Study design, duration, population and setting

This hospital-based longitudinal cohort study was conducted between March and June 2022 among adults (≥ 18 years) admitted in medical wards at BMC in Mwanza, Tanzania. BMC serves as teaching, consultancy and referral hospital for the Lake Zone regions (Mwanza, Shinyanga, Mara, Geita, Simiyu, Kagera and Kigoma) with 950 beds. Medical department at BMC has 188 beds distributed in private wing (n = 38), general male ward (n = 70), general female ward (n = 70) and adult intensive care unit (AICU; n = 10). A minimum sample size of 311 patients was calculated by [n = (Zα/2 +Zβ)2 × (p1 (1 - p1) + p2 (1 - p2) ÷ (p1 - p2)2] at confidence level of 90% and power of 80%, α = 0.05, β = 0.2, p1 = 5.86% and p2 = 2.1% [12].

Patients screening, enrollment and follow-up

All patients who consented to participate in the current study were screened for markers of infectious diseases such as full blood count and urinalysis at admission. This was done to rule out patients with subclinical infections before 48 h of admissions. The definition and screening of patients with clinical diagnosis of HCAIs based on previously documented protocol by HALT-3 (healthcare-associated infection in long-term care facilities) version 2.1 [13]. Other criteria for clinical diagnosis of HCAI were: all clinical symptoms and signs had to be new, non-infectious causes of signs and symptoms were excluded, suspicious of HCAI was not based on a single evidence as previously reported by Eilers et al., in 2012 [14]. In the current study, screening for HCAIs included clinical evaluation i.e., vital signs (body temperature, pulse rate, respiration rate and blood pressure) and laboratory investigations i.e., full blood count and urinalysis. Patients with normal (or negative) findings/results on screening tests were enrolled in the study and then followed-up on daily basis from 48 h after admission until final outcome (hospital discharge or death). Pre-tested structured questionnaire was used for collection of socio-demographic and clinical data whereas checklist was used during patient follow-up. However, after discharge patients were contacted by phone after every 7 days within 1 month to rule out if they develop HCAIs after hospital discharge. During follow-up, all patients who developed clinical presentation of HCAIs, the appropriate specimens e.g., urine for UTI-HCAI and blood for BSI-HCAI, were collected for aerobic bacteriological culture and antimicrobial susceptibility testing of isolated bacterial pathogens.

Culture, biochemical identification, and antibiotic susceptibility testingCulture

Respective clinical specimens from patients with clinical diagnosed patients of HCAIs were processed for implicating bacterial pathogens by conventional culture methods in Microbiology research laboratory at Catholic University of Health and Allied Sciences (CUHAS). Samples were directly inoculated on solid culture media (i.e., 5% sheep blood agar, chocolate agar, MacConkey agar, and Sabouraud dextrose agar; HiMedia, Maharashtra, India) except for blood samples which were firstly inoculated into brain heart infusion (BHI) broth by 1:10 and incubated for 24 h before being inoculated onto solid culture media [15, 16]. Negative blood cultures were monitored for 5 days before a final report [17]. Furthermore, urine samples were quantitatively inoculated onto solid culture media within 2 h after collection to rule out possible microbial contaminations [15, 16]. All incubations were done in ambient air at 35 ± 2 0C for 24–48 h.

Biochemical identification testing

Microbes’ colony morphology and characteristics on culture media were recorded and then Gram staining was performed to group bacteria into two major groups, Gram-positive cocci and Gram-negative rods, as well as Gram-positive budding yeast prior to biochemical identification testing (of bacteria) as previous documented [18]. Gram-negative rods were tested for sugar fermentation, CO2 and sulfur production on triple sugar ion agar; sulfur and indole production, and motility on sulfur-indole-motility agar; utilization of sodium citrate as the sole source of carbohydrate on Simmons citrate agar; and production of urease enzyme on Christensen’s urea agar. Gram-positive cocci were tested for catalase, coagulase and DNAse production; and sensitivity towards novobiocin and bacitracin; and hydrolysis of aesculin in presence of bile on bile aesculin agar. The media used for biochemical testing and identification of bacteria were from HiMedia, Maharashtra, India.

Antibiotic susceptibility testing (AST) and phenotypic detection of ESBL-GNB and MRSA

Disk diffusion method by Kirby-Bauer technique [19] was used for antibiotic susceptibility testing. Pure and fresh bacterial colonies were suspended in sterile 0.85% physiological saline and the turbidity of the suspension was adjusted to 0.5McFarland on Densi-CHECK device (bioMérieux, SA). Bacterial suspensions were inoculated on entire surfaces of Mueller Hinton agar (MHA; HiMedia, Maharashtra, India) plates, then antibiotic disks were seeded within 15 min. All antibiotic disks used by the laboratory were produced by HiMedia, Maharashtra, India. Ampicillin (AMP 10 µg), trimethoprim-sulfamethoxazole (STX 1.25/23.75 µg), ceftriaxone (CRO 30 µg), ceftriaxone-sulbactam (CIS 30/15µg) ciprofloxacin (CIP 5 µg), gentamicin (GEN 10 µg), nitrofurantoin (NIT 300 units; for urine isolates only), amikacin (AK 10 µg) and meropenem (MEM 10 µg) were tested against Gram-negative bacteria (GNB). Whereas, STX 1.25/23.75 µg, CIP 5 µg, GEN 10 µg, NIT 300 units, erythromycin (ERY 15 µg), clindamycin (CLI 2 µg) and linezolid (LZD 30 µg) were tested against Gram-positive bacteria (GPB). Inoculated and seeded plates of MHA were incubated aerobically at 37 ± 2℃ for 16–18 h. Zones of inhibitions around antimicrobial disks were measured in millimeter and interpreted to sensitive or intermediate or resistant by using cut-off values from Clinical and Laboratory Standards Institute (CLSI) 2022 [20]. However, in the current study we interpreted zones of inhibitions of CIS 30/15µg based on cut-off values for CRO 30 µg by CLSI 2022.

Strains of S. aureus with zone diameter of ≤ 21 mm around FOX 30 µg disk were interpreted as MRSA [20]. On the other hand, GNB with resistance towards ceftriaxone were phenotypically confirmed for ESBL production by examining a zone difference between CRO 30 µg and CIS 30/15µg disks. A zone difference of ≥ 5 mm was interpreted as ESBL phenotype. ESBL phenotypes were further confirmed for carriage of genes encoding for ESBL production however our findings were published previously [21].

Definition of technical term

Laboratory confirmed HCAI refers to isolation of pathogenic bacteria from biological specimen (e.g., blood, urine and pus) following diagnosis of HCAI based on clinical presentations.

Data management and analysis

Data were entered into Microsoft excel for cleaning and coding, then imported to STATA software version 15.0 for analysis. Categorical variables were presented in percentages and fractions while continuous variables were presented in median (interquartile range; IQR). Independent t-test was used to compare mean age and mean days of hospitalization between patients without and with clinical diagnosis of HCAIs. A p-value of < 0.05 at 95% confidence interval [95%CI] was considered statistically significant.