Levy Mwanawasa Medical University

BackgroundThe rapid emergence of generative artificial intelligence (AI) tools such as ChatGPT is transforming teaching and learning practices in higher education. This study assessed the awareness, attitudes, and usage patterns of ChatGPT among university students in Zambia and examined factors associated with students’ attitudes toward the technology.MethodsA multi-institutional cross-sectional study was conducted among 1,829 university students in Zambia using a structured questionnaire adapted from instruments informed by the Technology Acceptance Model (TAM). Data were analysed using SPSS version 26.0, with statistical significance set at p < 0.05.ResultsAmong the 1,829 participants, 81.7% were aged 18–25 years, 52.5% were female, and 92.4% were unmarried. Overall, 96.8% of the students had heard of ChatGPT, and 85.6% reported having used it before this study. Among AI-usage constructs, 74.8% of respondents perceived high risks associated with AI use, 73.3% perceived ChatGPT as easy to use, 64.0% perceived it as useful, and 56.8% reported behavioural intention to use AI tools. In multivariable analysis, students aged ≥40 years were more likely to report positive attitudes toward ChatGPT compared with those aged 18–25 years (aOR = 5.91; 95% CI: 1.23–28.33; p = 0.026). Technology/social influence was also significantly associated with positive attitudes (aOR = 2.04; 95% CI: 1.55–2.68; p < 0.001). Conversely, perceived risks were associated with lower odds of positive attitudes (aOR = 0.57; 95% CI: 0.43–0.75; p < 0.001). Regarding ChatGPT use, perceived usefulness significantly predicted higher usage (aOR = 1.49; 95% CI: 1.17–1.89; p = 0.001), while perceived risks were associated with reduced usage (aOR = 0.61; 95% CI: 0.45–0.84; p = 0.003).ConclusionAwareness and use of ChatGPT are widespread among university students in Zambia, with more than four out of five students reporting prior use. Perceived usefulness and social influence were positively associated with adoption, whereas perceived risks were linked to lower attitudes and reduced use. These findings highlight the need for higher education institutions to develop clear policies and integrate AI literacy into curricula to promote responsible and effective use of generative AI technologies in academic environments.

Klebsiella pneumoniae is a leading cause of neonatal sepsis in low- and middle-income countries, with antimicrobial resistance (AMR) significantly contributing to mortality. We used whole genome sequencing to explore the impact of an infection prevention and control (IPC) intervention on K. pneumoniae strains and transmission dynamics responsible for sepsis in a Zambian neonatal unit. Blood culture isolates were collected during the Sepsis Prevention in Neonates in Zambia (SPINZ) study, including a 7-month baseline period and 12 months following implementation of a low-cost IPC bundle. K. pneumoniae genomes associated with 411 neonatal infections were characterised, comprising 24 unique sequence types (STs) and dominated by ST307 (69.3%, n = 285). Nearly all isolates (99.0%) carried extended spectrum beta-lactamases, but few carried carbapenemases (2.7%). Most infections (95.6%) were associated with probable transmission clusters, ranging in size from 2-202 patients and spanning durations of 2-232 days. Most K. pneumoniae (n = 228, 70%) were isolated during the 7-month baseline period and formed six clusters, including one cluster of >200 neonates infected with ST307. Transmission of all strains was periodically suppressed by an IPC bundle; however not all strains were eliminated, and some were able to re-emerge later to re-establish infection and transmission, alongside newly introduced strains that formed additional transmission clusters. Some clusters were associated with rapid onset of disease (within 2 days of admission) and others with delayed onset, suggesting different sources of contamination (e.g., reagent vs environmental). These findings reinforce the need for sustained IPC efforts, and better understanding of environmental reservoirs of opportunistic pathogens in neonatal units to inform such efforts.

