relevant ethics committees.
Data analysis
Data were analysed using SPSS version 23. Descriptive statistics were used to describe the knowledge and opinion profile of the sample in respect of ICDs. Chi square (χ2) analyses were used to determine associations between knowledge and age, time since implantation of the device, the presence of atrial fibrillation and a history of myocardial infarction. Variables were dichotomised for this analysis, as indicted in Table 1. A power analysis with a power of 75%, an alpha of 0.05 and a large effect size (0.5) indicated that a sample of 28 was required. A p value of <0.05 was considered significant. Missing data were not inputted.
| Demographic profile | n (%) | Knowledge good | Knowledge poor | χ2 | p | |
|---|---|---|---|---|---|---|
| Gender | Male | 24 (80) | 9 (38%) | 15 (62%) | 0.000 | 1.0 |
| Female | 6 (20) | 2 (33%) | 4 (66%) | |||
| Marital status | Not partnered | 17 (57) | 8 (47%) | 9 (53%) | 0.938 | 0.171 |
| Partnered | 13 (43) | 3 (23%) | 10 (77%) | |||
| Employment status | Employed/well | 15 (50) | 8 (53%) | 7 (47%) | 2.29 | 0.128 |
| Unemployed/sick | 15 (50) | 3 (20%) | 12 (80%) | |||
| Health insurance | No | 21 (70) | 15 (71%) | 6 (29%) | 0.225† | |
| Yes | 9 (30) | 4 (44%) | 5 (56%) | |||
| Age | < 65 | 15 (50) | 2 (13%) | 13 (87%) | 5.18 | 0.021* |
| 65 and over | 15 (50) | 9 (60%) | 6 (40 %) | |||
| Education | Less than 10 years | 6 (20%) | 4 (67%) | 2 (33%) | 1.00 | |
| 10 years or more | 24 (80%) | 15 (63%) | 9 (38%) |
Results
Thirty patients participated in the study. The sample average age was 61.8±12.6 years. The majority (80%) were male and 100% were white Caucasian. On average, in the total sample it was 5.24±2.79 years since their ICD had been implanted and one-third (33%) of participants had previously received a shock from their ICD. Using chi square (χ2) analyses, the authors examined the relationship between six sociodemographic factors and 20 clinical profile factors and knowledge of ICD (Table 1). Age was the only factor significantly associated with knowledge in this study (p <0.05); those who were over 65 years had greater knowledge of their ICDs.
The authors measured whether participants had engaged in discussions about their illness trajectory and ICD deactivation with their next of kin, doctor or nurse (Table 2). Only 13% (n=4) had spoken with their doctor or nurse regarding deactivation and 10% (n=3) had the conversation with their next of kin.
| n (%) | ||
|---|---|---|
| Years since ICD implant | >3 years | 19 (63) |
| ICD battery replaced | Yes | 6 (20) |
| ICD leads replaced | Yes | 4 (13) |
| Received shocks from ICD | No | 20 (67) |
| Discussed change of battery with doctor/nurse | Yes | 5 (17) |
| Discussed change of battery with next of kin | Yes | 8 (27) |
| Discussed turning off shocks with doctor/nurse | Yes | 4 (13) |
| Discussed turning off shocks with next of kin | Yes | 3 (10) |
| Discussed illness trajectory with doctor/nurse | Yes | 10 (33) |
| Discussed illness trajectory with next of kin | Yes | 10 (33) |
| Told next of kin wishes regarding shocks if seriously ill/fatal disease | Yes | 7 (23) |
| Considered at some point asking nurse to turn shocks off | Yes | 3 (10) |
| Have religious faith/outlook to help manage as an ICD recipient | Yes | 17 (57) |
| Often think about questions about end of life | Yes | 7 (23) |
| I wish to discuss end of life | Agree | 10 (33) |
| I will raise the question of end of life myself when I feel the need to | Agree | 25 (83) |
| I will raise the question of end of life in connection to ICD implantation | Agree | 21 (70) |
| I will raise the question about end of life if I have a shock | Agree | 18 (60) |
| I will raise the question about end of life if I have multiple shocks | Agree | 25 (83) |
| I will raise the question about end of life upon repeatedly been hospitalised with recurring heart problems | Agree | 26 (87) |
| I will raise the question about end of life if suffering from disease with poor prognosis | Agree | 28 (93) |
| I will raise the question about end of life routinely upon return visits to ICD clinic | Agree | 11 (37) |
| I will raise the question about end of life if my heart disease deteriorates | Agree | 28 (93) |
| I will raise the question about end of life towards the end of life | Agree | 29 (97) |
| Battery replaced even if not received shocks (n=28) | Yes | 27 (96) |
| Battery replaced even if seriously ill from another disease (n=24) | Yes | 10 (42) |
| Battery replaced even if at an advanced age (n=21) | Yes | 15 (71) |
| Keep shocks even if dying of cancer/other disease (n=24) | Yes | 1 (4) |
| Keep shocks even if shocked daily (n=25) | Yes | 7 (28) |
| When find self at end of life want to decide myself if shocks to be turned off (n=26) | Yes | 23 (88) |
| When find self at end of life want doctor to decide to turn off shocks (n=29) | Yes | 14 (48) |
| When find self at end of life want next of kin to decide if shocks are turned off | Yes | 11 (42) |
With regard to knowledge, the mean knowledge score of participants was 6.15±2.5, reflecting that 59.3% had sufficient ICD knowledge (score ≥5). Of the 11 questions posed (Table 3), four were correctly answered by at least 70% of participants, and 90% correctly answered that the heart does not stop beating when an ICD is deactivated. Lowest knowledge scores tended to be associated with device functionality (questions F, H, I and K, see Table 3).
