Only sites able to offer a comprehensive range of the currently available AS treatments will participate in the registry. A study nurse is employed at the centres specifically to monitor echocardiography results and to communicate with the referring physician (RP) in a structured manner.

To be eligible for inclusion in the registry, patients must be aged 18 or over and diagnosed with severe AS by echocardiography. Severe AS is defined as having an aortic valve meeting at least one of the following criteria: 1) AVA < 1 cm², 2) valve area index <0.6 cm/m², 3) Vmax > 4.0 m/s, 4) mean transvalvular gradient > 40 mm Hg. All patients with previous AVR, or AS not considered severe according to the aforementioned criteria, will be excluded from the registry. Patients will be excluded if they have cognitive impairment that may limit their understanding of the factors involved in study participation, including the consent process. Enrolment will be consecutive.

All data are entered into the eCRF (s4trials, Berlin, Germany) and signed by the study nurse at the earliest opportunity, using registry identification numbers rather than names to protect identity. Automatic checks for plausibility and completeness are performed, and all data sets will be examined for irregularities by the data manager.

For each patient, baseline characteristics (including symptoms, co-morbidities, frailty, and EuroSCORE values), echocardiogram findings, and contact details will be documented at the point of individual study enrolment (where enrolment is defined as the time of informed consent directly after echocardiography-based diagnosis of severe AS). In addition, symptom status, co-morbidity, frailty, and eligibility for different treatment options (pharmaceutical management, SAVR, or TAVI) will be collected by the study nurse.

The main follow-up will be performed for all patients via a structured telephone call or contact with the RP 3 months after the date of each individual enrolment (i.e. a patient diagnosed and enrolled on day 9 will be followed-up at 3 months + 9 days). Follow-up data will include vital status, emergency admissions, referrals, and treatment since enrolment, as well as the sub-specialty of the healthcare professional responsible for making such a decision. Vital status at one year will be also collected via chart review and contact of the RP.

For the subset of patients included in the interventional phase only (see definitions below), data regarding the vital status of a particular patient, and the treatment plans or referral intentions of the RP (active decision not to treat, watchful waiting, conservative management, balloon aortic valvuloplasty (BAV), SAVR, or TAVI) will be obtained 1 week after the date of each individual enrolment. Watchful waiting is defined as the scheduling of a further outpatient assessment, stress echocardiography, exercise testing, transoesophageal echocardiography, catheterisation, computed tomography, or referral to a valve service or other clinician. Conservative management is defined as medical therapy only.

Study duration will be 15 months at each participating site, starting from the point of first patient enrolment. This study period is divided into 3 distinct phases (Fig. 1) with 3 different patient groups, as follows:

The primary outcome measure is the rate of SAVR or TAVI by the RP at 3-month follow-up in each of the 3 phases. Secondary outcome measures are as follows: 1) the rate of referral for exercise testing or stress echocardiography by the RP at 3 month follow-up, 2) time to next out-patient appointment or clinical review, 3) type of treatment plan selected (active decision not to treat, watchful waiting, conservative management, BAV, SAVR, or TAVI), 4) the appropriateness of treatment decisions made in collaboration with the valve specialist, and 5) survival outcome one year post-diagnosis.

We estimated that we would be able to capture a mean of 100 consecutive patients per site, 25 during the first observational phase, 50 in the interventional phase and another 25 during the second observational phase. The final sample size estimate was based on the aim to involve 25 sites in different countries and health care systems across Europe and to ask them to enrol patients throughout a year.

Intention-to-treat analysis including all enrolled patients will be carried out. Descriptive summaries will be used to present the collected evaluation data. For categorical variables (e.g. gender) frequency distributions will be given. Statistical significance will be determined using a chi-squared or a Fisher’s exact test. For numeric variables (e.g. patient age) minimum, maximum, mean, median, and standard deviation will be calculated. Statistical significance will be determined using a student’s t-test in the case of normally distributed data, or a Mann–Whitney U-test for non-normally distributed data. Linearized rates and actuarial probability statistics may be used where appropriate for adverse event reporting. Kaplan-Meier analysis will be performed for survival outcomes and, where appropriate, for adverse event outcomes.

In the first observational phase, the appropriateness of treatment decisions for the management of severe AS in current European clinical practice will be evaluated, and the scale of any deficit quantified. This is following several previous studies that have revealed consistently low rates of appropriate management in a considerable proportion of AS patients. In 2003, the large-scale Euro Heart Survey showed that 31.8% of patients with severe, symptomatic, single-valvular heart disease did not receive an intervention, and that surgery was denied in one third of AS symptomatic patients [20]. Later, in 2010, Freed et al. reported from a small study in 106 patients, that 69% of severe AS patients were not referred for valve replacement, of which 42% were symptomatic and 21% had LVEF <50%; therefore making them eligible for consideration under ACC/AHA guidelines [18]. More recent data have reported similar failings in monitoring valve disease, with a large but retrospective series identifying that only 59% of patients with moderate-severe left-sided valve disease received follow-up echocardiography within 60 days of the time recommended in existing guidelines [17]. Therefore, a deficit between eligibility and adequate management has been documented consistently, regardless of period, country, disease severity, or available treatment.

