A System to Classify Treatable Traits in Primary Care

In respiratory diseases, the presenting symptoms are often a combination of cough, dyspnoea and wheeze. These three symptoms can be present in a significant number of conditions, including airways disease, cardiac disease and lung parenchymal disease. Making an accurate and timely clinical diagnosis is a challenge. Furthermore, physical deconditioning and co-morbidities such as obesity often create further obstacles to diagnosis. Even in the context of a clinical diagnosis of airways disease, differentiating between asthma and COPD is often not a straightforward decision. It takes time to establish by evaluating a patient's symptoms and major risk factors like smoking or allergy suggesting a particular aetiology. The diagnosis is further refined by spirometry or measures of FeNO when available and, importantly, the patient's response to treatment (1).

There are several practical problems that make this approach less than accurate. Symptoms do not correlate with airflow limitation because co-exiting conditions like obesity and deconditioning and complicating factors such as anxiety and poor recall make symptom-based diagnosis imperfect (2). Diagnostic testing with spirometry is impractical and only provides a snapshot of lung function. This test relies on disease activity being present at the time of testing appointment. This feature is uncommon given the intermittent nature of symptoms in asthma. It is not uncommon for clinicians trying to interpret a set of lung function to hear a patient say something like "I am fine now, but I was awful two weeks ago". This means that people are often incorrectly labelled as having or not having asthma. A landmark study showed that incorrect labelling of people as having asthma but this could not be proven objectively in over 30% of patients assessed in a national study in Canada (3). The measurement of airway inflammation with FeNO is inaccurate unless treatment use is measured concurrently (4). Because of these practical problems with testing, clinicians often have to rely on symptoms to make the diagnosis of asthma.

Furthermore, in the context of an accurately established obstructive airway disease, practical issues persist. For example, differentiating between asthma and COPD, decision on referral to a secondary centre, tailoring treatment and determining if disease is controlled. Tailoring inhaled therapy to the individual patient is a further complex decision in this patient cohort. However, poor adherence to ICS/LABA treatment is common, on average it is less than 50% among patients in primary care (5-7). Therefore, the diagnosis of airways disease in primary care is inherently inaccurate. This inaccuracy means that decisions on treatment effectiveness are also inaccurate. These common but important limitations lead to overuse of corticosteroids, antibiotics and beta-agonists with poor symptom control potential medication related morbidity. It follows that diagnostic accuracy and appropriate inhaled treatment use in airways disease has real and significant implications for patient safety, adverse outcomes, cost and waste.

To address these problems the INCA team have developed algorithms to classify and align lung function, treatment use and symptoms. This data is delivered via a novel CE marked platform to non-specialists with specific "suggestion scripts" (8). The classification divides patients into 4 main groups based on whether the airway function is or is not controlled if the patient took their treatment and if they remain symptomatic (9, 10). These groups are;

1. The treatment intensify group: these people have persistent airflow obstruction/wide diurnal variation with good adherence. This trait requires additional bronchodilator and/or anti-inflammatory therapy and further tests.
2. The obstructed, but poorly adherent group: These people have persistent airflow obstruction and poor adherence to treatment or poor inhaler technique. Adherence is assessed based on a novel metric time above threshold. Management/treatment is therefore focused on promoting-adherence and correct inhaler technique.
3. The co-existing other condition group: This group have respiratory symptoms but no evidence of airflow obstruction. For these people airways disease is not the predominant problem and so should be evaluated with further tests and/or treatment of co-existing conditions.
4. Controlled airflow and controlled symptoms group: this group need to continue their current inhaled treatment.

This classification accounts for the common issues of poor adherence and inaccurate diagnosis in asthma, which are reported to occur as commonly as 50% and 30% respectively. The classification also accounts for some more nuanced issues that would arise if a clinician were to rely simply on assessing adherence or lung function. These include confirming that people with uncontrolled asthma have been adherent and on the other hand, even if poorly adherent, that controlled patients do not need advice on extra adherence. This classification may help a clinician to deliver a personalised, accurate and efficient consultation to people with asthma in primary care.

In this proposal the investigators will test the feasibility of this approach among patients attending community general practitioners, those newly referred for assessment in secondary care and those advanced nurse practitioner respiratory clinics with a physician's clinical diagnosis of asthma, more specifically those who have chronic respiratory symptoms who their treating physician believe to be due to asthma, but who have not yet undergone laboratory lung function testing to establish the correct diagnosis of asthma.

Hypothesis

The investigators hypothesize that the simultaneous measurement and alignment of inhaler use and airflow can be used to organise patients with clinical diagnoses of asthma into one of four groups described above. This classification can subsequently be used as the basis for adjustments to treatment and further diagnostic testing as needed. The investigators will assess the value of this approach in primary care by comparing it with the standard approach of using point in time measures with FeNO and spirometry coupled with self-report.