With one streamlined organisation actually designed for team-based care and healthcare information rationalised in central repositories, we can make much better use of existing resources to deliver more timely and effective healthcare to everyone in British Columbia. We can integrate and provide more timely, accurate, and comprehensive healthcare information in a data warehouse for public health, research, and healthcare system management. We can systematically deploy Artificial Intelligence to deliver faster, better care. The healthcare system would steward patient information. Patients would no longer be responsible for managing and communicating their healthcare information. A healthcare data warehouse and system-wide implementation of Artificial Intelligence would be the icing and the cherry on our information system cake.
Current State
To plot a course to a destination, we need an origin (Figure 1). This depiction of the current state is not perfect but it’s perfectly good as a starting point. Some systems and linkages are left out for simplicity. In my experience, we don’t need to waste time and resources describing the current state in excruciating detail. It’s more important to have a clear idea of where you want to be than it is to have a clear idea of where you are.
The current state of healthcare information and service delivery is information anarchy and ad hoc team care, except in hospitals and some clinics. The good news is that we have the technology; we just don’t use it well. Information is mainly stored in Commercial Off the Shelf (COTS), cloud-based applications. There are a few custom, cloud-based applications. All of them have proprietary data models.
The information management problem is acute in primary care where solo clinicians deliver ad hoc team-based healthcare. Physicians and nurse practitioners use a variety of COTS Electronic Medical Records systems (EMRs ), some share the same EMR, and some use artificial intelligence (ai), e.g. note taking. Lab, imaging, hospital discharge summaries, community health clinic visits, and vaccination information can be accessed through CareConnect. Pharmanet contains a record of all prescription drugs dispensed in the community but not in hospital. Patients may maintain their healthcare information in a cloud-based system, e.g. Apple Health. Information is exchanged as text (paper, eFax) or verbally by the patient (“PatientNet”). Patient information is fragmented and duplicated across multiple holdings, and may be lost when a clinician retires or dies.
In acute care, hospitals deliver team-based care supported by a COTS Electronic Health Records system (EHR). The EHR provides access to clinical information, and enables workflow and workload management. In British Columbia, Health Authorities use one of two EHRs (Meditech, Cerner), and each hospital or agency (e.g. BC Cancer) has its own instance of patient information, i.e. patient chart. When a patient is discharged, they may receive a briefing and prescriptions, and a discharge summary is placed in CareConnect. Clinicians can write notes in a patient chart in another facility that uses the same EHR, e.g. Vancouver General Hospital (VGH) cardiology can write a note in a patient chart at Lions Gate Hospital (LGH) for an LGH patient who received a procedure at VGH.
Ultimately, the patient is responsible for managing their medical information. This is a problem. They may not have been given their information, may be incapable of providing a complete or accurate medical history, and may not be believed even if they can. If the patient doesn’t have a primary care provider, there’s no one to maintain it in an EMR. Frankly, this is ridiculous.
The BC Center for Disease Control (CDC) compiles and integrates healthcare data from various sources into its Public Health Reporting Data Warehouse (PHRDW) to support public health. The BC CDC has data sharing agreements with the Health Authorities for acute care data, and with LifeLabs for lab data. It uses billing data uploaded to the Medical Services Plan (MSP) system as a proxy for primary care health information. Physicians upload data twice a month. The BC CDC also periodically surveys patients, e.g. BC Survey on Population Experiences, Action, and Knowledge (SPEAK).
Public health information isn’t timely, accurate or deep. Use of MSP billing data creates a reporting delay, analyses are limited to billing codes, and there is limited ability to “drill down” into patient data. Billing codes for novel conditions of interest (e.g. Covid) take time to set up and deploy. Delays in identifying the type and geographic extent of a rapidly spreading infectious disease can reduce the effectiveness of public health measures. Patients may be unable or unwilling to accurately complete surveys.
Healthcare system utilisation and performance data are compiled by the provincial and federal governments, and healthcare information is provided to external parties for research. For example, the University of British Columbia houses the Therapeutics Initiative, which provides evidence-informed recommendations on drug therapy, and a systems modelling group, which developed a COVID disease model to describe system state and evaluate policy options.
Rationalised Systems
Healthcare information systems can be designed and administered to support timely and effective team-based healthcare (Figure 2).
Information systems would be administered by a single government entity (Medical Information Management Services – MIMS), which would eliminate duplication and significantly reduce cost. Duplication of data, tests, and diagnoses is expensive; their loss is wasteful.