Background: Human immunodeficiency virus (HIV) recency assays are essential for distinguishing recent infections (≤ 12 months) from long-term infections, enabling targeted interventions. Western province, Zambia, has the nation’s highest HIV prevalence (16%), yet recent infection data in Mongu District remain unknown. Aim: To determine the burden of recent HIV infection and associated factors in Mongu District, Western province, Zambia. Setting: The study was conducted in Mongu district, a provincial town in Western province in Zambia. Methods: This cross-sectional analytical study was conducted from 12 February 2024 to 30 May 2024 in 19 health facilities in Mongu District offering HIV testing services (HTS), using Asante recency test kits to identify recent infections. A total of 771 newly diagnosed individuals were included, and logistic regression analysis was employed to identify factors associated with recent infections, with adjusted odds ratio (AOR) and 95% confidence interval (CI) calculated, and significance determined at a p-value 0.05. Results: Among 771 antiretroviral therapy (ART)-naïve individuals (median age 27 years; 71% female participants), 34% (n = 262) were identified as recent infections. Participants with medium (adjusted odds ratio [AOR] = 13.43, 95% confidence interval [CI]: 5.85–30.83) or low (AOR = 13.33, 95% CI: 5.67–31.33) HIV knowledge had higher odds compared to those with high knowledge (p 0.001). Unsuppressed viral load (VL) (≥ 1000 copies/mL) showed the strongest association (AOR = 852.1, 95% CI: 521.1–1432; p 0.001). Cluster of differentiation 4 (CD4) counts 200 cells/mm3 had 1848-fold higher odds (AOR = 1848.2, 95% CI: 192.8–17714.1; p 0.001). Clients who entered through the Prevention of Mother-to-Child Transmission had higher odds (adjusted odds ratio [AOR] = 2.67, 95% confidence interval [CI]: 1.11 – 6.11, p = 0.028) compared to the reference group. Conclusion: Mongu District exhibits alarmingly high recent HIV infections (34%), concentrated among youth, women, and individuals with suboptimal knowledge. Contribution: These findings underscore the urgent need for scaling recency testing, strengthening youth-focused prevention, and improving HIV literacy as critical public health priorities.

Dietary diversity is an important indicator of diet quality and nutritional adequacy in children; however, evidence among school-aged children in Zambia remains limited. This cross-sectional study assessed dietary diversity and associated factors among 638 upper primary schoolchildren aged 9–18 years from 12 schools in Lusaka District, Zambia, between August and October 2022. Data were collected using a 24-hour dietary recall, the Food Choice Questionnaire, and anthropometric measurements, and were analyzed in Stata version 14 using descriptive statistics and chi square tests at a 95% confidence level. Overall, 384 participants (60.2%) had high dietary diversity, and 254 (39.8%) had low dietary diversity. All participants consumed staple foods, but the intake of other important food groups was lower, including vegetables (57.5%), legumes (51.3%), fruits (50.2%), and animal products (49.7%). Higher dietary diversity was significantly associated with male sex (p = 0.009) and private school attendance (p = 0.004). No significant associations were found between dietary diversity and the perceived importance of vitamins and minerals, maintaining health, being nutritious, protein, good for skin and teeth, or fiber. The majority of upper primary schoolchildren in Lusaka had high dietary diversity; however, a substantial proportion still had low dietary diversity, highlighting the need for targeted school- and community-based nutrition interventions.

This study was conducted over 24 months (September 2021-Agust, 2023) to determine the time to resolution of Long COVID symptoms and the associated factors among people infected with SARS-CoV-2 during the 3rd wave (June - August, 2021) of COVID-19 in Zambia. Participants comprised a cohort of 303 people randomly sampled from 7 healthcare facilities from Lusaka and Southern provinces. Demographic and clinical data were collected at baseline and at a monthly interval from participants’ clinical records and interview. Descriptive analysis was conducted to summarise respondent demographic and clinical characteristics. Survival analysis (Kaplan Meier model) was conducted to determine the median time to symptom resolution, generate survival curves and the risk table. A Poisson model was fitted to compute the crude incident rate of recovery. Cox-Proportional Hazards regression analysis was conducted to determine the univariate and multivariate association between the covariates and the time to resolution of symptoms. Crude and adjusted hazard ratios, 95% confidence intervals and p-values were computed. A total of 236 events occurred during the follow- up period; the crude incidence rate was 1, 575.6 events per 1000 person years (95%CI: 1,363.2 - 1,784.4). The median time from diagnosis of Long COVID symptoms to resolution was 6 (IQR: 5-8) months. Vaccination status was significantly associated with the time to resolution of symptoms (aHR= 1.71, 95% CI: 1.02 - 2.86, p=0.04) after adjusting for age, number of children, occupation, pre-existing medical conditions, number of vaccine doses and hospital admission. Other variables did not show statistically significant association with the time to resolution of symptoms (p>0.05). These findings suggest that vaccination is associated with a shorter time to resolution of Long COVID symptoms. Netherless, causation can not be ascertained due to the observational design of the study. A study with an experimental design is needed to establish the causal relationship.