| Knowledge statement | Distribution n (%) | ||
|---|---|---|---|
| False | True | Don't know | |
| A. In order to turn off the shocks in an ICD the ICD must be removed by surgery | 17 (57)* | 4 (13) | 9 (30) |
| B. When the ICDs shocks have been turned off they cannot be turned back on again | 19 (64)* | 3 (10) | 8 (26) |
| C. When the ICD shocks have been turned off the heart stops beating | 27 (90)* | 1 (3) | 2 (7) |
| D. An ICD shocks can be turned off without me or my next of kin's knowledge | 22 (73)* | 5 (17) | 3 (10) |
| E. To turn off the shock in an ICD is the same as active euthanasia | 21 (70)* | 4 (13) | 5 (17) |
| F. In order to turn off the shocks in an ICD, a temporary deactivation can be done via a magnet | 5 (17) | 16 (53)* | 9 (30) |
| G. In connection with the shock, the shock is also transferred to those persons having direct bodily contact | 17 (57)* | 10 (33) | 3 (10) |
| H. When an ICD battery voltage is beginning to wear even the ICD's functioning worsens | 14 (47)* | 9 (30) | 7 (23) |
| I. An ICD always delivers a shock in connection with the end of life | 15 (50)* | 8 (27) | 7 (23) |
| J. In order to turn off the ICD shocks a reprogramming can be done via an external programmer | 2 (7) | 21 (70)* | 7 (23) |
| K. When the ICD shocks are turned off the pacemaker function in the ICD is also deactivated | 7 (23)* | 14 (47) | 9 (30) |
Most participants (77%) felt that they would not wish to maintain an active ICD if they were receiving shocks daily or had a serious illness (Table 4). Chi square (χ2) analyses were used to examine the relationship between opinions and knowledge of ICD; no significant associations were found.
| I would like: | Yes n (%) | No n (%) | Don't known (%) |
|---|---|---|---|
| To keep the shocks in my ICD even if I am dying from cancer or another serious illness | 1 (3) | 23 (77) | 6 (20) |
| To keep the shocks in my ICD even if I am receiving shocks daily | 7 (23) | 18 (60) | 8 (17) |
| To decide for myself if the shock function should be turned off | 23 (77) | 3 (10) | 4 (13) |
| The doctor to decide if the shock function should be turned off | 23 (77) | 3 (10) | 4 (13) |
| My next of kin to decide if shocks are turned off | 11 (37) | 15 (50) | 4 (13) |
Discussion
For patients with inherited and acquired cardiac conditions, ICDs are vital for the reversal of shockable rhythms, and the subsequent prevention of sudden cardiac death (Beattie, 2013). Although patients are regularly counselled about the benefits of ICDs (Raphael et al, 2011), research suggests that patient information in respect of device management and considerations about deactivation during serious illness or at the end of life are limited (Russo, 2011; Kirkpatrick et al, 2012). This was also reflected in this study where 17% and 20% of patients could not offer an opinion when asked if they would deactivate their ICD if receiving shocks daily, or if they had a serious illness.
Given that ICD shocks can cause severe pain and that, in about 35% of cases, shockable ventricular rhythms occur in the last hour of life (Boriani et al, 2019), it is essential that patients are given opportunities to consider issues around palliation and that they are supported to make decisions in this regard. However, the tailoring of end-of-life care plans to individual patients is not a simple task. It requires sensitivity, experience and competence on behalf of the health professional, and employers must ensure that staff who undertake these roles are adequately skilled to do so (Resuscitation Council (UK) et al, 2015).
It is normally the health professional that initiates discussions about ICD management and deactivation (Resuscitation Council (UK) et al, 2015). However, it is worth noting that in this study a maximum of one-third of participants had discussed ICD management issues such as battery changes, illness trajectories and shock disablement with a nurse, doctor or next of kin. Of those that held these discussions, it was predominantly illness trajectories that were discussed. This may indicate a need to identify opportunities for discussions and a need for clear lines of communication across the spectrum of multidisciplinary team members.
