Despite billions of dollars in funding, Ontario's digital health initiative has a data quality problem.

Breaking News


Despite billions of dollars in funding, Ontario's digital health initiative has a data quality problem.


The application of modern information technology to support the free flow of patient information around the circle of care is what digital health is all about. For patients, this implies that every health-care practitioner they meet in different places should be able to promptly and efficiently access important health record information.

Electronic health records, for example, are thought to improve patient-centered treatment, integrate care, and maintain the financial viability of our health-care system. However, despite billions of dollars spent over the previous two decades to facilitate speedy and safe sharing of health information, Ontarians are nonetheless confronted with the harsh fact that their health data are still scattered. The COVID-19 epidemic has shown even more data quality concerns.

Much of the available data on COVID-19 is a muddle, as revealed in a recent National Post piece. Data on infected cases and fatalities are not only delayed, but also incomplete. According to reports, Ontario provided contradictory figures from provincial medical professionals and municipal public health departments. It's no surprise that the Ministry of Health recognises that "uniform standards across sectors are absent, making it incredibly difficult to merge patient data or local systems with provincial ones."

It's a bitter pill to swallow after years of effort to enable quick and safe health data transmission.

Neither long-term nor effective

The Ontario government is exploring two measures to improve data quality, with examples including the accuracy and timeliness of data produced by various service providers. The first strategy focuses on leveraging common communication protocols to improve health data transmission across heterogeneous systems (systems built by multiple manufacturers that require different hardware and software configurations to operate).

This strategy, however, is neither scalable nor sustainable as communications among various systems become more complicated, time-consuming, and error-prone as additional systems are added to the mix. Inconsistent numbers of COVID-19 infected cases and deaths supplied by various levels of government are an example. Not to mention that these standards develop quickly, and older versions of the same standard are difficult to map and migrate to current ones.

The second method is based on the Digital Health Playbook's minimal common data set, a resource designed to help health-care companies construct their digital systems. To mention a few applications, the basic data set includes data classes (such as individual patients) and their related components (such as date of birth) for clinical notes, laboratory information, drugs, vital signs, patient demographics, and procedures.

These data sets, while adequate for the needs of family physicians whose primary role is disease control and prevention, are insufficient for treating complicated individuals with many health conditions, which need a tremendous quantity of health data from diverse health-care providers.

These two measures taken by the Ontario government to solve data quality concerns are not sustainable nor successful, and hence cannot be used to guide health digitalization policy.

As scholars concentrating on information technology in health governance, we recommend that a data strategy include four pillars:

1. Data quality requirements

To begin, data quality is an umbrella phrase that incorporates various qualities such as accuracy, accessibility, and timeliness. There are also trade-offs between these dimensions. For example, increasing the frequency of data reports may have an impact on data comprehensiveness, which necessitates time to cover all of the essential data.

While "fit for use" (meaning the quality of data meets the demands of its intended users) is deemed reasonable and pragmatic, it must be explicitly stated what quality criteria must be emphasised. Given limited resources and rising expectations to reduce health-care costs, determining which data quality criteria should be prioritised becomes increasingly important.

2. Long-term, scalable, patient-centered platform

Second, the health-care industry is not alone in coping with decades-old systems and the low-quality data provided by these systems, such as erroneous COVID-19 case numbers. Using lessons learned from banks and other organisations, the health-care industry might develop an open data platform that facilitates data sharing across health-care providers and allows patients to submit data through social media, mobile and wearable devices. Countries like the United Kingdom and Germany have begun to apply the open data platform concept.

3. Measurable improvement indicators

Third, quantifiable outcomes for initiatives to enhance data quality must be identified. Training programmes on best practises for data entry may be implemented, as could system features that allow for data quality verification (for example, completeness or consistency). Measurable outcomes would assure accountability and attainment of the desired goals, as well as informing future funding decisions.

4. The suppliers' improvement process

Finally, a data strategy must explicitly describe a data quality improvement and monitoring process in which data quality is regularly checked and analysed to ensure that data supports patient care and research. Because data quality is a shared obligation, the quality assurance process must take place both collectively among providers and individually inside each provider.

Meaningful interaction with all stakeholders is essential for defining and implementing the data strategy. Patients and providers, for example, must be involved in identifying the data needed to treat the diseases that consume the majority of our health-care expenditure, defining quality characteristics of the data, and specifying roles and responsibilities for preserving data quality.

In contrast to the Ontario government's band-aid approach, the four-pillar data plan is long-term, concentrated, and comprehensive. It would guarantee that data quality is prioritised in Ontario's health digitalization efforts. Following the concept, our health-care system would provide a scalable method for constantly improving data quality.

Without such a data plan, Ontarians risk losing another decade and billions of dollars.

Post a Comment