Health service use by community-dwelling patients with heart failure in northwest London, 2015–18: a retrospective cohort analysis

• Published in: The Lancet (British edition) Vol. 394; p. S64
• Main Authors: Kim, Dani, Hayhoe, Benedict, Aylin, Paul, Bottle, Alex
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 01-11-2019; Elsevier Limited
• Subjects: Cardiology; Clinical practice guidelines; Cluster analysis; Clustering; Cohort analysis; Congestive heart failure; Datasets; Electronic health records; Electronic medical records; Health care; Health services; Heart failure; Leadership; Mental health; Nurses; Nursing; Patients
• ISSN: 0140-6736; 1474-547X
• DOI: 10.1016/S0140-6736(19)32861-2

Heart failure management is complex and requires multidimensional care. Clinical guidelines advocate for the use of integrated care, but few studies have described overall health service use by patients with heart failure. We describe overall care usage by community-dwelling patients with heart failure in northwest London using a dataset linked across multiple health-care settings.
In this retrospective cohort analysis, we used Whole Systems Integrated Care (WSIC) data, a dataset of individual clinical events from primary care, secondary care, and other (defined as community, mental health, and social care) services in northwest London to analyse a retrospective cohort of patients with heart failure. Patients who remain in the WSIC dataset, were aged at least 18 years on Jan 1, 2015, with known age and gender, and whose diagnosis was not made after their date of death. The primary outcome was health-care usage of each of the three service categories overall as well as individual components of secondary care (eg, outpatient appointments) and other services (eg, nursing contacts). Usage patterns were identified using k-means cluster analysis using all distinct contacts for the whole study period and visualised by a heatmap. Nine health-care use variables were used to identify clusters, and differences between clusters in both clustering and non-clustering variables were analysed using Kruskal-Wallis and Pearson χ2 tests. This study was approved by the Discover Research Advisory Group.
Data from 39 301 patients with heart failure across 359 general practices between Jan 1, 2015, and Dec 31, 2018, were analysed. Almost all patients had some health service contact during the study period (21 226 [89·1%] of 23 828 in 2015; 24 876 [90·6%] of 27 443 in 2016; 28 663 [90·8%] of 31 554 in 2017; and 31 584 [91·1%] of 34 651 in 2018). Most commonly used services in 2018 were outpatient appointments (24 283 [70·1%] of 34 651 patients), general practitioner (GP) consultations (20 741 [59·9%] patients), unplanned accident and emergency visits (14 145 [40·8%] patients) and community services (13 762 [39·7%] patients), often for nursing. Use of cardiology-specific services ranged from 1005 (2·9%) patients for community visits to 8231 (23·8%) for outpatient visits. Five clusters of patients were identified, each with significantly different care usage patterns and distinct patient characteristics. Patterns included underuse of GP consultations, relative to the average use by the whole cohort, in patients living in more deprived areas; and increased use of other services in patients older than 75 years and those with three or more comorbidities.
Health and social care service use was widespread but varied. Patterns of care usage suggest differences across patient groups. This might be due to variation in patients’ access to care and insufficiently integrated care at earlier stages of heart failure. The limitations of our study are that electronic health records like WSIC are not coded for research and subject to some biases; k-means testing is sensitive to the input variables and selection of k; and our findings require further validation and might not be generalisable to regions beyond the northwest London area. Further research is needed to characterise the patterns we have identified.
Dr Foster (a Telstra Health company), the National Institute for Health Research (NIHR) Health Services Research, and the Imperial NIHR Biomedical Research Centre and the NIHR Collaboration for Leadership in Applied Health Research and Care for northwest London. The views expressed are those of the authors and not those of the NIHR.