Evaluation of Systemic Antifungal Prescribing Knowledge and Practice in the Critical Care Setting among ICU Physicians and Clinical Pharmacists: A Cross-Sectional Study

Management of invasive fungal infections (IFI) and subsequent treatment choices remain challenging for physicians in the ICU. Documented evidence shows increased practice of the inappropriate use of antifungal agents in the ICU. Continuous education of healthcare providers (HCPs) represents the cornerstone requirement for starting an antifungal stewardship program (AFS). This study aimed at evaluating knowledge gaps in systemic antifungal prescribing among physicians and clinical pharmacists in a critical care setting. A cross-sectional, multi-center, survey-based study was conducted in five tertiary hospitals located in Al-Ahsaa, Saudi Arabia between January and May 2021. A self-administered questionnaire was distributed among the targeted clinicians. A total of 63 clinicians were involved (65.5% ICU physicians and 34.5% clinical pharmacists). It was noted that a minority of the participating HCPs (3.2%) had overall good knowledge about antifungal prescribing, but the majority had either moderate (46%) or poor (50.8%) knowledge. The difference in overall knowledge scores between the ICU physicians and the clinical pharmacists (p = 0.925) was not significant. However, pharmacists showed better scores for the pharmacokinetics of antifungal therapy (p = 0.05). This study has revealed a significant gap in the knowledge and practice of clinicians as regards prescribing antifungal therapy in our area. Although the results cannot be generalized, the outcome of this study has exposed the need for a tailored training program essential for carrying out an AFS program.


Introduction
Fungal infections account for about 20% of documented intensive care unit (ICU) infections, of which Candida infections represent approximately 70 to 90% of total invasive fungal infections (IFIs) [1]. The incidence of candidemia has been rising over the decades due to an increase in the number of immunocompromised and critically ill patients [2]. Along with the steady growth of fungal infections among critically ill patients, there has been a relative rise in non-albicans Candida infections, these species exhibit resistance to azole antifungals and represent more challenging clinical situations [3,4]. In Saudi Arabia, studies have reported a high frequency of non-albicans candidaemia and an alarming rise of fluconazole-resistant Candida parapsilosis [4,5]. The antifungal resistance coinciding with a change in the epidemiologic pattern of candidemia identified in one study was described as alarming [5]. An outbreak report described the emergence of a cluster of Candida auris cases in the ICU of King Khalid Hospital [6].  Table 2, Section A, details the assessment of the knowledge and practices of the clinicians about antifungal therapy. The section included 11 questions, the first five of which assessed clinicians' knowledge and practices about empiric antifungal therapy in the ICU. According to the results of question one, which provided four case scenarios to assess the decision on starting empiric antifungal therapy, the majority (74.6%) of the participants adequately chose to immediately start empiric antifungal therapy for ICU patients who were still febrile after administration of broad-spectrum antibiotics; however, only 23.8% would start empiric antifungal therapy for dialysis patients exhibiting signs of septic shock ( Figure 1). These two clinical situations necessitated the initiation of antifungal therapy. On the other hand, 33.3% and 22.2% of participants' responses chose to start upon a positive β-D-glucan test and heavy growth of Candida species in urine culture, respectively, even though these treatment choices represent inappropriate utilization of antifungal therapy ( Figure 1). Pertaining to question 2, as shown in Table 2, those who spared the review of prior azole exposure for stable patients only constituted a minority (6.3%), but the majority (60%) mistakenly thought it should be checked for most patients. About one-quarter (25%) of clinicians stated that they face difficulties reviewing patients' histories to check for prior azole exposure. Question 3 assessed the first-line choice of empiric antifungal agent in case of suspected invasive candidiasis in a critically ill non-neutropenic patient. More than two-thirds of clinicians (68.3%) correctly identified echinocandin as the first-line choice ( Figure 2). Question 4 assessed the appropriate duration of antifungal therapy and 44.8% of clinicians correctly agreed that it requires 2 weeks of antifungal treatment even if the patients showed improvement ( Figure 2). Question 5 investigated clinicians' actions in case of a lack of response to antifungal therapy. Appropriately, more than three-quarters of clinicians (74.6%) agreed that antifungal therapy should be stopped if the patient showed no clinical response to the empiric antifungal treatment for 4 to 5 days ( Figure 2). Pertaining to question 2, as shown in Table 2, those who spared the review of prior azole exposure for stable patients only constituted a minority (6.3%), but the majority (60%) mistakenly thought it should be checked for most patients. About one-quarter (25%) of clinicians stated that they face difficulties reviewing patients' histories to check for prior azole exposure. Question 3 assessed the first-line choice of empiric antifungal agent in case of suspected invasive candidiasis in a critically ill non-neutropenic patient. More than two-thirds of clinicians (68.3%) correctly identified echinocandin as the first-line choice ( Figure 2). Question 4 assessed the appropriate duration of antifungal therapy and 44.8% of clinicians correctly agreed that it requires 2 weeks of antifungal treatment even if the patients showed improvement ( Figure 2). Question 5 investigated clinicians' actions in case of a lack of response to antifungal therapy. Appropriately, more than three-quarters of clinicians (74.6%) agreed that antifungal therapy should be stopped if the patient showed no clinical response to the empiric antifungal treatment for 4 to 5 days ( Figure 2).

