Impact of Point-of-Care Rapid Diagnostic Tests on Antibiotic Prescription Among Patients Aged <18 Years in Primary Healthcare Settings in 2 Peri-Urban Districts in Ghana: Randomized Controlled Trial Results

The overprescription of antibiotics is a significant driver of antimicrobial resistance (AMR), a major global health threat. In low- and middle-income countries (LMICs) like Ghana, where diagnostic capacity is often limited, acute febrile illnesses are frequently treated with antibiotics regardless of the underlying cause (just-in-case prescribing). This practice contributes to AMR and suboptimal treatment of infections. The proportion of antibiotic prescriptions in Ghanaian healthcare facilities is significantly high, ranging from 65% to 82% in primary facilities. Several studies have highlighted multidrug resistance to commonly prescribed antibiotics within the country. This inappropriate antibiotic use leads to increased morbidity and mortality, compounded by issues of non-adherence to prescriptions, including symptom remission, forgetfulness, and negative attitudes towards antibiotic use. Addressing this requires adapting case management of infectious diseases at the primary healthcare level. The introduction of point-of-care (POC) diagnostic tools, along with revised treatment guidelines, offers a potential solution. POC diagnostic testing (POCT), particularly rapid diagnostic tests (RDTs), can reduce uncertainty in diagnosing and treating infections, leading to a decrease in inappropriate antibiotic prescriptions in the short term, and to a reduction in bacterial resistance in the longer term. While microbiological diagnostic tests are used in hospitals, they are generally unavailable at the primary care level in LMICs. This study aimed to evaluate the effect of a package of POCTs, alongside clinical algorithms and communication training, on antibiotic prescriptions and clinical outcomes in children and adolescents presenting with acute febrile illness in Ghana.

The literature extensively documents the global burden of bacterial antimicrobial resistance and its devastating impact on mortality. The World Health Organization (WHO) highlights the urgent need to address this crisis. In Ghana, multiple studies have shown alarmingly high rates of antibiotic prescription in both primary and secondary/tertiary care settings. The prevalence of multidrug-resistant pathogens has also been reported through nationwide surveillance. Existing research emphasizes the link between inappropriate antibiotic prescribing and the rise of AMR. Furthermore, poor patient adherence to prescribed antibiotics contributes to the problem. Studies have identified various factors influencing adherence, including patient knowledge, attitudes, and the perceived severity of symptoms. The existing literature also supports the potential of POCTs to guide appropriate antibiotic use, reducing both inappropriate prescribing and AMR. Previous trials have shown the effectiveness of POCTs for specific infections like respiratory tract infections. However, the effectiveness of a comprehensive package of POCTs, coupled with clinical algorithms and communication strategies within a primary care setting in a LMIC context, remained largely unexplored prior to this study.

This open-label, centrally randomized controlled trial (RCT) was conducted in four public health facilities across two peri-urban districts in Ghana (Shai-Osudoku and Ningo-Prampram) between September 2020 and September 2021. The study enrolled children and adolescents aged 6 months to under 18 years presenting with non-severe acute febrile illness (fever >37.5°C or history of fever within 7 days). Participants were randomized 1:1 to either an intervention or a control arm. The intervention comprised a package of pathogen-specific and non-pathogen-specific POCTs for common fever causes in Ghana, a clinical and diagnostic algorithm, and training and communication (T&C) packages to improve adherence to prescriptions. POCTs included RDTs for malaria, influenza A/B, respiratory syncytial virus (RSV), Streptococcus pneumoniae, group A streptococci, and Salmonella Typhi, along with CRP and white blood cell (WBC) counts. The algorithm guided antibiotic prescription based on POCT results and clinical findings. The T&C package included training for healthcare workers on the algorithm and patient-centered communication strategies to enhance adherence. The control arm received standard-of-care. The primary outcomes were the proportion of patients with favorable clinical outcomes at day 7 and antibiotic prescription at day 0. Secondary outcomes included adherence to the algorithm by healthcare workers and adherence to prescriptions by patients. The sample size was calculated to detect a 30% reduction in antibiotic prescriptions in the intervention group, but the final sample size was 1512 (761 intervention, 751 control) due to lower than expected recruitment, necessitating adjustments to the statistical analysis plan to maintain sufficient power. Data were collected using offline electronic case report forms and analyzed using R version 4.2.1. Descriptive analysis, chi-square tests, and relative risk calculations were performed.

A total of 1512 patients were enrolled (53.5% male, 76.3% <5 years old). The most common symptoms were fever (94.4%) and cough (64%). Most participants (74%) received a presumptive diagnosis of respiratory illness. In the intervention arm, most POCTs were performed, though positivity rates were generally low for pathogen-specific tests. Malaria was found in 11.4% of intervention participants and 11.2% of controls. The most common identified infections were influenza A/B (10%), S. pneumoniae (5.1%), group A Streptococcus (4.6%), RSV (2.8%), and typhoid (1.6%). By day 7, almost all participants (99.7% intervention, 99.4% control) had favorable outcomes. The overall relative risk (RR) of antibiotic prescription in the intervention arm was 0.89 (95% CI, .79 to 1.01), not statistically significant. However, significant reductions were observed in subgroups: a 14% reduction in children under 5 (RR, 0.86; 95% CI, .75 to .98), a 15% reduction in non-malaria patients (RR, 0.85; 95% CI, .75 to .96), and a 16% reduction in patients with respiratory symptoms (RR, 0.84; 95% CI, .73 to .96). Adherence to antibiotic prescriptions was high in both arms (approximately 90% by pill count). Adherence to the algorithm by healthcare workers showed inconsistencies, with some inappropriate prescriptions. Four participants were withdrawn due to severe adverse events.

While the overall reduction in antibiotic prescriptions in the intervention group was not statistically significant, the substantial reductions observed in specific subgroups strongly suggest the benefit of the intervention package. The findings align with studies in Burkina Faso and Uganda conducted during the same period. Although the reduction was smaller than anticipated, it still represents a meaningful step toward curbing inappropriate antibiotic use and combatting AMR. The antibiotic prescription rates in both arms (around 40%) remain higher than the WHO-recommended level of <30%, highlighting the ongoing need for interventions to optimize antibiotic stewardship in primary care. The high adherence rates observed in both arms suggest other factors might influence antibiotic use beyond the intervention package. The study period coincided with the COVID-19 pandemic, impacting patient presentation and potentially affecting the prevalence of tested infections, particularly respiratory illnesses. Despite these limitations, the results support the integration of affordable and accessible POCTs into primary healthcare systems to improve the quality of clinical decision-making regarding antibiotic use.

This RCT demonstrated that a package of POCTs, diagnostic algorithms, and communication training can effectively reduce antibiotic prescriptions in primary healthcare settings, especially in children under 5 and those with non-malarial respiratory infections. Policymakers in LMICs should prioritize the integration of simple, affordable POCTs into primary care to guide appropriate antibiotic use and combat antimicrobial resistance. Future research should explore the long-term impact of this intervention on AMR and investigate the factors contributing to high adherence rates irrespective of treatment arm.

The study faced limitations, including a lower than anticipated recruitment rate, leading to a smaller sample size than initially planned. This reduction impacted the power of the study to detect smaller differences. The study period coincided with the COVID-19 pandemic, potentially altering the usual pattern of respiratory infections. The low positivity rates for several pathogen-specific POCTs might reflect the pandemic's impact on disease prevalence. Potential contamination between the control and intervention arms, though efforts were made to minimize it, could also have influenced the results. The observed high adherence rates, irrespective of intervention, warrant further investigation to determine the underlying factors.