The need for timely conversations regarding ICD deactivation has been well supported by guidelines and research (Sheehan et al, 2011; Lewis et al, 2014; Resuscitation Council (UK) et al, 2015; Brady, 2016; Hill et al, 2016; Boriani et al, 2019). These conversations should support an informed decision-making process that may potentially reduce the number of patients receiving shocks in their final days of life. Cardiology professional societies (Lampert et al, 2010; Resuscitation Council (UK) et al, 2015; Boriani et al, 2019) provide healthcare staff with algorithms and guides to assist in the management of withdrawal of ICD therapy and decision-making for people approaching end of life. These can help pave the way for staff who manage this difficult process.
The HRS and the professional guidelines (Resuscitation Council (UK) et al, 2015; Boriani et al, 2019) highlight that effective communication and clear documentation are essential components of critical care. Failure in relation to these is the main reason for complaint and litigation (Resuscitation Council (UK) et al, 2015). They also recommend that specific procedures be adhered to, such as consent and accurate documentation of conversations regarding ICD management and deactivation. As device recipients have been shown to rely on their doctor/nurse to make decisions, health professionals should initiate these conversations to help reduce patient burden (Sanchez-Toscano et al, 2015). Despite the availability of guidelines and strong recommendations, there is evidence to suggest a lack of adherence to best practice (Sheehan et al, 2011; Lewis et al, 2014; Resuscitation Council (UK) et al, 2015; Hill et al, 2016).
Early and effective communication is also important to reduce the stress patients may experience, particularly towards the end of life (Groarke et al, 2012; Hill et al, 2015). Ongoing education positively affects patients' levels of knowledge (Cinar et al, 2013). Sanchez-Toscano et al (2015) recommended that a continuous education programme be provided for patients. It should be commenced prior to device implantation and be progressed in accordance with the patient's needs and in advance of cognitive decline (Strachan et al, 2011). However, the challenge is to pinpoint when exactly medical treatment is no longer viable for a patient (Boriani et al, 2019), so that the timing is right to open these discussions. Some patients prefer to discuss ICD deactivation only when their health deteriorates (Thylén et al, 2014), while others prefer to avoid all such forms of discussion (Herman et al, 2013).
To date there has been no education programme identified for patients with ICDs in Ireland and this study has highlighted this need. Research and guidelines suggest that initial discussions about ICD deactivation should not be left until the person is nearing death; ideally it should occur before implantation. The need to further investigate how patients perceive dying and deactivation of their ICD devices is widely acknowledged (Sheehan et al, 2011; Goldstein et al, 2014; Lewis et al, 2014; Brady, 2016; Hill et al, 2016). Delaying communications about decision-making in the final days of patients' lives, is inappropriate, given the importance and implications of the decision.
Limitations and strengths
Of the 120 patients invited to participate, only 30 accepted the invitation. The female population is under-represented, with only 6 female participants. In addition, the study was confined to one research site. Despite this, the study was powered, robust, ethically sound and used previously validated and reliable research instruments.
The participation rate of only 25% in this study was low and considerably lower than in other similar studies. However, the authors contacted patients by post on only one occasion to invite them to participate in the study. It is recognised that mail methods of recruitment are less effective than personalised recruitment (Kelley et al, 2003) and that reminders are a useful way of increasing responses, but the authors considered it inappropriate to send repeat invitations, given the sensitive nature of the topic and the need to synchronise data collection with patients' pre-scheduled clinic appointments. Furthermore, the letters of invitation were sent from the hospital, but the hospital had no direct involvement with the study. It was therefore important that patients felt the hospital was providing impartial information, and not promoting involvement through issuing repeat invitations.
A further possible explanation for the low participation rate was related to a proviso from the ethics committee. To reduce patient burden, patients in this study were informed in the patient information letter that they would be ineligible to participate if they had been involved in a drug trial study within the previous 12 weeks. Although exact figures are unknown, this exclusion criterion may have reduced the number of patients who were potentially eligible to participate.
Conclusion
There are rapid advances in medical care and with that, increasing numbers of ICD devices implanted worldwide. Consequently, the number of ICD recipients will continue to grow in the next decade. It is incumbent on health professionals to ensure that, where acceptable to patients, they are fully educated about their ICDs. Furthermore, towards the end of life, open dialogue and support are fundamental to assist patients and families with decision-making in respect of ICD deactivation. In this study, knowledge levels among participants was somewhat reflective of other international studies and research suggests that patients who are well informed make better decisions. However, the sample was small and further research is warranted in this area. Nevertheless, based on this study and the literature, the authors conclude that for ICD recipients, there is scope to improve the adequacy of education and communication in respect to deactivating ICDs at the end of life.