Discussion

In our cross-sectional survey, we evaluated the knowledge and attitudes regarding antimicrobial resistance and antimicrobial stewardship among nurses, pharmacists, and physicians working at a national cancer center in sub-Saharan Africa. Most respondents considered antibiotic knowledge very important to their clinical job and desired additional training in antibiotic use. We found that nurses were less familiar with the terms “antimicrobial resistance” and “antimicrobial stewardship” than pharmacists or physicians. Finally, we identified numerous barriers to diagnosing infections, including the ability to regularly monitor patient temperatures and to obtain blood cultures. These findings will inform our ongoing efforts to develop ASPs that account for the unique barriers encountered at UCI and in other cancer treatment programs throughout sub-Saharan Africa.

A primary goal of antimicrobial stewardship is to decrease the emergence, selection, and spread of antimicrobial resistance by optimizing antimicrobial use. ^{10,Reference Fishman15} To this end, it is important to educate clinicians about the causes of antimicrobial resistance. ^{Reference Mendelson and Matsoso16} More than 80% of our survey respondents recognized antimicrobial resistance as a significant problem at UCI, which was higher than in previous surveys of healthcare workers from sub-Saharan Africa. ^{Reference Abahamye17,Reference Gebrehiwot and Tadiwos18} This finding may reflect our ongoing efforts to educate staff about the high prevalence of MDR gram-negative bacterial infections at UCI. ^{Reference Lubwama, Phipps and Najjuka2,Reference Lubwama, Adams and Muwonge19} Among the UCI staff, there was general recognition that inappropriate antibiotic use contributes to antimicrobial resistance. Hand hygiene was not as well recognized as a contributing factor and represents an opportunity for improved education. Similar surveys from sub-Saharan Africa show that hand hygiene is an underrecognized contributor to resistance. ^{Reference Abahamye17,Reference Gebrehiwot and Tadiwos18,Reference Kheder20} Yet, good hand hygiene reduces transmission of resistant bacteria within the healthcare setting. ²¹ For example, in a recent study from Mulago National Referral Hospital in Kampala, resistant gram-negative bacteria primarily spread through direct contact between patients and healthcare providers or nonmedical caretakers. ^{Reference Mboowa, Sserwadda and Bulafu22} It is also important to address environmental limitations that occur in low-resource settings (eg, lack of alcohol-based hand gel, inadequate handwashing facilities), which remain an ongoing challenge at UCI.

The appropriate use of microbiologic diagnostics to guide therapeutic decisions, or diagnostic stewardship, is another important aspect of antimicrobial stewardship. ²³ Diagnostic stewardship includes the ability to rapidly identify patients with suspected infections and to obtain the appropriate microbiologic laboratory tests. The ability to consistently monitor temperatures was seen as a key limitation to identifying patients with infections at UCI. In high-resource settings, temperatures are routinely measured 3–4 times daily and are, therefore, not included as a core component of diagnostic stewardship. However, in sub-Saharan Africa, 1 nurse may care for 30–70 patients, ^{Reference Mula, Middleton, Muula, Solomon and Varga24} making it challenging to consistently measure vital signs. Developing strategies that account for staffing limitations could improve fever detection, facilitate blood culture collection, and decrease time to antibiotics. We are currently collaborating with UCI nurses to train patient family members to measure temperatures and alert nurses when a fever occurs. Since time from fever onset to the administration of guideline-recommended antibiotics is associated with outcomes, ^{Reference Koenig, Schneider, Morgan, Ammann, Sung and Phillips25} improved fever detection could decrease mortality for patients receiving cancer treatment in sub-Saharan Africa.

Obtaining blood cultures for patients with suspected infections is also an important component of diagnostic stewardship. Blood cultures decrease the use of unnecessary broad-spectrum antibiotics by allowing clinicians to tailor a patient’s antibiotic therapy to the causative organisms. ^{10,23,Reference Koenig, Schneider, Morgan, Ammann, Sung and Phillips25} They are also critical for developing a hospital antibiogram, in which the antimicrobial susceptibilities of organisms isolated at that hospital are aggregated into easily referenced tables. ^{10,23,Reference Jacobs, Hardy and Semret26} The antibiogram allows clinicians to track changes in antibiotic susceptibility and adapt evidence-based institutional treatment guidelines accordingly. ^{Reference Pollack and Srinivasan27} In our study, patient cost was the most frequently identified barrier to obtaining blood cultures, which is consistent with findings from previous studies in sub-Saharan Africa. ^{Reference Cox, Vlieghe and Mendelson28} In Uganda, blood cultures cost up to $10.00 USD (ie, 4–5 days’ average wages), ²⁹ which is prohibitively expensive. Since blood cultures are essential for diagnosing and treating infections in patients with cancer, policy makers should consider these a core component of cancer care and factor these into the cost of cancer treatment.