The heart of the rationalised healthcare information system is one COTS EMR for primary care, one COTS EHR for acute care, and a BC Health Data Warehouse for public health and for healthcare system management. The EMR and EHR would be single logon, clinical information consoles configured to meet the needs of clinicians and manage workflow. A data warehouse is optimised for analysis and reporting, and takes pressure off production systems. In our public health system, data would be uploaded every 24 hours from production clinical, lab, and financial systems into the data warehouse. Pharmanet would become a complete record of medication dispensed, plus vaccinations, in the community and acute care.
Rationalisation and centralisation of healthcare information would give patients control over their data for their care, and could also improve the care of others. Private and public sector organisations can only use data for the purpose for which it was collected. Patients could authorise use of their healthcare information to improve care of their children or siblings. This would simplify research into hereditary diseases (e.g. cancer) and clinical recommendations for medication (pharmacogenetics). They could also authorise use for research and public health. Public health would then have accurate, timely and complete information for infectious disease surveillance and could easily “drill down” to obtain more specific information.
More accurate, integrated, timely, and comprehensive information would mean better healthcare for individuals and better public health. Healthcare information would be safeguarded in authoritative, complete, and integrated information repositories accessed by standard applications. Ready access to more complete and accurate information would mean less frustration, better decisions, and better outcomes. Clinicians would spend less time compiling information and more time delivering patient care. Patients would no longer be responsible for information management, and would no longer be asked for information the health system already had.
Future State
Rationalised systems are scalable and support expanded team-based healthcare (Figure 3). The patient would authorise role-based access to their information by other healthcare professionals, e.g. physiotherapists, dentists, ophthalmologists. Emergency services would have access to patient information. We could systematically implement Artificial Intelligence to improve healthcare.
Increased scope of practice for community pharmacy would be addressed by treating medication provision and management as separate services, as they are in acute care. Pharmacists in private sector pharmacies would continue to dispense medications and make minor diagnoses on a fee for service basis. Community Clinical Pharmacists on salary would be a new service to provide consults, perform medication reviews, manage medication treatment protocols, and diagnose minor ailments virtually or in-person at Community Health Clinics. Healthcare professionals (esp. physicians) would have priority for consults from Community Clinical Pharmacists.
Accessible, authoritative and comprehensive patient and public health information would enable us to systematically deploy artificial intelligence (AI) to improve and expand healthcare. AI is only as good as the information on which it is based. Garbage in, garbage out, as we used to say. AI can support clinicians, emergency responders, public health, and patients by quickly providing an analysis, preliminary diagnosis, or recommendations for treatment and further tests. It can also be used by clinicians to improve their knowledge and practice. Clinicians could use AI to evaluate their prescribing against best practices (Prescribing Review), and to follow-up patients where diagnosis and treatment changed (Diagnosis Reviews). Reviews could count towards Continuing Education required for license renewal.
The healthcare information system would contain a very large, timely, and authoritative dataset to support clinical research and public health modelling by the BC CDC and other authorised parties. For example, the Therapeutics Initiative at UBC would have access to clinical data for assessing drug therapy. The UBC Modelling Group would have access to real-time infection data to calibrate and test predictive models used to evaluate public health measures.
The healthcare system would be better managed, and would be accountable to patients and healthcare providers. The data warehouse would contain performance measures, e.g. time to diagnosis by specialists, patient reported experience measures (PREMs), patient reported outcome measures (PROMs). The Provincial government would have performance measures to manage workload and manage the healthcare system. The Federal government would have performance measures to evaluate and benchmark provincial healthcare systems. Patients and healthcare providers would have access to this information.
Next Steps
It is essential that healthcare systems respond to needs of clinicians. Although the system would be comprehensive in design, it would be implemented in phases. Clinician involvement is critical. Pilots would be used to refine systems, benchmark performance, and develop transformation strategies and plans. We can learn from past EHR rollouts.
We now have all the elements of a healthcare system transformation project. With this project, we can have a public system that delivers efficient, cost-effective, high quality healthcare to all British Columbians. A healthy population is essential for a healthy society and a productive economy
Where do we start? Leadership, direction, and mandate. We need leadership focused on transforming our healthcare system into one designed for team-based care, not patching one where team-based care is ad hoc. The focus must be on improving value and outcomes, not simply reducing costs. In British Columbia, this means one health authority, not six. We have a rational framework within which we can make progress at a tolerable pace. We can eat the moose one steak at a time.
Healthcare System Transformation Project
Performance Measures and Improved Accountability to Patients and Healthcare Workers
Improving Primary Care with improved Information Management
A New Day in Primary Care (story)
Streamlining the Health Organisation
Putting the Pieces Together (story)
Team-based Healthcare for Everyone, Everywhere
Primary Care in Lake Cowichan (story)
Overdose! (story)
Stewardship of Healthcare Information
greg steer
Enterprise Coach