Aims: We aimed to identify subgroups in the patient population with different trajectories of long-term readmission rates. The study also aimed to assess common causes and their sequences of readmissions for each subgroup. Methods: Patients with a primary diagnosis of heart failure (HF) in the period 2008-2009 were identified using nationally representative primary care data linked to national hospital data, which contain information on 10.5 million patients. HF patients were followed up for 5 years. Group-based trajectory models and sequence analysis were applied. Results: The model categorised the HF population (n = 9466) into 5 subgroups: low-impact (66.9%); two intermediate ones (27.4%); chronic high-impact (2.3%) with steady high annual readmission rates; and short-term high-impact (3.4%) with rapid decline in readmission rates. The groups were defined by their trends of yearly number of readmissions. The all-cause 5-year mortality was highest in the short-term high-impact group (n = 185, 72.8%), followed by group 2 (intermediate users) (n = 744, 58.8%), low-impact (n = 4244, 56.9%), chronic high-impact (n = 88, 37.6%) and group 1 (intermediate users) (n = 401, 30.3%) (p < 0.01). Compared with low-impact users, high-impact users were associated with higher mortality, bereavement episodes, and more out-of-hours GP visits. The chronic high-impact users had distinct sequences of causes of emergency admissions most often consisting of chest infection, ischaemic heart disease, and cardio-pulmonary signs and/or symptoms. Conclusion: Chronic high-impact users constitute a small proportion of total patients, but they have increasingly high use of healthcare services. Short-term high-impact users represent largely end of life patients. They require prompt involvement of the palliative care team to reduce unnecessary readmissions to hospital.