Loss Of Weight as predictor of CANcer in primary care (LOWCAN)

The LOWCAN projects conducted globally investigate unexpected weight loss as a symptom of cancer. The LOWCAN projects aim to inform clinical guidance for primary care doctors about which patients with unexpected weight loss should be offered investigation for cancer and other serious diseases and who can be safely spared investigation.

Existing research suggested that weight loss was a "low-risk, but not no-risk" symptom, estimating that between 1 in 200 and 1 in 30 patients with unexplained weight loss would be diagnosed with cancer.

 It was not clear how "weight loss" should be interpreted by general practitioners - how much weight loss was important and over what time-period? Should weight be routinely measured or should they wait for the symptom to be volunteered by their patient? Was the diagnostic value of unexpected weight loss independent of other symptoms and test results?

 Without clarification of these points, doctors in primary care may adopt too high a risk threshold for referral (choosing to watch and wait inappropriately and potentially missing cancer cases) or too low a threshold (overburdening resources in secondary care and exposing patients to unnecessary investigation).

WHAT WE ARE DOING:

The LOWCAN projects are three separate areas of investigation:

Systematic review: To identify, assess the quality of, and synthesise all the previous research on the diagnostic value of weight loss as an indication of cancer.

UK studies: Using UK electronic health records data to examine the extent to which coded weight loss indicated cancer in patients presenting to general practice. Additional symptoms and test results were investigated to identify which helped to identify which patients with unexpected weight loss require further investigation. Subsequent research is now underway to investigate whether these findings remain valid in a separate UK dataset (see BLOTTED).

International studies: Using electronic health records data from the US and Australia this research aims to understand if measured weight loss, or weight loss recorded in the free-text of consultations, indicates cancer in the same way as coded weight loss. This research also investigates whether the UK findings can be replicated and implemented in other health systems to understand the impact of primary care professionals using these findings in the clinic.

Expected impact

Our findings are intended to inform future iterations of clinical guidelines on cancer investigation for patients attending their GP, such as the update of the 2015 NICE NG12 Suspected Cancer guidelines.

Furthermore, in the NHS, and in many other countries, clinical pathways are in development to investigate patients with non-specific symptoms such as unexpected weight loss (see SCAN): our findings will inform referral criteria to these pathways. We will also conduct research to integrate our findings into the computer systems that GPs use.