KardiaMobile for detecting atrial fibrillation

The committee noted considerable evidence from 6 comparative studies, including 3 randomised controlled trials that showed improved atrial fibrillation (AF) detection using the single-lead KardiaMobile. The clinical experts agreed that monitoring with KardiaMobile could increase infrequent AF event detection because it could record an AF event whenever symptoms are presented. The external assessment centre (EAC) advised that the evidence base was strongest in people with undiagnosed palpitations and people with a history of AF who need to monitor their AF recurrence. The experts noted that, in clinical practice, KardiaMobile has been most commonly used in people presenting with palpitations for detecting infrequent AF events. The committee concluded that the population groups in the evidence base reflected the use of the KardiaMobile device in a wide range of relevant clinical contexts, but it considered the most persuasive clinical cases were symptomatic people with suspected paroxysmal AF.

The committee noted that all the evidence on the clinical effectiveness of KardiaMobile was on the single-lead device. It was advised that the single-lead device is commonly used in clinical practice to detect AF, and the use of the 6‑lead device is limited in the NHS. The clinical experts agreed that the 6‑lead device provides heart rhythm from multiple angles, and it could have incremental benefits in some people to detect other arrhythmias when a good quality electrocardiogram (ECG) trace is available. For AF detection, the committee and experts concluded that the single-lead device is suitable in most patients.

The clinical experts advised that AF is a chronic condition. After AF is diagnosed, people are likely to be on medications such as anticoagulation or rhythm control drugs to reduce the risk of stroke and control symptoms for a long time. The experts noted that KardiaMobile could improve medication management; for instance, some medications can only be used when the normal heart rhythm is restored. However, no direct evidence was available on the clinical benefits of KardiaMobile after diagnosis of AF. For example, there was no evidence for the association between early AF detection and reduction in longer-term outcomes such as stroke events. The committee understood the limitations of the evidence base and concluded that more research would be of value that explores the use of KardiaMobile after AF diagnosis, including the impact of using KardiaMobile on clinical outcomes.

Evidence from published studies and patient experts shows that KardiaMobile is easier to use compared with other ECG monitors such as the Holter monitor. People with experience of using KardiaMobile found the device to be accessible at symptom onset and that it allowed improved access to care when needed. It has been well accepted across people of different age groups if they have a compatible mobile device. The committee concluded that KardiaMobile is a convenient device that people can use at home to monitor their heart rhythms.

Patient selection is important and should be guided by clinical judgement. The clinical experts emphasised that devices need to be offered to people on an individual basis guided by clinical assessment of individual circumstances. Key factors to consider include risk of developing AF, age, comorbidities, and the availability of primary and secondary care resources to interpret ECG traces. Furthermore, other factors such as the compatibility of mobile devices and patient preference also need to be considered. They noted that widespread use of KardiaMobile in the NHS without careful patient selection may place extra demand on local services. The committee concluded that healthcare professionals should assess individuals and indications when considering whether to prescribe KardiaMobile.

The committee noted that one of KardiaMobile's advantages over some other technologies is that it is a portable device that provides real-time ECG traces and heart rhythm classification. Despite this, the EAC confirmed that clinical interpretation of all recorded ECGs is needed, in line with the device instructions for use, to limit the effect of false negative and false positive results. The clinical experts added that expertise in interpreting ECG traces is essential to ensure the accuracy of AF diagnosis. Also, the experts added that a considerable proportion of unreadable and unclassified ECG recordings would be interpretable by experienced healthcare professionals to inform clinical decision making. The committee concluded that ECG data generated by KardiaMobile should be reviewed by an experienced healthcare professional before a diagnosis is made.

People need a smart device compatible with the KardiaMobile app. The company provides a list of compatible smart devices. The clinical experts said that in their experience, most people have access to a smart device, and alternative ambulatory monitors are offered if a person does not have a compatible device. The clinical experts also noted that KardiaMobile may not be suitable for people who cannot stay still or have problems holding the device; for example, people with tremors may have difficulty with recording an accurate trace.

The clinical experts highlighted the issue of unreadable ECG traces. They explained that the way people use the device is likely to affect the quality of ECG recordings. In clinical practice, healthcare professionals often provide support for people to set up the device, allowing them to also advise on effective use. There are also self-help videos that explain how to use the device. The clinical experts noted that a lack of experience using the device may lead to unreadable ECG recordings. The committee concluded that training is important to make sure people use the device correctly and minimise possible interference while taking the recording.

The committee understood that the company's original cost model was complex. The model was simplified during consultation. Data from 6 comparative studies was used to estimate the cost impact using KardiaMobile for 2 separate cohorts: 1) undiagnosed people presenting with palpitations and 2) people with previously diagnosed AF at risk of AF recurrence. The committee noted that the results from the company's cost model showed that KardiaMobile is cost saving for each cohort compared with Holter monitor over a 5‑year time horizon. The committee considered that some assumptions and parameters, such as 100% diagnostic accuracy for KardiaMobile, may be unlikely or may not fully reflect clinical practice. The committee concluded that the company's modified model was relevant to the decision problem, but still featured limitations, and considered the EAC's additional cost modelling a more appropriate basis for its decision making.

The committee felt that the EAC's original approach using a cost calculator did not capture the cost impact of using KardiaMobile. The EAC was therefore asked to develop a new cost model, to evaluate using KardiaMobile to detect AF in people presenting with undiagnosed palpitations and people being monitored for AF recurrence. From these further analyses, the committee accepted that KardiaMobile is slightly cost saving compared with standard care for patients with undiagnosed AF (people with symptomatic palpitations). This saving is driven by a reduction in diagnostic costs because the cost of KardiaMobile is lower than that of Holter monitor. The clinical experts explained that the benefits of KardiaMobile in this population are likely to be realised because KardiaMobile is better at detecting infrequent AF events compared with Holter monitor. The committee concluded that there are likely to be cost benefits using KardiaMobile in symptomatic patients with suspected paroxysmal AF.

For people who need to monitor AF recurrence, including those who have had an ischaemic stroke or a transient ischaemic attack without current evidence of AF, the results of the EAC's analysis show that KardiaMobile is likely to be cost incurring. This was driven by an increase in using anticoagulation for preventing strokes. The committee understood that the model included only a small selection of patients who are at low risk of developing strokes (CHA2DS2‑VAScC score of 1) and that medication was the only intervention included in the model. Expert advice suggested that clinical management of patients being monitored for AF recurrence is complex and varied widely because of patients' comorbidities and their medical and treatment history. In clinical practice, there is no clear care pathway for monitoring AF recurrence. The committee accepted the limitations of the current model and concluded that more information is needed about the care pathway for AF recurrence monitoring and its associated resource use.