Participants' Responses to Questions about Targeted and Prophylactic Antifungal Therapy
As shown in Table 2, Section A, questions 6-10 assessed HCPs' knowledge and practice about definitive or targeted antifungal therapy. Question 6 assessed the practice of de-escalating empiric echinocandin to fluconazole upon blood culture results revealing Candida albicans. Based on the HCPs' responses, we found that about 57% were sure that the antifungal therapy should be de-escalated to fluconazole if the patient improved after

Participants' Responses to Questions about Targeted and Prophylactic Antifungal Therapy
As shown in Table 2, Section A, questions 6-10 assessed HCPs' knowledge and practice about definitive or targeted antifungal therapy. Question 6 assessed the practice of de-escalating empiric echinocandin to fluconazole upon blood culture results revealing Candida albicans. Based on the HCPs' responses, we found that about 57% were sure that the antifungal therapy should be de-escalated to fluconazole if the patient improved after administration of an echinocandin such as caspofungin, with blood culture results revealing Candida albicans sensitive species ( Figure 3); whereas, about 29% would still proceed unnecessarily with echinocandins. In a similar situation, question 7 assessed HCPs' practice in case of patient improvement on empiric echinocandin and blood culture growing Candida glabrata species, where echinocandins should be pursued. The responses revealed that 54% believe that they should continue with an echinocandin ( Figure 3). Nevertheless, about 27% were not sure as they have not come across Candida glabratainduced infections. As for Candida krusei isolated from blood cultures in question 8, 54% would still favor continuing with echinocandin as long as the patient improved; whereas, about 27% were not sure about the appropriate action. Only 3.2% would de-escalate to voriconazole even though this is the appropriate choice ( Figure 3). Likewise, in question 9, which assessed clinicians' actions in case of confirmed candidemia, 57% were sure that it requires immediate antifungal treatment for any ICU patient with confirmed candidemia (Table 2). Logically, neither waiting for the identification of Candida species nor waiting for a repeat culture are accepted practices. Question 10 assessed targeted therapy for invasive aspergillosis, where 38% of clinicians agreed that voriconazole should be used if any case of invasive aspergillosis comes across in their practice ( Table 2). Pertaining to the prophylactic use of antifungal agents referred to in question 11, 55.6% and 52.4% of clinicians adopted the approach in patients with recent perforated intra-abdominal surgery therapy and in those with high Candida scores and a high rate of ICU invasive candidiasis, respectively ( Table 2).

