The Community IntraVenous Antibiotic Study (CIVAS): a mixed methods evaluation of patient preferences for and cost effectiveness of different service models for delivering outpatient parenteral antimicrobial therapy

Background: Outpatient parenteral antimicrobial therapy (OPAT) is widely used in most developed countries, providing considerable opportunities for improved cost savings. However, it is implemented only partially in the UK, using a variety of service models. Objectives: The aims of this research were to (1) establish the extent of OPAT service models in England and identify their development; (2) evaluate patients’ preferences for different OPAT service delivery models; (3) assess the cost-effectiveness of different OPAT service delivery models; and (4) convene a consensus panel to consider our evidence and make recommendations. Methods: This mixed-methods study included seven centres providing OPAT using four main service models: (1) hospital outpatient (HO) attendance; (2) specialist nurse (SN) visiting at home; (3) general nurse (GN) visiting at home; and (4) self-administration (SA) or carer administration. Health-care providers were surveyed and interviewed to explore the implementation of OPAT services in England. OPAT patients were interviewed to determine key service attributes to develop a discrete choice experiment (DCE). This was used to perform a quantitative analysis of their preferences and attitudes. Anonymised OPAT case data were used to model cost-effectiveness with both Markov and simulation modelling methods. An expert panel reviewed the evidence and made recommendations for future service provision and further research. Results: The systematic review revealed limited robust literature but suggested that HO is least effective and SN is most effective. Qualitative study participants felt that different models of care were suited to different types of patient and they also identified key service attributes. The DCE indicated that type of service was the most important factor, with SN being strongly preferred to HO and SA. Preferences were influenced by attitudes to health care. The results from both Markov and simulation models suggest that a SN model is the optimal service for short treatment courses (up to 7 days). Net monetary benefit (NMB) values for HO, GN and SN services were £2493, £2547 and £2655, respectively. For longer treatment, SA appears to be optimal, although SNs provide slightly higher benefits at increased cost. NMB values for HO, GN, SN and SA services were £8240, £9550, £10,388 and £10,644, respectively. The simulation model provided useful information for planning OPAT services. The expert panel requested more guidance for service providers and commissioners. Overall, they agreed that mixed service models were preferable. Limitations: Recruitment to the qualitative study was suboptimal in the very elderly and ethnic minorities, so the preferences of patients from these groups might not be represented. The study recruited from Yorkshire, so the findings may not be applicable nationally. Conclusions: The quantitative preference analysis and economic modelling favoured a SN model, although there are differences between sociodemographic groups. SA provides cost savings for long-term treatment but is not appropriate for all. Future work: Further research is necessary to replicate our results in other regions and populations and to evaluate mixed service models. The simulation modelling and DCE methods used here may be applicable in other health-care settings. Funding: The National Institute for Health Research Health Service and Delivery Research programme.