Summary

The United Kingdom has lower survival rates for many types of cancer than the rest of Europe,1 partly due to delays in diagnosis.2 When people present with non-specific symptoms, general practitioners (GPs) may find it difficult to determine whether further clinical investigations are needed. This may delay access to treatment, limit treatment options and reduce survival rates.2 3 To address this, Macmillan Cancer Support developed a Cancer Decision Support tool.4 It calculates a person’s risk of having an undiagnosed cancer based on symptoms, medical history and demographic data,5 and helps GPs consider whether further testing or specialist referral is needed. It uses two algorithms that have accurately identified people at risk of having an undiagnosed cancer and supported earlier diagnosis by GPs.6-11 A pilot study found that the tool raised awareness of cancer symptoms among GPs and encouraged further investigations.12 It has since been rolled out across the country.4

Challenge

Survival rates for many types of cancer are worse in the United Kingdom (UK) than in the rest of Europe,1 partly due to delays in diagnosis.2 For many cancers, symptoms can be vague and non-specific, meaning that general practitioners (GPs) may find it difficult to determine whether further clinical investigations or a referral are needed.2 3 Delays in diagnosis have a negative impact on the lives of people with cancer, their families and caregivers – the UK Department of Health has reported that increased public awareness of cancer symptoms and earlier diagnosis could save an estimated 5,000 lives every year.13

Solution

Macmillan Cancer Support developed a Cancer Decision Support (CDS) tool aimed at GPs to improve the early detection of cancer in the UK.4 The tool calculates a person’s risk of having an undiagnosed cancer based on symptoms, medical history and demographic data pulled from their medical records.5 The calculations are based on two algorithms developed and validated by UK researchers – QCancer® and Risk Assessment Tool (RAT).2 6 14

The CDS tool was launched in 2013, initially focusing on lung, colorectal, pancreatic, gastro-oesophageal and ovarian cancer.5 It has since been expanded to include blood, breast, cervical, uterine, kidney, prostate and testicular cancer. While it does not diagnose cancer, it helps GPs ‘think possible cancer’ and consider further testing and specialist referral. To do this, it uses:

  • automatic prompts that alert GPs when a person’s risk of a specific cancer is above 2% (medium risk)
  • a symptom checker where GPs can select symptoms from a list to calculate a person’s risk of having an undiagnosed cancer
  • a risk stratification tool that calculates a score for every patient in a practice, allowing GPs to screen their patient population based on their risk of different types of cancer.5

Using this information and their clinical judgement, GPs can choose to continue monitoring the person or proceed with further investigations and/or a referral.

What has been achieved

Both of the CDS tool’s algorithms have been shown to support early diagnosis of cancer:

  • Use of the RAT increased lung cancer diagnoses by 37%, with a 19% increase in early diagnoses (stage 1 and 2)6
  • QCancer® accurately identified people at risk of undiagnosed colorectal,7 ovarian,8 gastro-oesophageal,9 kidney10 and pancreatic cancer11 when tested in general practice data sets.

From March to November 2013, Cancer Research UK piloted the CDS tool in 439 GP practices in England.12 While the study did not measure the impact of the tool on the number of cancer diagnoses made by GPs, it did find that:

  • the cancer risk calculated by the CDS tool was the same in 54% of people, higher in 31% of people and lower in 15% of people when compared with the risk perceived by GPs
  • 20% of people were referred, 23% of people underwent further testing and no further action was taken in 47% of people when GPs used the CDS tool
  • for 19% of people, GPs reported that they would not have taken further action had they not used the CDS tool.12

Some GPs reported that repetitive automatic prompts may discourage them from using the tool (‘prompt fatigue’) and that consultation length may also be a barrier to the initiative.12 People with cancer were concerned about the impact of decision-making tools on the quality of their interactions with GPs.

Since its launch in 2013, the CDS tool has been rolled out across the UK through two GP information technology systems: EMIS Web and INPS Vision.4

Next steps

There are plans to further improve access to the CDS tool by embedding it in TPP SystmOne (IT system).

Further information

References:

  1. De Angelis R, Sant M, Coleman MP, et al. 2014. Cancer survival in Europe 1999–2007 by country and age: results of EUROCARE-5—a population-based study. Lancet Oncology 15(1): 23-34
  2. Hippisley-Cox J, Coupland C. 2013. Symptoms and risk factors to identify women with suspected cancer in primary care: derivation and validation of an algorithm. British Journal of General Practice 63(606): e11-e21
  3. Morgan I, Wilkes S. 2017. Improving early diagnosis of cancer in UK general practice. British Journal of General Practice 67(659): 276-77
  4. Macmillan Cancer Support. Cancer Decision Support (CDS) tool. Available here: https://www.macmillan.org.uk/about-us/health-professionals/programmes-and-services/prevention-early-diagnosis-programme/cancer-decision-support-tool.html
  5. Macmillan Cancer Support. The Macmillan Cancer Decision Support (CDS) tool: Frequently asked questions Available here: https://www.macmillan.org.uk/_images/cds-faqs_tcm9-295413.pdf
  6. Hamilton W, Green T, Martins T, et al. 2013. Evaluation of risk assessment tools for suspected cancer in general practice: a cohort study. British Journal of General Practice 63(606): e30-e36
  7. Collins GS, Altman DG. 2012. Identifying patients with undetected colorectal cancer: an independent validation of QCancer®(Colorectal). British Journal of Cancer 107(2): 260-65
  8. Collins GS, Altman DG. 2013. Identifying women with undetected ovarian cancer: independent and external validation of QCancer® (Ovarian) prediction model. European Journal of Cancer Care (English Language Edition) 22(4): 423-29
  9. Collins GS, Altman DG. 2013. Identifying patients with undetected gastro-oesophageal cancer in primary care: External validation of QCancer® (Gastro-Oesophageal). European Journal of Cancer 49(5): 1040-48
  10. Collins GS, Altman DG. 2013. Identifying patients with undetected renal tract cancer in primary care: An independent and external validation of QCancer® (Renal) prediction model. Cancer Epidemiology 37(2): 115-20
  11. Collins GS, Altman DG. 2013. Identifying patients with undetected pancreatic cancer in primary care: an independent and external validation of QCancer® (Pancreas). British Journal of General Practice 63(614): e636-e42
  12. Moffat J, Ironmonger L, Green T. 2014. Clinical Decision Support Tool for Cancer (CDS) Project Evaluation Report to the Department of Health. Available here: https://www.macmillan.org.uk/_images/cds-executive-summary_tcm9-291978.pdf
  13. Department of Health. 2007. The cancer reform strategy. London: Department of Health
  14. Hippisley-Cox J, Coupland C. 2013. Symptoms and risk factors to identify men with suspected cancer in primary care: derivation and validation of an algorithm. British Journal of General Practice 63(606): e1-e10