More information about text formats
We read the extremely important paper by Crosbie et al1 with interest as it has potential implications for population screening for lung cancer. However, the paper contains some ambiguities and inconsistencies and it would be very helpful to obtain clarification from the authors in order to interpret their findings in a population screening context.
Firstly, in the methods section, it is stated that ever smokers aged 50 to 74 years registered at participating general practices were invited to a community based lung health check (LHC). It is not stated whether every individual registered as an ever smoker was invited. However, assuming this was the case it appears from the flow chart that a total of 16,402 invitation letters were sent out, but if the aim was to send these only to individuals registered as ever smokers it is not clear why 6,476 letters were sent to never smokers.
In any event, the flow chart indicates that letters were sent to 9,926 smokers and that 2,613 attended the LHC. Thus the uptake of the first filter was 26.3% which does not resonate with the first statement in the results section i.e. “Demand was extremely high”.
There are also two apparent inconsistencies in the data presented; in table 1 the number of attendees is stated as 2,541 yet in the flow chart it is 2,613. In addition, in the legend for the flow chart it is indicated that the overall numbers are based on General Practitioner recorded smoking status for 15,072 individua...
There are also two apparent inconsistencies in the data presented; in table 1 the number of attendees is stated as 2,541 yet in the flow chart it is 2,613. In addition, in the legend for the flow chart it is indicated that the overall numbers are based on General Practitioner recorded smoking status for 15,072 individuals and yet 16,402 letters were sent out.
With complex data like these it is easy to have small inconsistencies but it would be really helpful to have clarification of the exact method whereby potential high risk individuals were identified and invited for assessment along with precise numbers.
We should like to make a further point about the paper’s use of the term ‘false positive’. This is confined to screened individuals who are referred to the cancer clinic and who are not diagnosed with cancer. The paper reports that 2.8% of LDCT positive individuals met these criteria.
However, a larger group had indeterminate LDCT results, 12.7% in the flowchart. The majority of these had negative results on repeat CT at three months. It is debatable whether these should or should not be grouped with the false positives. However, as the flowchart helpfully shows, this group of individuals are an output of LDCT. It is worth noting here that any health decrement arising from these results should not be neglected in an evaluation of screening as a result of not being characterised as false positives.
1. Crosbie PA, Balata H, Evison M, et al. Implementing lung cancer screening: baseline results from a community-based ‘Lung Health Check’ pilot in deprived areas of Manchester. Thorax 2019;74:405–409