My second post ever on this site was titled “Diagnosing cancer is hard”, where I argued that making a diagnosis of cancer (or making an appropriate referral) is difficult for primary care physicians in the setting of non-specific complaints that are normally *not* due to cancer. This week an excellent article on the subject was published at The BMJ called “Rethinking diagnostic delays in cancer: how difficult is the diagnosis?”, and I urge everyone to read the article.
Dr. Lyratzopoulos and colleagues do an excellent job describing which cancers are easier, intermediate, and harder to suspect and diagnose. They also discuss potential reasons why “multiple consultations” (three or more visits to a primary care physician before referral to a specialist is made) occur and how that leads to a delay in diagnosis and erodes the trust of patients in the healthcare system. I was quite honestly shocked that 80% of patients eventually diagnosed with cancer had only one or two consultations with a primary care doctor before the appropriate referral was made (presumably one to hear the complaint, and a second visit to go over test results stemming from the initial complaint). I was also surprised that the traditional ominous “red flags” have only a 5% (in women) or 10% (in men) positive predictive value for cancer.
What I feel is missing from the analysis in this BMJ article is the type of situation I recently encountered. A middle-aged man saw his family doctor for a complaint of back pain, and was reassured it was mechanical (using the evidence of a normal chest x-ray and spine x-ray). Three months later he developed acute onset bilateral leg paralysis from multiple-level spinal cord compression from an undiagnosed lung cancer. I would argue that this situation, a single visit with reassurance and eventually a catastrophic diagnosis, is more upsetting for patients than multiple follow up visits (usually with additional tests in between) leading to a moderate delay in referral to an oncologist. Situations like this are not adequately captured in the BMJ article since technically the number of visits is two (the trip to the GP and the emergency room visit post-paralysis).
My takeaway points from this discussion are that:
1) Multiple visits to primary care doctors are not necessarily a bad thing since persistence or progression of symptoms along with close follow up may reduce delays in a cancer diagnosis
2) We need a way to also capture delays in diagnosis in days/weeks/months rather than simply number of primary care appointments
3) Primary care physicians have an incredibly hard job balancing judicious use of resources and the appropriate level of testing indicated for vague or non-specific complaints related to an undiagnosed cancer
Overall I am quite happy to see this situation being studied and published, and I think all primary care doctors should keep the list of easier, intermediate, and harder cancers to suspect and diagnose in the back of their minds.