17.1 What does it mean if a cytologist has a d’ value of zero?
A cytologist with a d’ of zero indicates that they are not able to discriminate between normal and abnormal cells.
17.2 What would a response curve diagram look like for a cytologist with perfect discrimination skills?
There would be no region of overlap between the normal and abnormal response curves.
17.3 What happens to the false positive rate and false negative rate when a cytologist shifts their criterion to a higher value?
The false positive rate decreases and the false negative rate increases.
17.4 What does it mean if a cytologist has a criterion value of zero?
A criterion value of zero indicates an unbiased response. In other words, the decision to report a cell as normal or abnormal is based solely on the morphological features of the cell.
17.5 How can a cytologist increase their diagnostic sensitivity without a change in d’?
The cytologist can choose to decrease their criterion (i.e. shift criterion to the left in figure 17.4) and respond “abnormal cells present” more frequently. This will, of course, increase the false positive rate and decrease diagnostic specificity.
17.6 How can a cytologist increase their diagnostic specificity without a change in d’?
The cytologist can shift their criterion to a higher value (i.e. move criterion to the right in figure 17.4) and respond “abnormal cells absent” more frequently. The price paid for this improvement in specificity will be a decrease in diagnostic sensitivity.
17.7 Assuming that cytology response curves have a normal distribution, where on the response curve diagram should a cytologist place their criterion if they are seeking to achieve an equal number of false positives and false negatives?
The cytologist should place criterion at the intersection of the normal and abnormal response curves. The assumption that must be made here is that the two overlapping curves have a normal distribution with equal variance.
See chapter 16 to improve your understanding of statistical terminology in cytology, or alternatively the more general text in the Fundamentals of Biomedical Science Series: Blann A. Data Handling and Analysis. Oxford University Press. 2015)
17.8 List a few examples of workplace distractions that might adversely affect a cytologist’s performance as they carry out their microscopy duties.
Potential workplace distractions are many and varied. Some of the more common causes of distraction are noise, personal electronic devices, extremes of temperature and inappropriate lighting.
17.9 Why do you think cytologists set a low criterion at the start of their training?
The underlying explanation for this phenomenon has not been fully investigated, but we can speculate that trainee cytologists set a low criterion because of their desire to achieve and maintain diagnostic sensitivity. This desire may be reinforced by the individuals responsible for their training.
17.10 Can you think of any examples of non-analytical learning (perceptual learning) in everyday life?
All sighted humans learn about their visual world via processes that are best described as perceptual, e.g. the way infants learn to recognise their mother. The same is true of the way we learn to taste and smell and how we learn to recognise sounds.
17.11 Explain the difference between random error, systematic error and sporadic error.
In cytology, random errors are unpredictable fluctuations in human performance resulting from cytologists’ inability to maintain a stable criterion and/or d’. In contrast, systematic errors are consistent inaccuracies in the same direction (e.g. persistent false positive or false negative reporting). Sporadic errors are intermittent or isolated mistakes.