OpenCancer is born!

OpenCancer was founded to tackle the problems we face with clinical data collection in cancer. There is universal agreement that good data collection underpins modern medical practice, yet we struggle to achieve this. OpenCancer aims to tackle a few of the root causes:

  1. Develop open data models that match clinical, research and reporting needs.
  2. Share best practice in the development of clinical pathways.
  3. Develop reusable technological elements by providing tools and APIs for use cases such as prognostic scoring and auto-identification of trial patients.

Data Modelling

We pride ourselves on a very detailed level of cancer data collection in the NHS. However, this process has been allowed to proliferate in an uncontrolled manner such that different registries stipulate data items that are confusingly similar, but not the sameBy example for a single cancer patient we are asked / obliged to submit closely related data fields to the following registries:

  • Cancer Outcomes and Services Dataset (COSD)
  • Systemic Anti-Cancer Therapy (SACT)
  • National Radiotherapy Dataset (RTDS)
  • National audits such as The National Prostate Cancer Audit (NPCA) and The British Association of Urological Surgeons (BAUS).

There are also newer projects such as The 100,000 Genomes project and The NIHR Health Informatics Collaborative, which overlap with these datasets and mandate the collection of even more data.

Very few hospitals collect this data in clinically-facing systems. Instead we separate this process into the back office, often employing non-clinical staff using non-clinical systems. This results in a significant overhead to the delivery of clinical care, and invariably leads to inconsistencies in data between the different systems.

“Data collection remains problematic in medicine.”

This has to change.

The only sustainable solution is to collect this data as a consequence of the delivery of care. This would not only address the issues of administrative overhead and data provenance, but would incentivise the clinical team to capture good quality data.

“If the data I enter is useful to me and my patient (and not just used for a registry), I will make sure I enter high quality data.”

The data would be collected once and reused through the clinical journey, but would also feed many different reporting datasets.  To achieve this goal, we must agree one data model for cancer, which needs to extend and be compatible with the data model for the delivery of all aspects of care across the NHS. This data model would be used in all clinical systems which collect cancer data, and would also underpin the structure of the data used for registry reporting and research.

“We need a single information model which underpins clinical care, registry data collection and research.”

Clinical Pathways

To translate the information models to real world use, one must consider the different steps of the clinical pathway, and the role of data at each step. Eg At the point of referral from the GP:

  • What data does the GP need to send to allow initial triage?
    • Can this be extracted automatically from the GP system?
    • Is there extra information (such as patient preference) which needs to be entered manually?
      • Who will enter this data?
  • What additional tests may be useful before the initial clinical visit?
    • Can these be requested automatically?
    • Can we pull these results in automatically?

“This pathway modelling is just as important as the information modelling to ensure good data capture, and most importantly good clinical care.”

With OpenCancer, we aim to document and share exemplars of good clinical modelling, to maximise learning from these projects. These pathway models may act as templates for institutions who are to undergo a similar service redesign.

Reusable Resources (including APIs)

We aim to provide useful web-based resources for both patients and healthcare professionals, including:

  • Prognostic and risk scoring calculators.
  • Clinical trial eligibility checkers.
  • Coded lists for common cancer related classification systems such as the CTCAE (Common Terminology Criteria for Adverse Events).

These will be provided as:

  • Simple web pages and apps for patients and clinicians.
  • APIs for incorporation of this functionality into third party apps, generic electronic patient record systems or specialist systems.

Currently every health IT platform, from single departmental systems to megasuite electronic health records, has to build this functionality for their own system. This is expensive, inefficient and ultimately unsustainable at a national scale.

The provision of these tools at a national level has the potential to improve healthcare in an equitable fashion, and accelerate the benefits of the digitisation of cancer care.

Leave a Reply