Table 3. Inclusion, exclusion and diagnosis selection criteria for each dataset ^*.

	Inclusion criteria (ICD-10 codes in brackets[54])	Definition- if not met, entry excluded
Malaria dataset	Diagnosis of ‘malaria (free text)’ or ‘plasmodium falciparum malaria (B50)’ or positive blood film	History of fever or fever on presentation AND positive blood film
LRTI dataset	Diagnosis of ‘pneumonia (J18)’ or ‘bronchiolitis (J21)’ or ‘Acute bronchitis (J20)’ or ‘acute respiratory infection (free text)’ or ‘resolving pneumonia (free text)’	Cough or shortness of breath or chest pain or breathless AND raised RR or crackles or crepitations or decreased air entry or wheeze or signs of severe disease i.e. chest indrawing or nasal flaring or grunting or signs of dehydration or head nodding or lethargy or decreased coma score
Diarrhoeal disease dataset	Diagnoses- ‘unspecified bacterial intestinal infection (A04.9)’ or ‘giardia (A07.1)’ or ‘viral gastroenteritis (A08)’ or ‘diarrhea (free text)’ or ‘diarrheal disease (free text)’ or ‘non infective diarrhoea (K52.9)’	Diarrhoea listed as a presenting complaint

*Numbers in brackets are ICD-10 codes