Background The 2014 British Thoracic Society (BTS) and Scottish Intercollegiate Guideline Network (SIGN) guidelines recommend a treatment escalation/de-escalation approach to asthma management. This guideline is well established, but more evidence is needed on the prescribing patterns leading to treatment escalation/de-escalation, including how practice compares with BTS/SIGN guidelines.

Methods We conducted a retrospective, observational, longitudinal study of asthma patients treated in primary care using the Clinical Practice Research Datalink. Aims were to: classify patients according to BTS/SIGN guidelines; describe the proportion of patients transitioning between steps; describe patient demographics and clinical characteristics per group. Inclusion criteria were: patients of acceptable quality standard in CPRD; Read code for asthma diagnosis/treatment; age 12–80 years on 1 st June 2016; ≥1 year of data prior to 1 st June 2016. Patients taking part in asthma studies/clinical trials during the study period were excluded.

Results 6 07 212 patients with an asthma Read code were identified. The average age was 43 years and 55.1% were female. Most patients (31.4%–47.9% from study beginning to end) were in step 1 and received short-acting bronchodilators. 18.0%–20.1% of patients were in step 2, 16.2–23.2% in step 3, 0.9%–1.8% in step 4 and 0.8%–0.6% in step 5. 12.0–6.4% of patients were diagnosed but untreated. Most patients (80.1%–94.4% over the study duration) remained stable in a given treatment step in each 6 month interval. The proportion of patients stepping up reduced from the beginning of the study (10.5% in December 2007 to 2.2% in June 2016); by June 2016 more patients stepped down than up (3.7% vs 2.2%).

A small proportion of patients stepped up to the next consecutive stage of therapy (0.4%–5.9%), the smallest number moving from step 3 to 4 (table 1). A small proportion of patients stepped down to the next consecutive stage, but non-consecutive steps were not uncommon. The most common step down was from step 5–3 (12.7%).

Conclusion Most patients remained stable in a step throughout the study. Only a small proportion of patients moved through consecutive steps in the treatment pathway, as recommended by guidelines. Step 4 was underused, with more patients moving from step 3–5 or 5–3 than 3–4.