Participants' Responses to Pharmacokinetic Properties of Antifungal Therapy
Table S1 (Supplementary Materials) describes the assessment of knowledge and practice of HCPs regarding the pharmacokinetics of antifungal therapy. This section included eight questions. Pertaining to question 12, fortunately, it was found that almost four quarters (78%) of the clinicians were sure that fluconazole is the antifungal agent needed to be prescribed for a urosepsis patient with a positive Candida albicans urine culture. Similarly, for another scenario in question 14, where an ICU patient is at higher risk of IFI with a Candida score above three and a creatinine clearance (CrCl) of 30 mL/min, clinicians correctly believe that the dose of fluconazole needs to be renally adjusted (66.7%), as shown in Figure 4. Unfortunately, for question 15, a case of IC in a pregnant female, the majority of clinicians (63.5%) were uncertain which antifungal agent provides the safest option, and only 19% of them were likely to give amphotericin B (Figure 4). Furthermore, two-thirds (60%) of HCPs correctly knew that amphotericin B deoxycholate

Participants' Responses to Pharmacokinetic Properties of Antifungal Therapy
Table S1 (Supplementary Materials) describes the assessment of knowledge and practice of HCPs regarding the pharmacokinetics of antifungal therapy. This section included eight questions. Pertaining to question 12, fortunately, it was found that almost four quarters (78%) of the clinicians were sure that fluconazole is the antifungal agent needed to be prescribed for a urosepsis patient with a positive Candida albicans urine culture. Similarly, for another scenario in question 14, where an ICU patient is at higher risk of IFI with a Candida score above three and a creatinine clearance (CrCl) of 30 mL/min, clinicians correctly believe that the dose of fluconazole needs to be renally adjusted (66.7%), as shown  Figure 4. Unfortunately, for question 15, a case of IC in a pregnant female, the majority of clinicians (63.5%) were uncertain which antifungal agent provides the safest option, and only 19% of them were likely to give amphotericin B (Figure 4). Furthermore, two-thirds (60%) of HCPs correctly knew that amphotericin B deoxycholate is known for its inherent renal toxicity, and 65.5% believe that echinocandins do not require renal dose adjustments, as presented in questions 16 and 17, respectively ( Figure 4). In question 18, regarding the only azole that needs dose adjustment in hepatic impairment, only 20.6% of the clinicians were able to correctly identify voriconazole ( Figure 4); meanwhile, 36.5% were not sure about it. Furthermore, clinicians were not confident when prescribing the right fluconazole dose for the treatment of candidemia. Only 31.7% of them were sure that fluconazole 800 mg loading dose (LD) followed by 400 mg per day was the proper dose ( Figure 4). We also learned that 70% of clinicians do not rely on the oral route when administering fluconazole at their current ICU setting irrespective of the bioavailability, as shown in question 13 (Table S1). Although fluconazole has high bioavailability and is not affected by food, HCPs' responses can be explained by the septic status of the majority of ICU patients which dramatically affects gastrointestinal absorption.

Participants' Overall Knowledge and Practices Scores
The assessment of systemic antifungal prescribing knowledge and practice among physicians and pharmacists was given in Table 2. It can be observed that the prevalence of physicians who demonstrated correct answers by starting antifungal treatment immediately when managing ICU patients with confirmed candidemia was statistically significantly higher than pharmacists (p = 0.003) and using a prophylactic antifungal agent in patients who had undergone a recent perforated intra-abdominal surgery (p = 0.016) as reported in question 9 and 11, respectively. Vice versa, pharmacists showed better correct ratings regarding antifungal agents for a pregnant woman with IC (p = 0.001), and the azole that needs dose adjustment in hepatic impairment (p = 0.023), as seen in questions 15 and 18, respectively. Furthermore, the overall pharmacokinetics knowledge score was higher in pharmacists (p = 0.050) ( Figure 5). Considering the overall level of knowledge and practice, it was poor among 50.8% of HCPs, moderate in 46%, and only 3.2% had a good knowledge level. Interestingly, the overall knowledge scores of the ICU physicians and the clinical pharmacists (p = 0.925) were not significantly different and their antifungal therapy score was similar as well (p = 0.906) ( Figure 5).