To improve rational antibiotic decision making, it is necessary to understand provider prescribing practices. Among UCI clinicians, patient white blood cell count was most frequently considered to be “very important” when choosing antibiotics. This is fitting because patients with neutropenic fever are at high risk of infection-related complications. ^{Reference Freifeld, Bow and Sepkowitz11} Fewer respondents identified HIV status or CD4 count as important. However, there is evidence that patients with HIV are at higher risk of developing treatment-related infections than their HIV negative counterparts. ^{Reference Park, Kim and Hwang30} Those with HIV have high rates of antibiotic exposure ^{Reference Nwosu, Avershina, Wilson and Rudi31,Reference Gaskell, Feasey and Heyderman32} and are at increased risk of developing infections with resistant bacteria. ^{Reference Gaskell, Feasey and Heyderman32,Reference Mulu, Yizengaw and Alemu33} The microbiology of febrile illness also varies by HIV status. For example, in sub-Saharan Africa tuberculosis is a leading cause of sepsis among those with HIV. ^{Reference Moore, Jacob and Banura34} Because one-third of patients with cancer in sub-Saharan Africa also have HIV, ^{Reference Bender Ignacio, Ghadrshenas, Low, Orem, Casper and Phipps35} understanding the relationship between HIV status and treatment-related infections will inform locally relevant guidelines for cancer-related infections.

Educational interventions improve healthcare workers’ ability to prescribe the appropriate antibiotics. ¹⁰ Almost two-thirds of our respondents received antibiotic teaching within the past year. Most received teaching during clinical care activities and used Internet searches and discussions with colleagues to answer antibiotic questions, indicating that just-in-time learning was a primary educational strategy. Respondents also considered readily accessible materials such as smartphone applications, teaching on ward rounds, and informational handouts to be “very helpful” learning formats. Although continuing medical education is traditionally delivered through didactic lectures, ^{Reference Sullivan36} active learning strategies (eg, audit and feedback, case-based learning) and multimodal interventions increase knowledge retention and translation into clinical practice. The COVID-19 pandemic highlights the importance of using multimodal strategies, particularly for those living in low-resource settings. Online learning platforms are expensive and challenging to use when there is inconsistent Internet access. On-demand mobile tools (eg, smartphone apps) increase access to critical information during routine patient care and improve knowledge sharing across educational sites. ^{Reference Lall, Rees, Law, Dunleavy, Cotič and Car37} Incorporating these tools into our ASP could improve evidence-based infectious disease management at our center.

Overall, in our study, nurses were less familiar with the terms “antimicrobial resistance” and “antimicrobial stewardship” than physicians or pharmacists. At UCI and in other hospitals in low-resource settings, nurses are often responsible for identifying patients with infections and initiating antibiotics. Thus, it is critical for oncology nurses to understand how to select the appropriate antibiotics. ^{Reference Mula, Middleton, Muula, Solomon and Varga24} Most stewardship interventions target physicians and pharmacists; thus, nurses are underutilized members of the stewardship team. In high-resource settings, nurse-driven protocols have been developed to facilitate blood culture collection and antibiotic initiation for patients with neutropenic fever. ^{Reference Salter38} In sub-Saharan Africa, similar protocols have been developed to manage bacterial meningitis. ^{Reference Wall, Mukaka and Denis39} These protocols improve guideline adherence, decrease time-to-antibiotics, and improve patient outcomes. Developing nurse-led protocols and training nurses in the principles of antimicrobial stewardship are ways to adapt ASPs for patients receiving cancer treatment in low-resource settings.

Our study had several limitations. This survey was completed at a single cancer center in sub-Saharan Africa, and the results may not be representative of other cancer centers. Given the numerous infectious diseases initiatives underway at UCI, knowledge of antimicrobial resistance and antimicrobial stewardship may be higher at UCI than in other cancer treatment programs. Those who responded may also have more training or inherent interest in the topic. In addition, some respondents did not answer every question, which influenced our ability to compare answers across questions. Because the survey took place during the COVID-19 pandemic, the frequency of education, concerns about supply availability, and issues with antibiotic shortages may not be typical. However, studies completed in low-resource settings before the COVID-19 pandemic had similar findings regarding antimicrobial education and antibiotic supply availability. ^{Reference Gebrehiwot and Tadiwos18,Reference Knowles, Sharland and Hsia40}

In this survey, we evaluated the knowledge and perceptions of antimicrobial resistance and stewardship among physicians, nurses, and pharmacists working at a single cancer hospital in sub-Saharan Africa. UCI staff were knowledgeable about the term antimicrobial resistance and eager to learn more about antibiotic use. We also identified several areas for targeted interventions. These include increasing education regarding the importance of hand hygiene and appropriate antibiotic selection, allocating resources for blood cultures, implementing strategies to improve fever detection, and incorporating nurses as key members of the antimicrobial stewardship team. Our findings will inform the development of ASPs for cancer treatment programs in low-resource settings.