Revised Abstract

Electronic Health Record software (EHR), is used by medical professionals regularly to interact with patient records. The functionality of this software is key to public health, however the quality of this software does not match it's importance, or its cost. Innovating and developing healthcare software is a challenge due to healthcare's complexity, risk, and In response these challenges, this thesis proposes Barnett: a novel system to store, share, and interact with health records across institutions. Notably, this system moves control from the vendor to the user - through iterative, crowd based improvement, and ownership. This allows the system to fit unique, and varied user needs. Barnett was developed through an interdisciplinary qualitative research process, grounded in perspectives from design, software engineering, healthcare; and interviews with healthcare professionals. The final system will be assessed by expert stakeholders, and in relation to the values and criteria identified in exploratory phases of research. The survey of current EHRs, healthcare models, and the design process indicates that developing a system which embraces the uniqueness of patients, institutions, and users, is potentially a more sustainable approach to electronic health records.


that developing software accross the entire healthcare system instead of in standalone products is potentially a more sustainable approach to healthcare records. that developing EHR software which embraces the uniquness - allows innovation - crowd sourced improvement - that developing a system which enables crowd innovation is potentially a more sustainble approach to healthcare records. - that developing a system which embraces the uniqueness of patients, institutions, and users, is potentially a more sustainable approach to electronic healthcare records.

Abstract

Through reviewing literature and interviewing industry professionals, I identified three main challenges when designing software for healthcare.

Healthcare is complex, and varied. As healthcare has become more effective, it's complexity has increased. The information stored in medical records, and what the user needs varies accross location, role, enviroment, context. Regionally software is bespoke, and unique - and will continue to change in the future. Put simply - every single medical institution is slightly different.

Healthcare is risky. Managing risk is tricky, but software failure is unacceptable in healthcare. Designing software which is robust, and well tested, is necessary.

Healthcare is becoming team based. Healthcare information is transforming from single institutions to a 'team' of doctors around a singular patient. This is a fundamental change in our healthcare system, and cannot be slowly integrated into our existing system

Barnett is an experimental electronic health record system (EHR). It's experimental because it has considered the application of unique ideas and new technology at every level. It differentiates itself from existing EHR systems by being different. It is currently a design proposal.

It is a general purpose EHR system. This means it manages how health records are stored, shared, and interacted with by the end user. It however is not a fully fledged Patient Management System - for example it doesn't deal with staff scheduling, inventory management, or invoicing. It aims to do one thing well - manage health records.

Unique Qualities

Embrace Uniqueness. No two patients are the same. They have unique requirements of their care, so why should their records look the same. A newborn baby's record would look very different to an elderly cancer patient - Barnett accepts this by its design. Likewise, the medical staff interacting with these records are unique - a nurse with poor vision has different needs to a paramedic working under stress.

Likewise, institutions are unique. Regionally, healthcare can look very different, and when looking from country to country, the requirements are completely different. Barnett accepts, and works around this uniqueness, instead of trying to find a 'singular model' which fits all use cases.

Enables mass improvement. By design, every part of Barnett can be replaced, changed, improved, and made available to all users. This creates a system which enables iterative improvement - where small, simple improvements can add up to a huge effect. For example, say a physiotherapist wants to track muscle strength for rehabilitation, but there isn't a place in her patients records to record that. She could define her own storage format, to fit her needs, and add it to her patient's records. Or even better, she could search the registry of modules, and find something somebody else has made which fits her needs. If a nurse is dissatisfied with the interface of the immunization section, he could design his own interface, and make it available to the community. This creates a system where the improvement of the software is crowd sourced.

Removed control. Barnett is not owned, controlled, or run by a singular company or government. It can be used by anybody, for whatever they choose. It aims to be improved and run by constant development and improvement by the community of users. This shifts the power - if a medicial institution wants to change Barnett to fit their unique needs, they can, without relying on a third party.

Additionally, Barnett uses a distributed network for the sharing of records. This means the network can be implemented, and setup in any configuration that is needed, without intervention from a governing power. For example, if two medical institutions want to share records with each other, they can make the connection themselves, while maintaining a robust, failure resistant network.

Be a system. Barnett doesn't aim to be just software - it aims to be a system. Atul Gawande has defined three skills a good system requires:

Every storage format in Barnett is required to be able to be anonymized for statistical research. This means success and failures can be recognized via a 'top down' approach. Likewise, success and failure can be recognized by the staff interacting with the records on a daily basis - this is a 'bottom up' approach.

Combined with the ability of mass improvement, this means solutions can be quickly devised, prototyped, tested, and either implemented or iterated upon.

Enable team based healthcare. Patients no longer see a single healthcare provider. They are surrounded by GPs, specialists, pharmacists, hospitals, laboratories. Getting all these individuals who may possible never see each other to collaborate is a challenge. Barnett's access model appears to the user as a cloud based system - instead of every institution having singular, separate medical records, instead they are all accessing a shared, health record.

Barnett facilitates this collaboration by ensuring that privacy is respected - for example, your dermatologist wouldn't need access to the visitation notes from your GP - unless, they are explicitly granted access by you or your GP.

Barnett also provides unique team communication tools - a writing format (MedDown), which is specifically designed for describing clinical information.

Manage Risk. Risk management is tricky for healthcare software, as there is no way to 'validate' a lot of healthcare information. For example, its easy for a computer to validate if an email address is correct - but there is no way to validate if a blood pressure of 140/90mmHg is correct.

However, Barnett tests and validates wherever it can. The network is designed to be as robust and as available as possible. Interfaces can be tested to ensure input = output. Values can be tested to see if they're in the right format. Improvements can be tested to ensure no data is lost. Prescribed medications can be tested against clinical databases to avoid drug interactions.

All changes to any record are kept, forever, as well as a record of who accessed and changed what. This way mistakes, and errors can easily be reversed, without information loss.