Participants' Overall Knowledge and Practices Scores
The assessment of systemic antifungal prescribing knowledge and practice among physicians and pharmacists was given in Table 2. It can be observed that the prevalence of physicians who demonstrated correct answers by starting antifungal treatment immediately when managing ICU patients with confirmed candidemia was statistically significantly higher than pharmacists (p = 0.003) and using a prophylactic antifungal agent in patients who had undergone a recent perforated intra-abdominal surgery (p = 0.016) as reported in question 9 and 11, respectively. Vice versa, pharmacists showed better correct ratings regarding antifungal agents for a pregnant woman with IC (p = 0.001), and the azole that needs dose adjustment in hepatic impairment (p = 0.023), as seen in questions 15 and 18, respectively. Furthermore, the overall pharmacokinetics knowledge score was higher in pharmacists (p = 0.050) ( Figure 5). Considering the overall level of knowledge and practice, it was poor among 50.8% of HCPs, moderate in 46%, and only 3.2% had a good knowledge level. Interestingly, the overall knowledge scores of the ICU physicians and the clinical pharmacists (p = 0.925) were not significantly different and their antifungal therapy score was similar as well (p = 0.906) ( Figure 5).

Participants' Responses to Institution-Related Factors Affecting Antifungal Prescribing Practice
In Table 3, 50.8% of the HCPs indicated that their institution had antifungal sensitivity testing and 39.7% had surrogate antifungal tests such as the β-D-glucan or galactomannan test. The most prevalent Candida species, as reported by the institution, was Candida albicans (69.8%). Approximately 61.9% would de-escalate fluconazole based on the isolate sensitivity reported in culture. Nearly 60% of the HCPs reported that the average number of days for fungal culture results was 3 to 5 days. Only 27% of the HCPs use antifungal agents for the purpose of prophylaxis. The most preferred route for prescribing fluconazole in an ICU setting was the IV to oral route (60.3%).

Participants' Responses to Institution-Related Factors Affecting Antifungal Prescribing Practice
In Table 3, 50.8% of the HCPs indicated that their institution had antifungal sensitivity testing and 39.7% had surrogate antifungal tests such as the β-D-glucan or galactomannan test. The most prevalent Candida species, as reported by the institution, was Candida albicans (69.8%). Approximately 61.9% would de-escalate fluconazole based on the isolate sensitivity reported in culture. Nearly 60% of the HCPs reported that the average number of days for fungal culture results was 3 to 5 days. Only 27% of the HCPs use antifungal agents for the purpose of prophylaxis. The most preferred route for prescribing fluconazole in an ICU setting was the IV to oral route (60.3%).

Socio-Demographic Factors Affecting Participants' Knowledge and Practices Score
In Table 4, consultants were more associated with having better knowledge and practice scores (F = 3.131; p = 0.032). Likewise, the private sector and other hospitals were more associated with having better knowledge and practice scores (F = 2.883; p = 0.032). However, the difference in knowledge and practice according to years of experience did not reach statistical significance (p = 0.084). In Table 5, post-hoc analysis indicates that there was a significant difference in the knowledge and practice score between consultants and residents (p = 0.019); however, the comparison of knowledge and practice among other job descriptions did not reach statistical significance (p > 0.05).   Table 5. Post-hoc analysis of the multiple mean differences in the score of knowledge and practice according to the job description (N = 63) .

(I) Job Description (J) Job Description Mean Difference (I-J)
Sig.

