JMIR Publications

ABSTRACT

Background:

Artificial Intelligence (AI) is mandated to offer support also in finding valid diagnosis.

Objective:

In the context of an ongoing project, we present the available data for cough as a predictor of several diagnoses in primary care as a possible supplement of a machine based diagnostic algorithm. Furthermore, we discuss the applicability of such an algorithm against the background of the specifics of general practice.

Methods:

PubMed and the Cochrane Library were searched with defined search terms, supplemented by the search for grey literature via the “German Journal of Family Medicine” until 29th November 2022. Inclusion criterion was the explicit analysis of cough as a predictor of any conceivable disease. Exclusion criteria were articles other than original, language other than English or German, and the not mentioning of cough as a predictor.

Results:

In total, 235 records were identified for screening, of which 24 articles were identified to meet our inclusion criteria. Most results (eight) were found on chronic obstructive pulmonary disease. The others distributed among the topics asthma or unspecified obstructive airway disease, different infectious diseases, dyspepsia/ gastroesophageal reflux disease, and bronchogenic carcinoma. Positive odds ratios were found for cough as a predictor of chronic obstructive pulmonary disease, influenza, COVID-19 infection, and bronchial carcinoma whereas the results for cough as a predictor of asthma and other not specified obstructive airway disease were inconsistent.

Conclusions:

The example of “cough” does not provide a sufficient database to contribute odds to a Machine Learning based diagnostic algorithm in a meaningful way. Therefore, so far the value of such an algorithm is questionable in the non-diagnosis-centered setting of primary care.