Implementation and adoption of the first national electronic health record: a qualitative exploration of the perspectives of key stakeholders in selected English care settings drawing on sociotechnical principles
Cresswell, Kathrin Martina
Introduction: Internationally, there is increasing interest in the potential of information technology to enhance the quality and efficiency of healthcare. Many countries are currently actively pursuing electronic health record implementations. However, the introduction of such systems often has significant consequences for users’ work practices and organisational functioning due to the complex processes involved in implementing and adopting new technology. Problems may be exacerbated in a national implementation context if users feel that systems are imposed and offer insufficient customisability due to a focus on achieving widespread interoperability. England has embarked on a large-scale national implementation of electronic health records. One of the procured systems was Lorenzo, which was to be built while it was being implemented. Investigating the implementation and adoption of Lorenzo is of particular interest as, in theory, the approach of “co-creating” a system in collaboration with the National Health Service (NHS) should help to increase software usability and thereby facilitate integration with work practices. I sought to understand the views and experiences of users as well as organisational consequences of introducing Lorenzo, and how these evolved over time in the complex environment of a national electronic health record implementation. Methodology and methods: I conducted a qualitative longitudinal investigation in purposefully selected secondary and community care settings which were implementing early Lorenzo functionality. I conceptualised the settings as case studies. Data collection was theory-driven in that it utilised a methodological framework, which was developed specifically for the purposes of my study and based on the existing theoretical and empirical literature. Using this framework with multi-sited ethnography helped me to examine the immediate environment in which Lorenzo was implemented without neglecting the organisational and political context in which local developments were situated. Data collection consisted of interviews with Lorenzo users and managers in case study sites; interviews with external stakeholders (including policy makers, system developers, and independent sector representatives) from outside NHS Trusts; non-participant observation of staff meetings and use of the technology; as well as a combination of field notes, documents pertaining to Trusts and wider political developments, and press statements. Data collection and thematic analysis were informed by a sociotechnical Actor-Network Theory-based approach highlighting the interrelated nature of technical and social dimensions. The study also drew on other related theoretical frameworks that helped to address some of Actor-Network Theory’s theoretical and practical shortcomings. Most helpful in this respect were Strong Structuration Theory, the Social Shaping of Technology, and the Theory of the Diffusion of Innovations (theoretically); and multi-sited ethnography and case studies (practically). I employed inductive and deductive analytical techniques utilising thematic tables for organising and interpreting the data. Individual case studies were analysed first in order to examine local dynamics, before cross-case comparisons were made and findings were integrated with data obtained from outside case study sites. Results: I collected data between 2009 and 2011 in three case study sites. The complete dataset comprised interview data from a total of 66 different participants within Trusts, 14 interviews with stakeholders from outside case study sites, 38.5 hours of non-participant observation, 149 pages of press statements, 31 pages of field notes, and a range of national and local Trust documents. The three sites differed in demographics and local implementation strategies, and hence presented diverse stories of sociotechnical change unfolding over time within their complex individual contexts. However, there were also similarities, not least the fact that all were implementing the same system and that they were operating within constantly evolving political and economic contexts. Users found it difficult to integrate Lorenzo with their everyday work practices as the software was perceived to be not fit-for-purpose. Over time, these difficulties attenuated to some extent, particularly in the smaller-scale deployments in sites that had invested significant time and resources to adapt the software to fit with their everyday practices. Lorenzo implementation also had significant consequences for organisational functioning, which was often hampered by local restrictions in software customisability associated with national arrangements. Conclusion: I have developed a theoretically informed methodological framework and applied this to explore sociotechnical processes involved in the implementation and adoption of Lorenzo. In doing so, I identified potentially transferable theoretical insights into local and national developments over time and based on these proposed mechanisms involved in the implementation and adoption process. Overall, my findings help to explain why the adoption of Lorenzo was much slower and on a smaller scale than originally anticipated. The interplay between social (political, individual and organisational) and technical factors was central to implementation progress. At the root of many problems encountered were difficulties with integrating systems with work practices of users and more general organisational functioning. In relation to Lorenzo, co-creating national software with strong user involvement was hampered due to different requirements in individual settings and wider, political and economic constraints. Based on the English experience, there may be some important transferable lessons for similar ventures in other countries. Most importantly, national implementations need to build on a solid basis of local technology adoption by allocating sufficient time for individual users and organisations to adjust to the complex changes that often accompany such service redesign initiatives.