Situational analysis of antibiotic prescriptions in Kenyan neonatal units for antimicrobial stewardship: a retrospective longitudinal study

Abstract

Background: High antibiotic use in neonatal units may drive antimicrobial resistance and cause harm including mortality. We used data from 22 Kenyan neonatal units to (1) describe the proportion with antibiotic prescriptions at admission; (2) assess the predictors of non-first line antibiotic prescription; (3) estimate antibiotic use, and (4) explore postadmission antibiotic switching.

Methods: Retrospective longitudinal study from 1st September 2020 to 31st October 2023. Antibiotics were classified as first line (penicillin plus gentamicin only), third generation cephalosporins (ceftazidime or ceftriaxone) or others. The proportion of antibiotic prescriptions were computed, and a multilevel logistic regression model used to analyse predictors of non-first line prescription. Antibiotic use was quantified by days of therapy (DOT) and length of therapy (LOT).

Findings: Most neonates-62.6% (51,883/82,834)- received at least one antibiotic prescription at admission. Overall, first line antibiotics constituted 86% (44,636/51,883) but third generation cephalosporin use reached 100% in two facilities temporarily. The odds of non-first line prescription was greatest for outborn neonates (Odds ratio 2.27, 95% CI 2.12-2.43) while the estimated antibiotic consumption was 418 (389-500) per 1000 patient days by LOT and 744 (691-869) by DOT. From exploratory data post admission switching was most commonly to third generation cephalosporins.

Interpretation: There is a high use of antibiotics potentially related to severity of illness at admission. Adherence to national guidelines for first line antibiotics is however generally high. Estimation of neonatal antibiotic prescription patterns and use over time and place is feasible and will be important in assessing the effectiveness of antimicrobial stewardship in Kenya and elsewhere in reducing antimicrobial resistance.

Funding: This work was funded by the Wellcome Trust.

Keywords: Antibiotic use: antimicrobial stewardship; Antibiotics; Neonatal.

Fig. 1

Patient flow diagram.

Fig. 2

Top four diagnoses for neonates with antibiotic prescriptions. The areas of overlap are restricted to the top four diagnoses and depict instances where two or more of these were documented for a group of patients e.g. 1166 had low birth weight, respiratory distress syndrome, and neonatal sepsis. Where there is no overlap shows those who had only one of the top four diagnoses. Patients in either of these two categories (overlap versus no overlap) may also have another diagnosis in the “others” category.

Fig. 3

Trend of antibiotic prescriptions at admission from September 2020 to October 2023.

Fig. 4

Variation in antibiotic prescription patterns at admission in nine selected hospitals over time. Individual dots represent the monthly proportion and the bold lines the trend line by LOWESS technique. A-I is sequential labelling of the nine selected hospitals in this figure while H e.g., H1 corresponds to all 22 hospitals in supplementary figure 1. CIN, Clinical Information Network; First Line, penicillin and gentamicin. Third Gen, third generation; Others, other antibiotics.

Fig. 5

Exploring post admission antibiotic switching amongst patients with an antibiotic prescription at admission. First line: Gentamicin and Penicillin prescribed together without other antibiotics; Partial: Gentamicin OR Penicillin only prescribed (but not both); Mixed: Gentamicin OR Penicillin prescribed together with other antibiotics; Others: Any other antibiotic.