Discussion
This study was conducted to determine the knowledge and practices of clinicians regarding antifungal therapy prescribing in an ICU setting. To our knowledge, this is the first study in Saudi Arabia that investigated clinicians' prescribing knowledge and practice toward antifungal therapy. Based on our findings, we confirmed that even regular prescribers in our area are in need of continuing education because, from this study, we observed a major gap in the knowledge and practices of clinicians in this domain.
Although most of our participants (75%) admitted to adding an empiric antifungal for patients who remain feverish despite broad-spectrum antibiotic therapy, only about one quarter (24%) were found to have started it initially in dialysis-dependent patients exhibiting signs of septic shock. About 33% of participants thought the positivity of a β-D-glucan test per se warrants the addition of antifungal therapy, and 22% thought that urinary colonization with Candida species is an indication to start antifungal agents even in a vitally stable patient. These practices contradict the published practice guidelines where empiric antifungal therapy is recommended in patients who have signs of septic shock and who are at risk of IC such as dialysis in our scenario [6]. It is worth mentioning that empiric therapy based solely on colonization with Candida species appears inadequate and false positivity remains a significant limitation of the β-D-glucan test, especially in critical illness [6].
On the other hand, clinicians demonstrated good knowledge and practices regarding initial antifungal treatment. Overall, 68.3% of them were confident that echinocandin (e.g., micafungin, caspofungin, anidulafungin) was the primary drug for the initiation of empiric antifungal treatment. Furthermore, 44.4% of them knew that the treatment should be continued for at least 2 weeks even if the patient is clinically stable. Likewise, three-quarters of the clinicians (75%) were aware that the antifungal treatment should stop if there was no clinical response to the therapy in 4-5 days. Conversely, in one study conducted by Aldrees et al [15] in an academic tertiary care center in Riyadh, Saudi Arabia to assess the appropriate utilization of caspofungin, more than half (56.3%) of the empiric treatment group continued beyond five days from initiation, even though there was no evidence of invasive Candida infections. The discrepancy can be explained by the fact that responding to patient scenarios in surveys differs from tracing real-world practices.
Regarding de-escalation practices, 57.1% of clinicians were sure that antifungal therapy should be altered from echinocandin to fluconazole if the patient was clinically stable and the blood culture grew isolates like fluconazole-sensitive Candida albicans; whereas, in the case of Candida glabrata, 54% of clinicians were aware that the prescription of echinocandin has to be continued until the patient improves. These findings are in accordance with clinical practice guidelines for the management of candidiasis [6]. In a retrospective study carried out in a 30-bed mixed ICU, located in the University Hospital of Lille, France, de-escalation was done in only 20% of the patients receiving empiric antifungal therapy for suspected Candida infection [25]. Though a higher percentage (57.1%) of our participants selected the correct approach of de-escalating echinocandin to fluconazole in Candida albicans susceptible strains, this percentage might have been exaggerated by the survey nature of our study; hence, it may be even lower in real-life practices. Furthermore, clinicians could not distinguish the right prescription for clinically stable patients with blood culture isolates such as voriconazole-sensitive Candida Krusei. Nearly 60% of them would continue with the echinocandin as long as the patient improved when, in fact, it needs to be de-escalated to voriconazole (3.4%) [6].
Comparing our results to the report of Valerio et al. [20], which was conducted in four European countries, we found some similarities; about 30% of physicians failed to distinguish Candida urinary infection from colonization, as did 22% in our study population. A low proportion of European physicians (41%) were able to identify the need for empiric antifungal added to antibiotics in patients with sepsis and femoral catheter-related infections. In a similar scenario of sepsis in dialysis-dependent patients in this study, we reported that only 24% of clinicians started antifungal therapy. Another survey-based study conducted in a 1550-bed tertiary care hospital in Spain identified serious gaps in the knowledge of prescribing physicians about important aspects of the diagnosis, prophylaxis, and treatment of IFIs [26].
Pertaining to invasive aspergillosis (IA), 38% of clinicians knew that voriconazole is a better antifungal agent than either liposomal amphoteric B (25.9%) or echinocandin (12.1%) for the management of IA. These results were concordant with Valerio et al.'s study [21], where 57% of the European physicians chose voriconazole as a first-line agent, as adopted by the international guidelines [27].
It is important to note that only about one-half of clinicians (51%) indicated that antifungal sensitivity tests were available in their current institution but 35% indicated that they were not available and 14% were not sure about the availability of susceptibility testing in their institutes. This was noted in a study conducted to assess the epidemiology and burden of IFI in the Arab League countries, where data concerning in vitro antifungal susceptibility testing were lacking in many of the documented available studies [28].
It can be further observed that Candida albicans is the most dominant Candida species reported by HCPs in our study (70%) and others were not dominantly reported including Candida tropicalis (8%) and Candida parapsilosis (4.8%). Literature from Saudi Arabia reported that Candida albicans is the most commonly detected Candida species [28][29][30]. However, in the study by Aldardeer et al. [4], Candida glabrata was the most commonly specified blood culture of Candida species followed by C. albicans, which is not consistent with our results.
The overall level of knowledge and practice was poor among 50.8% of our study HCPs, moderate in 46% and good in only 3.2%. This was concordant with a cross-sectional survey evaluating Nigerian resident doctors' knowledge and awareness about IFIs, where only two (0.002%) out of the 1046 respondents had a good level of awareness of IFIs [22]. Moreover, the study reported statistically significant differences in knowledge about IFIs among the various cadres of doctors as the level of knowledge increased with seniority, which was also confirmed in our research as consultants had better knowledge and practice scores [22].
The utilization of systemic antifungal agents has increased significantly in most tertiary centers. However, AFS has received very little attention [26]. Assessing knowledge and tracing practices of prescribing physicians and clinical pharmacists represents the first step essential for the development of an AFS program [26]. The application of an AFS program is associated with appropriate antifungal drug use, improved resource utilization and cost savings [18,19].

