Background: A limitation of service delivery in primary care in the United Kingdom is that services are often organised to manage discrete long-term conditions, using guidelines related to single conditions, and managed in clinics organised around single conditions. However, many older patients have more than one condition (so called multimorbidity). Qualitative research suggests that these patients experience ‘hassles’ in their care, including multiple appointments, poor co-ordination, and conflicting recommendations. However, there is limited quantitative evidence on the ‘hassles’ that patients with multimorbidity experience, or factors predicting ‘hassles’ in patients with multimorbidity. Methods: We conducted a cross sectional study, mailing questionnaires to 1460 patients with multimorbidity identified from the disease registers of four general practices in the UK. Patients were asked to complete a range of self-report measures including measures of multimorbidity, measures of their experience of multimorbidity and service delivery. Data were analysed using regression modelling to assess the factors predicting ‘hassles’ in patients with multimorbidity. Results: In total 33 % (n = 486) of patients responded to the baseline survey. The ‘hassles’ most often reported by patients related to lack of information about conditions and treatment options, poor communication among health professionals, and poor access to specialist care. There was a significant relationship between numbers of conditions, and reports of ‘hassles’. In multivariate analysis, 5 variables predicted more ‘hassles’: more long-term conditions, symptoms of anxiety and depression, younger age, being in paid employment, and not having a discussion with their GP in the last 12 months. Conclusion: Hassles are frequently reported by patients with multimorbidity in primary care. A priority for future research should be on the development of new models of care that better cater for these patients. This research highlights core hassles that need to be addressed, and the patient groups that are most at risk, which may aid in the design of these new models.