Using Macmillan cancer decision support tool
Dr Anthony Cunliffe
Diagnosing cancer in Primary Care can be difficult. A GP with a 2000 patient list will see approximately 6-8 new cancer diagnoses a year but will see many patients with symptoms or signs that could represent cancer (Rubin, G. et al, 2015). There are many different types of cancer, and a GP may see only a few cases of the more common cancers each year, and very few of the rarer cancers in a whole career.
With this in mind, we all need guidance and tools to help us consider cancer as a possible diagnosis when perhaps the presenting symptoms are vague or there is an absence of ‘red flag’ symptoms. Through my interest in cancer in Primary Care and my role within Macmillan I was aware of the cancer decision support tools developed by Professors Willie Hamilton and Julia Hippisley-Cox and started to use the electronic version of these developed by Macmillan Cancer Support, Macmillan’s cancer decision support tool (eCDS).
The tool relies on the Read coded symptom information in the patient’s record. The tool can then detect these codes and combine them with patient demographic information to calculate a percentage risk score of the patient having, an as yet, undiagnosed cancer. The GP is then encouraged to use their clinical decision making to decide whether to take further action with the patient, taking in to account numerous factors including the reasoning for the percentage risk score which is provided in the form of a ‘pop-up’ box alert. The tool has been found to be useful for those patients presenting with multiple or vague, symptoms which would not normally prompt a GP to consider cancer as a possible diagnosis.
The eCDS tool can be particularly useful when a patient has seen multiple doctors with different symptoms that alone would not be alarming. The tool detects all symptoms coded by any of the various doctors thereby highlighting the overall clinical picture. If there is a significant risk of cancer based on the coding entered by the various different doctors, eCDS will display a percentage risk score alert as soon as the patient’s notes are opened, allowing the GP to assess further the possibility of cancer as a diagnosis. For this assessment, the GP can use the ‘symptom checker’ function of the tool to add in other symptoms perhaps not previously coded to see if this affects the calculated score. With the score in mind the GP can decide if further actions are needed, and use their clinical decision making to determine if other illnesses may be causing the symptoms that contribute to the percentage risk score. However, the tool can support the decision to arrange investigations within or direct from Primary Care or check the symptoms with the NICE urgent suspected cancer guidelines and, if appropriate, make a referral.
I have found eCDS helpful with Practice Nurses carrying out holistic patient reviews. Practice nurses ask routine questions regarding the patient’s bowel habits and weigh the patient which has sometimes led to worrying issues such as significant change in bowel habit and unexpected weight loss being coded. The eCDS tool was able to pick up these codes and produce a percentage risk score alerting me to a possible cancer diagnosis when I next opened the patient’s file.
Primary Care Practitioners can worry about getting too many alerts within IT systems but I have not found this to be a problem with the eCDS tool. It is not intrusive and when I have not been concerned about the patient, I have simply not acted on the alert. There are times when it has made me consider a diagnosis that I had not previously thought about, but I have still decided not to take further action. On these occasions I consider it a good thing that eCDS made me reflect on the patient’s symptoms and signs, even though I have then reassured myself that no further action is needed. In this sense it is indeed a tool to support clinical decision making.
For more information on the eCDS tool please visit our webpage or email us at firstname.lastname@example.org.