This study aimed at determining the rationality of antibiotic prescriptions and associated in-patient treatment outcomes in children aged 2–59 months with severe pneumonia at Bwizibwera Health Center IV, Mbarara District in south-western Uganda from 1st May 2018 to 30th April 2019.

In this study, 75.1% of antibiotic prescriptions were irrational among children under five in Bwizibwera HC IV, Mbarara District. Another study conducted in Western Uganda revealed 61.9% of antibiotic prescriptions [5]. These high percentages could be attributed to availability of antibiotics at the time of prescription. Our study was conducted in Government health facility with frequent stock out of essential medicines among others antibiotics, while in the other study of Akunne [5] was done in a private hospital with availability of antibiotics of choice.

Earlier studies found that despite the availability of treatment guidelines, Uganda’s health care system is still challenged with high rates of irrational antibiotic use [17,18,19]. This could be due to poor implementation of government policies and guidelines which at the end affects rationality of antibiotics prescription.

In the studies conducted in Turkey and Mongolia slightly smaller percentages of irrational antibiotic prescriptions were reported (56.5%) and (56.6%) respectively [7, 8]. The study population in Turkey and Mongolia were higher compared to our current study population in Uganda. This could preliminarily explain the observed differences.

In other countries, irrational prescriptions of antibiotics were observed among children [3, 6, 9] of 33.4%, 35.1% and 46% respectively. These differences could be accorded to difference in geographical location, treatment guidelines, competence of prescribing staff and availability of antibiotics.

Our study revealed that 24.9% of antibiotic prescriptions were rational. This was based on the right regimen, right duration and frequency of drug administration. The percentage of rational antibiotic prescriptions in the current study is lower than that reported earlier in Tanzania of 44% [12].The difference in the reported percentage of rational prescriptions by the study in Tanzania and our study could have risen due to the fact that the Tanzanian study was a multi-center study and involved patients with several disease conditions other than just pneumonia.

A study conducted in public health care facilities in Uganda reported that rational prescription was 12.4% [14] and this is lower than the 24.9% revealed by our study. The different in percentages could be due to antibiotic prescription in one condition and in one health facility while Trap et al., [11] looked at all levels of health care facilities in Uganda and the general performances in the country.

In Turkey, the rate of rational antibiotic use was reported to be 11.3% [7]. While in our current study rational antibiotic prescription was 24.9%. This could be attributed to difference in treatment guidelines between Uganda and Turkey and also difference in study design.

The outcomes of antibiotic treatment were categorized into good and unfavorable outcomes in the current study. A child was considered to have a good treatment outcome if he or she improved and was discharged within 7 days.

Unfavorable treatment outcomes were considered when there was development of complications, referral to the hospitals, self-discharged and deaths.

This study reported 2 deaths (20%) out of 10 unfavorable outcomes which is comparable with a study in Indonesia of 7 (15.2%). This borderline similarity could be due to poor choice of the antibiotics, wrong dosage, dose and route of administration and other empirical error.