Caring for cough : guiding diagnosis in primary care

Proper diagnosis in the very common presented acute cough or other LRTI-like symptoms allows GPs to adequately treat and inform their patients about the expected course of their disease. Management depends on the assumed cause of the LRTI, which includes potentially life threatening pneumonia, as well as self limiting bacterial or viral infections and underlying obstructive lung diseases. The challenge for individual GPs is not only in recognizing patients in need of antibiotic treatment (in case of suspected pneumonia), but more broadly in identifying whether it concerns a self-limiting condition, or a disease for which targeted treatment and/or follow-up are indicated. The studies in this thesis were performed to guide the GP in this diagnostic dilemma and the main results were as follows. First, we found that most cases of radiographic pneumonia are not identified based on clinical judgment alone. A diagnostic symptom and signs model had moderate clinical value in discriminating between high and low risk for pneumonia. In contrast to PCT, CRP had added diagnostic value merely by correctly excluding pneumonia. Second we found that previously unrevealed obstructive lung disorders (asthma and COPD)were not uncommon among patients presenting with acute cough, but standardised brief diagnostic models using symptoms and signs derived from patients with chronic cough poorly predicted unrevealed asthma and COPD. Regarding detecting influenza the previously reported diagnostic influenza model (‘Flu Score’) showed diagnostic value for identifying influenza cases especially in patients presenting with acute cough during periods with higher influenza prevalence. For all investigated diagnostic strategies however (e.g. pneumonia, asthma/COPD and influenza) estimated disease probabilities were relatively low in positive cases. In the last chapter we discussed some critical issues regarding the translation of our main results into clinical practice. First we regarded difficulties of implementation of diagnostic models. Next, we discussed whether the patient’s prognosis should be the starting point for treatment, instead of his diagnosis. Finally, advantages and drawbacks of using CRP in daily practice were discussed. In summary, simple and robust symptom combinations as well as C-reactive protein testing may serve as diagnostic tools to assist GPs in their management decisions of patients with acute cough, especially in ruling out pneumonia. However discriminative value of most models remains modest and therefore these models should be used with caution and always in combination with the physician’s own clinical assessment of the individual patient