Limitations
Our study is somehow limited by the relatively small sample size of participating clinicians; however, this is partly explained by the selection of only ICU specialists in the five tertiary hospitals from which we gained approval in our area. Furthermore, access to ICU staff was greatly limited by the restrictions of the COVID-19 outbreak during which our study was conducted and even some ICU staff declined participation owing to workload. Our study was also conducted in one governorate in Saudi Arabia (Al-Ahsaa), which renders the generalization of results to other areas or other countries impractical.

Study Design and Setting
A cross-sectional, multi-center, survey-based study was conducted to evaluate ICU staff physicians' and clinical pharmacists' knowledge and practice of prescribing systemic antifungals in an ICU setting. This study was conducted among five tertiary hospitals located in Al-Ahsaa, Saudi Arabia between January and May 2021. The five hospitals were King Fahad Hospital Hofuf (KFHH), Almoosa Specialist Hospital, Al-Ahsaa Hospital, Prince Saud Bin Jalawy Hospital (PSBJ), and King Faisal Hospital.

Study Population and Data Collection
ICU staff physicians including consultants, specialists, and residents as well as ICU clinical pharmacists of the aforementioned hospitals were invited to complete a paperbased or electronic self-administered survey. The study investigators paid visits to the study site to meet and interview the clinicians in person. The objectives of the study were explained to the participating clinicians. Participation was voluntary and verbal consent was sought from all participants. Potential participants were selected by purposive sampling. The participating HCPs were allowed 15-20 min to complete the survey tool. Owing to the constraints imposed by the COVID-19 outbreak, especially in critical care areas, an electronic self-administered questionnaire was further provided to the ICU staff that the investigator did not have access to.