A combination of patient and physician-based factors has been proposed to explain the gap between guidelines and clinical practice [17]. Clinicians in the Euro Heart Survey reported advanced age, severe left ventricular dysfunction, recent myocardial infarction, severe coronary disease, prior neurological disease, and a decrease in symptoms under medical treatment as justifications for not referring a patient for SAVR [20] but several of these potential reasons are highly debatable. Additionally, Freed et al. found that the lack of referral for SAVR or TAVI in symptomatic patients was mainly due to failure of the treating physician to attribute symptoms to AS (29%), followed by an elevated surgical risk (15%), despite 40% of the latter patients having low EuroSCORE parameters [18]. Furthermore, existing methods for determining surgical risk (such as the logistic EuroSCORE model and on-line STS Risk Calculator) have many pitfalls and limitations, and appear not to reliably predict the true likelihood of mortality in all patients; particularly those undergoing TAVI [21]. Thus, a RP with poor knowledge of the limitations of current assessment tools in valvular heart disease may erroneously prevent eligible patients from undergoing lifesaving intervention.

Taken together, these results suggest that appropriate medical attention may be delayed or withheld in a significant proportion of patients due to poor decisions made by RPs acting independently rather than in the environment of a ‘heart team’ discussion. Observational phase 1 of the present study will supply further insight into the validity of this assumption in contemporary practice, and provide data on the type of physician currently responsible for making treatment decisions. Furthermore, justifications for non-referral will be collected to identify variables associated with inadequate treatment decisions. This may be useful for educating physicians and improving appropriateness of future treatment.

Several different strategies for quality improvement have been suggested to enhance patient care. The need for a well-organised valve service to improve monitoring, evaluation, and timing of intervention for patients with AS has been highlighted in several recent publications [22‐24]. A key element of this valve service is to ensure effective, multidisciplinary communication between team members, including echocardiographers and clinicians, co-ordinators, interventional cardiologists, and cardiothoracic surgeons [23]. This should, in theory, maximise the number of patients receiving the appropriate therapy, increase referral rates for intervention and thereby minimise morbidity and mortality. However, with cost effectiveness always a key concern, healthcare institutions favour economical interventions with a clear business case [22].

A previous study by Taggu et al. described the positive outcomes of a quality of care intervention for patients with valvular disease, reporting that the agreement between clinical follow-up and best practice guidelines rose by 51% following the establishment of sonographer-led valve clinics for patients with stable valve disease [24]. This highlights the capacity for improvement in the care of patients with valvular disease, although main focus was on monitoring rather than on intervening. In the interventional phase of the present study, the effect of employing a single, dedicated echocardiography clinic nurse to notify the RP of a finding of severe AS will be evaluated. Outcomes such as referral rate will be compared to those seen in observational phase 1 to detect any improvements. This type of intervention via a phone call or other non-automated method is known as facilitated relay of patient data [25]. In a previous, large systematic review of quality improvement strategies in the care of hypertensive patients, implementation of simple facilitated data relay was shown to reduce systolic and diastolic blood pressure by a median of 8.0 and 1.8 mmHg, respectively, and result in 25.1% more patients attaining systolic blood pressure within their target range [26]. This suggests that such an intervention can have a substantial impact on patient care, though it has never been tested in the context of severe AS. Should the present study demonstrate significant improvements in the appropriateness of patient referral for timely and appropriate intervention under the facilitated data relay intervention, implementation would be simple and low-cost, making it an attractive strategy for improving quality of care.

For the third phase of the present study, the facilitated relay of diagnosis will be stopped and its legacy effect on patient management evaluated. This will indicate whether the intervention was efficient at changing the behaviour of the RPs and indicate the potential for short interventions to be effective in the longer-term.

Firstly, the largely observational nature of the IMPULSE study means that investigators will have little control over the characteristics of the severe AS patients enrolled. Considering that 3 distinct populations will be used for different phases of the study, this may result in an uneven distribution of treatment decisions between phase groups. Accordingly, the appropriateness of RP decisions may reflect mistakes due to more challenging patient variables, rather than active deviations from guidelines. However, this is unavoidable if a snapshot of real-world practice is to be obtained, and the high patient numbers planned for enrolment over a reasonable time period (minimum 3 months) should accommodate for such imbalances.

The study does not involve an investigation into the contribution of patient/family preferences to the treatment decision-making process. It will therefore not be possible to exclude this important variable from the assessment of the appropriateness of the RP’s decision. In the present analysis we will focus on evaluating guideline adherence, with future work potentially delving into the underlying reasons for RPs deviating from the recommendations.

The relatively short study duration means that guideline adherence will only be evaluated up to 3 months post-diagnosis. This means that long-term consequences of appropriate/inappropriate management are beyond the scope of the current study. Additionally, phase 3 will only be suitable for assessing the short-term legacy of the quality of care intervention, and further studies will be necessary to establish any longer-term effects. However, for providing an initial view into the efficacy of facilitated data relay to improve initial care decisions in severe AS patients, 3 month follow-up is sufficient.

Lastly, during the facilitated data relay intervention stage, the physician is not only informed of diagnosis outcomes, but also of their patient’s participation in a study. This may result in a feeling that their professional capacity is being evaluated, prompting more conscientious decision-making. Although clearly an advantage for study participants, this may result in over-estimation of the effect the quality of care intervention would have under real-world, non-study conditions. However, in order to comply with ethical guidelines, such a notification is unavoidable.