Hidden in plain sight : personal health records and the invisibility cloak of disadvantage

This thesis with publications sets out to present evidence which will answer the following research question: RQ1: What is the relationship between socioeconomic disadvantage and personal electronic health records? This research question has been prompted by a concern that personal electronic health record systems as they are currently being realised in Australia are unlikely to meet the needs of those at a socioeconomic disadvantage. These individuals are often 'disempowered, disengaged and disconnected', and have been largely invisible in the process of health records development. If this concern is valid, any enhancement in the provision of healthcare which results from the use of personal health records is likely to bypass less capable citizens and patients. The research investigations contributing to the evidence in this thesis are presented as a number of peer reviewed publications; four have already been published and a further two are under review. The thesis highlights how ordinary citizens have been inhibited in their adoption and use of personal health records by the use of highly specialised language, and in Australia by the way in which health records policies and procedures have been developed and implemented. Evidence is presented that validates a concern that even greater challenges exist for disadvantaged users. These users tend to display lower levels of text literacy, technical literacy and health literacy, all of which have been identified as barriers to the adoption and continued use of personal electronic health records. The thesis presents detailed evidence identifying those areas in Tasmania whose populations have higher levels of socioeconomic disadvantage, higher use of public hospital services, and proxy measures suggestive of lower literacy. From a health informatics perspective, this thesis argues that those involved in the design, implementation and evaluation of personal electronic health record systems may have neglected a crucial requirement for such systems to be fit for purpose in the context of their intended use and intended users. The research was conducted in two phases. Phase 1 of the research used exploratory investigations to delineate, test and validate ideas and concepts which were then exposed to critical consideration by peers involved in health informatics research, in the following publications: 1. Language Games and Patient-centred eHealth. This publication explored the way in which the use within ehealth systems of specialised medical language and terminology, SNOMED CT in particular, can be a barrier, even for ordinary citizens who read well and are familiar with technology. For potential users of such systems who are already at a disadvantage the barrier can appear insurmountable. 2. Citizens, patients and policy: a challenge for Australia's national electronic health record. This publication examined the development of Australia's ehealth policy, of which a personally controlled electronic health record is an integral component. Despite extensive consultation with information technology and health industry stakeholders, and with organisations representing health consumers, there was little evidence of direct engagement with ordinary citizens or patients. 3. The PLU problem: are we designing personal ehealth for People Like Us? This publication raised concerns that the approaches taken in the design, implementation and evaluation of personal health records were likely to ignore the preferences, needs and capabilities of disadvantaged patients. The publication further suggested that this limited focus risks diverting resources from those more conventional options for health service delivery which currently provide services for disadvantaged patients. 4. Personal Health Records are designed for People Like Us. This literature review assessed the extent to which the design, implementation and evaluation of personal health records identified the particular needs and capabilities of disadvantaged users. There were very few examples where patient capability as a barrier to use of the system was considered; in some cases, patients with low literacy skills were actively excluded from the process, and effectively rendered invisible. These publications from Phase 1 identified several aspects of personal electronic health record systems that constrain their usefulness for disadvantaged patients. Given that some time had elapsed during completion of Phase 1, two additional questions emerged from these preliminary findings: RQ2a: What is the current evidence about barriers to the uptake and continued use of personal electronic health records? RQ2b: Is it possible to identify a group of disadvantaged healthcare users in Australia likely to face higher barriers to the adoption and use of personal electronic health records? In Phase 2 of the research two further investigations were undertaken in order to address these additional questions. The first investigation involved a review of literature about personal health records and barriers to uptake and continued use, with an emphasis on recent publications. The second investigation used cluster analysis to conduct an empirical analysis of local data about socioeconomic disadvantage, healthcare use and proxy measures of identified personal health record barriers. This analysis involved 2.1 million de-identified hospital records correlated with geo-location, census data and measures of social disadvantage. The two publications from Phase 2 report the outcomes of these investigations: 5. Barriers to meaningful use of personal health records by patients: A structured review (In review) This literature review sought publications about personal health records and barriers to adoption and continued use, predominantly from 2004 to 2014. Analysis of these publications identified a range of barriers which included age, race or ethnicity, income and socioeconomic status, education, text technical and health literacy, Internet and computer access, and disability. The review also identified a number of evaluations which may have introduced a selection bias by actively excluding low capability subjects. 6. Who will benefit most from a personal electronic health record? Analysing Tasmanian data about disadvantage, public hospital use and barriers to the adoption and use of personal health records. (In review) This empirical study of data from 96 geographic areas in Tasmania analysed three of the Australian Bureau of Statistics' Socio Economic Indicators for Areas (as measures of disadvantage), 800,000 records of public hospital admitted episodes and associated ICD10 diagnosis coding, and records for 1.3 million emergency department attendances as indicators of public hospital use (around eight years' data), and area measures of education, Internet access and qualifications (as proxy measures for text literacy, technical literacy and health literacy). Cluster analysis showed two subgroups of areas with disadvantaged, low capability users who were higher consumers of public hospital care, and proxy measures for barriers to personal health record adoption and use, in contrast to a privileged capable subgroup using much less care. Choropleth maps confirmed the dichotomy between these clusters. In summary, this research has identified inequity in the design, implementation and evaluation of personal electronic health records, and has confirmed (in a Tasmanian context) that citizens who experience socioeconomic disadvantage also use public hospital services at a higher rate. These citizens also have a higher incidence of chronic disease and are subject to capability barriers likely to limit their ability to benefit from personal health records as they are currently configured. None of this is new, but what drives this research is the apparent lack of interest in applying these concepts, about disadvantage, health and divides, to the design and implementation of personal health records. The concluding remarks for this thesis note that: 1. Current approaches to research on personal electronic health records mean that the socially disadvantaged are invisible, often being discounted as 'non-adopters' or 'not qualifying' for a study. Without special attention, personal electronic health record systems will continue to ignore the 'disempowered, disengaged and disconnected'; 2. Patients from disadvantaged neighbourhoods in Tasmania use public hospital services to a greater extent than those from privileged neighbourhoods, and display characteristics which are indicative of barriers to personal health record use; 3. There is a risk that disadvantaged patients will receive worse healthcare as a result of a focus on personal electronic health records as they are currently realised. Diversion of health resources to the implementation of personal health records may result in an increased inequity in healthcare outcomes; 4. User centred design for personal health records may help to address this issue, but only with the participation of a truly representative group of potential users, including the disadvantaged. In conclusion, policymakers, informaticians, health service managers and healthcare providers should look carefully behind the cloak of invisibility which hides the disadvantaged, and ensure that any benefits generated by ehealth innovation are being shared equitably. Surely it is in the exceptional degree of blindness and inattention among the proponents, developers and implementers of personal health records to that which is 'hidden in plain sight' - that those whose healthcare is most in need of improvement are the least likely to benefit from having (at least theoretical) access to a personal health record. This thesis provides evidence which removes that cloak of invisibility, and shines a light on this cluster of disadvantaged patients. Will anyone notice?