Questionnaire
The survey was designed by the authors in accordance with the IDSA 2016 updated Clinical Practice Guideline for the Management of Candidiasis [6]. Four of the main authors are academicians and three of them are hospital clinical preceptors in the area of critical care as well. Our survey was revised and validated by two infectious diseases and critical care professors as well as a pharmacology professor at King Faisal University. For the purpose of testing the validity, length, clarity, and comprehensibility of the survey, we piloted the questions among thirteen ICU staff physicians and two clinical pharmacists who provided their comments and amendments were subsequently made to the items of the survey. The internal consistency of the questionnaire was measured using Cronbach's alpha (a = 0.73), which was indicative of acceptable reliability.
The survey tool had 26 items; 19 of which assessed antifungal therapy knowledge and 7 of which explored institution-related practices of the participating clinicians. The survey was divided into four parts; part 1 covered some demographic characteristics of the participants. Part 2 of the survey assessed knowledge and practice of prescribing systemic antifungals in the ICU setting with regards to the decision to initiate and agent choice for empiric, prophylactic or definitive treatment in a given patient scenario. Part 3 of the tool assessed dosing and some pharmacokinetic parameters of the most commonly prescribed antifungal agents. The last part explored some institutional factors affecting antifungal therapy prescribing.

Sample Size and Ethical Approval
Our study was conducted in 5 tertiary centres in our area; the total number of ICU staff in these centres was 75. We employed the Raosoft calculator; the sample size was computed at a 5% margin of error, and a 95% CI. The calculated sample size was 63, and we were able to recruit this number. Ethical IRB approvals were obtained from King Fahad Hospital (KFHH) and Almoosa Specialist Hospital in Hofuf, Al-Ahsaa. The King Fahad Hospital IRB approval extended our privilege to collect data from other MOH hospitals in Al-Ahsaa.

Statistical Analysis
The prescribing knowledge and practice related to systemic antifungals was assessed using 19-item questionnaires, where the correct answer for each question had been identified and had been coded with 1 and the incorrect answer had been coded with 0. Knowledge questions 1 and 10 had multiple responses with 2 and 3 correct answers, respectively, giving a total knowledge questionnaire of 22 items. The total knowledge score was obtained by adding all 22 items and a possible score range from 0 to 22 had been generated. The actual score range based on participants' ratings has a range from 4 to 19 points.
Categorical variables were measured as frequency and proportion (%) and continuous variables were expressed as mean and standard deviation. The comparison of knowledge and practices according to the socio-demographic characteristics of HCPs was performed using an independent sample t-test or one-way ANOVA test. Post-hoc analysis has been carried out using the Tukey HSD test. A p-value of 0.05 was considered statistically significant. Normality testing was performed using the Shapiro-Wilk test as well as the Kolmogorov-Smirnov test. The data follows the normal distribution. Thus, the parametric tests were applied. All data analyses were performed using the Statistical Packages for Software Sciences (SPSS) version 26 (IBM Corporation, Armonk, New York, USA).

Conclusions
This study has revealed a significant gap in the knowledge and practice of clinicians toward prescribing antifungal therapy in our area. The overall level of knowledge and practice was poor among 50.8% of HCPs, and only 3.2% of HCPs had a good knowledge level in the present area of study. The overall knowledge and practice scores were almost identical across physicians and clinical pharmacists (p > 0.05), but the score of pharmacists for the pharmacokinetics domain was significantly higher compared to that of physicians (p = 0.05). Our findings, therefore, reflect the situation of antifungal prescription by clinicians in the region of study. These findings have exposed the need for a tailored training program essential for carrying out an antifungal stewardship program. Although our findings cannot be generalized, the knowledge gaps observed may partly explain the inappropriate use of antifungal agents, which consequently contributes to a global increase in antifungal resistance, adverse outcomes and increased costs. Compared to numerous antibiotic prescribing evaluation studies, there is a paucity of antifungal prescribing investigations making this research the first of its kind in the region. Hence, to fully evaluate this situation and utilize the applicability of our findings, further regional studies utilizing a larger population are needed in order to gain a better insight into the knowledge and practices of clinicians as regards prescribing antifungal therapy in our region.