Table 5. Clinical case analysis using prior risk assessment and Bayesian incorporation of clinical symptoms to determine posterior risk.
| Case | Sex | Age | Smoking status | Age started | Age stopped | Smoking duration | Smoking intensity (cigarettes/day) |
Symptoms | LR+ | Calculated prior risk % | PPV Calculated on the basis of individual posterior risk % |
PPV based on presenting symptoms in the published cohort[17]* |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A | M | 68 | Smoker | 20 | NA | 48 | 10 | cough + fatigue |
3.45 | 1.86 | 6.14% Moderate risk |
0.63% Low risk |
| Case A represents a low risk patient based on symptoms alone and therefore would not require further investigation or referral. When we take into account prior risk defined by age, sex, smoking status and intensity, this patient is at greater risk then the moderate risk patient in Case B below and does require further investigation (chest X-ray). | ||||||||||||
| B | F | 61 | Never | NA | NA | NA | NA | dyspnoea + haemoptysis | 27.98 | 0.124 | 3.36% Moderate risk |
4.90% Moderate risk |
| Case B represents a patient with moderate risk when considering symptoms alone. Here consideration of prior risk has little effect on the risk status. | ||||||||||||
| C | M | 58 | Ex | 17 | 47 | 30 | 10 | loss of appetite + haemoptysis |
449.74 | 0.2697 | 54.88% High risk |
45.28% High risk |
| Case C represents a high risk patient based on symptoms and even with a negative chest X-ray this patient would require further investigation to exclude lung cancer[17], as 20% of all chest X-ray requests from primary care in confirmed lung cancer patients are negative[33, 34]. The current cut-off of urgent cancer referrals in the UK is PPV>3% so this patient would be considered at high risk and should be investigated further, regardless of the chest X-ray findings. | ||||||||||||
* from raw data provided by one of the referenced authors
Likelihood ratio = LR+ Positive predictive value = PPV