No sensitivity analysis was carried out. In addition, the authors highlighted a number of limitations to their study:
(1) hospital stay and death are only indirect measures of the severity of PCP
(2) patients receiving prophylaxis may have better access to medical care than patients not receiving prophylaxis
(3) patients not receiving primary care may have concurrent untreated illnesses that could contribute to mortality or prolonged hospital stay
(4) the need for hospitalization is not necessarily a good indicator of severity of disease
(5) the selection criteria used would tend to identify a disproportionate number of sicker patients requiring hospitalisation.