Fighting the Pandemic – With Data
Why cancer data has never been so important
by Sarah Miller, Head of Informatics – East of England Cancer Alliances
We currently see data everywhere. Numbers. Rates. Coronavirus infections, hospital admissions, deaths. It can feel overwhelming. In my living memory I cannot remember a time when the general population were so exposed to such a myriad of daily data, epidemiology and public health information.
Generating this volume of data at this level of frequency is important at times of a national incident like COVID-19, but it is also challenging for those of us working behind the scenes. Even the most robust of health data collection systems and automated analytical processes are prone to error, which is why there is a quality assurance process that usually takes days or weeks. Yet the pressure has been high to get things right in almost real time. From our desks, those of us working in data and analysis roles are grateful to our colleagues on the frontline of the NHS who toil for patients directly affected by COVID-19 and we ask, what can we do to help?
We analysts, data scientists and informatics experts across the NHS England and Improvement East of England region have helped in the best way we know how. We have worked long hours squinting at numbers on computer screens with little respite since March. We have collated, cleaned, aggregated, counted, processed, engineered, visualised and analysed multiple datasets between now and March. We have communicated beyond our analytical silos – with our counterparts across various specialities and departments, both across the region and outside of the region. Our aim has been to robustly quantify the impact of the pandemic, for the benefit of our patients and our NHS services.
For cancer, our knowledge and awareness of data has never been so important. The field has experienced some unintended consequences of COVID-19 – based on the redeployment of cancer staff to support the COVID-19 response; the cessation of aerosol-generating diagnostic procedures; the re-purposing of theatre space in hospitals and a reduction in patient access to, and interaction with, primary and secondary care services during the first lockdown.
What did we do to help? From March, my Informatics Team had to become familiar with new cancer-related datasets almost overnight; ones that I had not used in the 15 years I have worked in cancer statistics. These were mainly operational cancer datasets – patient waiting lists, backlogs, referral activity and treatment numbers. We had to rapidly understand how the numbers have changed during the pandemic. These datasets now, rightly, form part of our everyday working life. This enables me to update dozens of senior colleagues every week with cancer service progress across all of our systems and trusts.
Since March 2020, my Informatics Team has produced more than 250 data packs and reports to update national, regional and local colleagues with cancer recovery progress. This is breath-taking.
During COVID-19, with our Level 4 Critical Incident Cancer Clinical Cell, I developed a brand new cancer data collection template, to bridge the gaps in our knowledge about cancer services – ordinarily, this would take 6-12 months to develop and embed; instead it took around four weeks! Without the buy-in and support of our 14 trusts’ Cancer Managers and Cancer Leads, this would not have been possible. During a highly pressurised time, these NHS staff have also been capturing vital data for us to track almost real time cancer service progress. I cannot thank them enough.
My job is to look for anomalies, outliers, data quality and patterns in the cancer datasets to flag to my colleagues. By presenting the numbers as evidence, it has stimulated crucial discussions with the national team, regional colleagues, systems, commissioners and trusts. Even data with limitations is useful because I can ask: What is missing? Why is it missing? Is the data definition clear enough? How can we support improvements? The Cancer Alliances therefore have a vital role in terms of informatics, in that we have a helicopter view of all cancer services in the region. As such we can benchmark, we can compare, we can share best practice across geographical boundaries.
Cancer referrals are an example metric that we have followed closely since March 2020. Many cancer patients begin their journey on a cancer pathway via their GP. Around 40% of patients in the East of England are diagnosed with cancer through this “two-week wait” GP referral route (2ww) – with the expectation that 93% of these patients are seen within 14 days.
According to the latest published management information data, cancer referrals dropped to around 2,000 patients per week in the East of England during April 2020. The latest data shows that the number of referrals in the region recovered to more than 100% of the pre-COVID baseline during October and November 2020 – with around 6,700 two week wait referrals per week across all cancer pathways. As of December 2020, referral numbers reduced, which is the usual trend around the Christmas period.
The latest referrals data shows that referral numbers are slightly lower to pre-COVID levels by 21st December (Figure 1). I recommend applying caution when interpreting this data, because it is operational and management information and, as such, there will be inherent data quality issues until the data is validated. However, it is useful as an indicative data source and its value is highest when we place it alongside the local contextual narrative from provider trusts and systems.
Figure 1. Two-week wait referrals, NHS Digital e-referral service dataset, January 2020 to December 2020
Breast, skin and colorectal referral pathways in the region increased beyond their pre-COVID levels across many provider trusts in late 2020. Our trusts have been working hard to extend their clinics, including opening some on weekends to account for these extra patients. For pathways like lung, which has recovered at a slower rate – partly due to symptom overlap with COVID-19 – our systems have increased patient and public communications and the Cancer Alliances have taken an active role to disseminate the NHS and PHE “Help Us, Help You” campaign, focused on encouraging patients with suspected cancer symptoms to contact their GP.
This is why data has never been so important. Without it, we would only have an informal, anecdotal perspective on the impact of the coronavirus on cancer pathways. Interventions and campaigns would be based on conjecture and not evidence. However, with it, we can set the scene, quantify the changes, amend service provision and drive evidence-based conversations to reduce any variation. Ultimately, we want to ensure fair and positive outcomes for cancer patients in the region.
My next steps? I intend to incorporate the patient voice into our informatics products and to assess the impact of COVID-19 on inequalities and outcomes for our patients on a cancer pathway. The patients and their data are at the heart of everything we do. We will continue to fight this pandemic for them, using their invaluable health data.