Data Strategy for health and social care

please see overall reflections

see overall reflections

see overall reflections

Please find below an SBAR paper written regarding primary care cancer data. This highlights some of the issues with data from a primary care perspective

Primary Care Data for Cancer Control Efforts
Recommendations for Data

Authors: Douglas Rigg, Lorna Porteous, Sam Harrison, Chris Williams, Catherine Thomson

Situation

Data from primary care is invaluable for assessing the cancer pathway, from symptom presentation to treatment monitoring to community-based care. At present data from primary care is not easily extracted from GP clinical systems, nor linked to the national cancer registry in Scotland.

There is a clear and pressing need (particularly highlighted by the COVID 19 pandemic) to assess primary care activity within the context of cancer care. For example, symptom presentation, referral activity, test activity, cut by demographic groups and geography, could provide an invaluable source of information for early diagnosis efforts (prioritisation and monitoring). Community-based care, including those patients discharged from specialist care for “best supportive care”, and cancer treatment monitoring are other parts of the pathway that could benefit from increased understanding of activity.

Data from a novel dataset (MAINROUTE1) in England, has been a useful resource for assessing changes to public help-seeking behaviour and GP referral activity during the pandemic.

In addition, making relevant data available to GP clusters can help drive quality improvement activity, support decision making and sharing of good practice. This also helps reduce unnecessary variance of clinical practice between local practices of similar patient demographics.
(ref:https://ihub.scot/media/9095/20220330-gp-cluster-report-v10.pdf)

This would help support action point 2.10 in the Scottish Government’s “Effective Cancer Management: Framework” document: “An audit of Primary Care referrals should be undertaken in collaboration with General Practice clusters at least annually and discussed with local Cancer Management Groups to support compliance and any lessons shared.” https://www.gov.scot/publications/frameworkeffective-cancer-management/pages/4/

1 https://www.phc.ox.ac.uk/research/cancer-research-group/monitoring-attendance-investigations-referralsand-outcomes-during-covid-19-mainroute



Background

Data from primary care systems is currently not collected to a nationwide repository or analysed in Scotland. SPIRE (Scottish Primary Care Information Resource) has been a national approach to collecting data for secondary purposes. SPIRE allows small amounts of information from GP practice records to be used to help doctors’ surgeries, NHS Scotland and the Scottish Government to improve
care and plan services. SPIRE can also help with research into new treatments for particular conditions or diseases, to monitor outbreaks (of flu for example) and to develop new medicines. It allows this information from GP patient records to be securely transferred and safely processed by Public Health Scotland (PHS).

SPIRE can also be used in research. Bespoke searches and reports can be set up but technical expertise is needed to build the searches, there is an approval process to be cleared, and the data itself has limitations in regard to quality and completeness.

There are efficiencies in the management and extraction of primary care data underway, notably with the purchase and implementation of Albasoft’s data extraction pump which should enable new technical approaches to accessing primary data. Recently purchased too was a terminology server which is likely to assist the granularity, and agility of, updates to primary care data classifications.

As noted above, there is concern about the current consistency and completeness of primary care data within clinical systems. Data such as referral activity and investigations (both radiology and laboratory) should be consistent and available, but symptoms and examination findings are usually recorded in freetext rather than codes that would allow for searching and data extraction. There is also additional added complexity with the different clinical systems in use currently. Data fields of
interest to cancer include, risk factors (e.g. weight, smoking status), symptoms and clinical signs associated with cancer, and tests undertaken within primary care. Work is underway within SPIRE to draw out and evaluate risk factor information. The quality of data on signs and symptoms of cancer is unclear, but with the introduction of the new terminology server it is hoped that an eventual move from READ to SNOMED-CT codes within primary care systems maybe smoother.

Separate to the technical challenges, there is the key challenge of ensuring that the primary care workforce enter the data that is required to produce meaningful activity reports. With the removal of the QOF there is no longer the incentive to ensure general health data, such as smoking status, is routinely recorded. Not all clinicians will routinely use discrete codes for clinical symptoms or signs. There may be opportunities with newer clinical systems to pull ‘code words’ directly from primary care free text notes.

For patients receiving palliative care, the national “Palliative Care DES” remains in place. This ensures ongoing maintenance of palliative care registers with in practices with coding. Key Information Summaries (eKIS) could be used as a proxy for “anticipatory care planning”. Other markers of palliative care such as prescribing of “Just in Case” medication could also be extracted if required. This would
enable us to build a picture of palliative care delivery in the community.

Assessment

There are several current opportunities to drive forward progress on the availability and quality of primary care data for cancer specifically including:
• Due to COVID, there is renewed interest and drive to understand disruption to primary care, as related to cancer, within the Cancer Governance Structures (i.e. Cancer Data Programme Board and Early Diagnosis Programme Board).
• The new 10-year cancer strategy for Scotland is underway, with major focuses on data, early diagnosis and throughout primary care.
• Palliative Care strategy for Scotland is also underway and inclusion of impact on Primary Care is important
• Technological and system advances as noted above

Recommendations

We recommend the set-up of a Cancer and Primary Care Data Task and Finish working
group to scope and deliver activity in this area, which could include:

• Developing an overview of the current state of the primary care and cancer data
landscape with more detailed recommendations. This overview should highlight strands of activity, who is accountable for delivery and timelines/milestones. This should look at all aspects of primary care from presentation to delivery of palliative care.
• Scoping of a pilot project looking at the quality of symptom and signs data in primary care datasets currently (could include all cancer relevant variables?). Within this consider challenges and opportunities related to:
o Past incentives to improving quality of data
o Clinical coding, including READ-SNOMED mapping
• To liaise within and/or consider a collaboration with the MAINROUTE team in England to facilitate learning and sharing of challenges/solutions
• Develop a Cancer Cluster Activity Report for GP Clusters to support Quality Improvement Activity.

This recommendation has been drawn together to address the perceived/real fragmentation and duplication of work in this space. Historically there has been a lack of progress in this area due to lack of ownership and accountability within and between organisations and individuals.

References:

1. Consultations for clinical features of possible cancer and associated urgent referrals before
and during the COVID-19 pandemic: an observational cohort study from English primary care
2. https://www.nature.com/articles/s41416-021-01666-6.pdf?origin=ppub SPIRE – accessed 4th
May https://spire.scot/