Factors influencing decision making and antibiotic prescribing patterns for the treatment of carbapenem-resistant Enterobacteriaceae (CRE) among non-infectious physicians in Thailand: a qualitative study

Carbapenem-resistant Enterobacteriaceae (CRE) infection is associated with elevated mortality rates, which are higher than the odds in carbapenem-susceptible Enterobacteriaceae infected patients [1]. Currently, antibiotic options for CRE treatment are very limited. Polymyxin, fosfomycin, tigecycline, and aminoglycosides are the mainstay treatments [2, 3, 4]. In addition, several therapeutic strategies, e.g., using high-dose of antibiotics [3], double carbapenems [5], and combined antibiotics [4, 5] are used to treat severe CRE infections to maximize treatment success. Although there are CRE treatment guidelines by the Infectious Diseases Society of America Guidance in 2020 [6], the resistance of CRE to antibiotics in each area is different due to the variety of common carbapenemase genes [3]. Therefore, the treatment of CRE infections was diverse locally. The superiority of the efficacy of each treatment option is not very well defined [7], and the treatment of CRE is highly dependent on the clinical experience of the physicians [8].

Because of the difficulty in treating CRE infection, inappropriate CRE treatment is common [9, 10, 11]. The prescription of inappropriate empirical treatment can lead to poor overall treatment outcomes [12] and high mortality [9]. Also, inappropriate specific antibiotic treatment, i.e., the delay of coverage antibiotics and the use of non-coverage antibiotics, is associated with poor clinical outcomes for CRE treatment [13]. The complexity of CRE treatment further depends on several other factors, e.g., physician knowledge, attitude, and perception of CRE [14, 15], engagement of healthcare providers [16], and intervention efforts in medical decision-making [17]. Since appropriate antibiotic use is important for CRE treatment and physician decisions on how CRE infections are treated, empirically and specifically, are critical for the success in the management of CRE-infected patients. We aimed to find physicians’ attitudes toward CRE infection situations and investigate how noninfectious disease (non-ID) physicians select antibiotic treatment patterns for CRE infection, to better understand factors influencing decision-making and antibiotic prescribing for CRE.

Of the 36 residents or physicians-in-training invited, 25 (69%) participated (Table 1). The interviews revealed 13 themes related to factors influencing decision-making and antibiotic prescribing patterns for CRE (Table 2).

Table 1
Characteristics of 25 non-infectious physician participants

• Click for larger image
• Click for full table
• Download as Excel file

Table 2
Themes and illustrative quotations identified from semi-structured interviews with 25 inpatients physicians

• Click for larger image
• Click for full table
• Download as Excel file

Theme 6. Approaches for the empiric treatment of CRE

Most physicians did not think of CRE as a causative organism until the culture result was reported. However, when the empiric was essential, regimens often recommended by experienced ID specialists included carbapenem, colistin plus fosfomycin, colistin-based, and aminoglycoside-based combinations (Table 3). Factors that affected the selection of antibiotics included the site of infection and the history of antibiotics used for the treatment of the previous CRE infection. Some physicians always wait for recommendations from ID specialists.

Table 3
Empirical treatment regimens for CRE infections

• Click for larger image
• Click for full table
• Download as Excel file

In this study, participating physicians believed that appropriate antibiotic use was important for the treatment of CRE patients, agreeing with the study by Zilberberg et al. [13]. Also, physicians agreed that selecting antibiotics for the treatment of CRE infection was difficult because of multiple antimicrobial use and limited options, agreeing with other studies [2, 7]. Several factors that were also reported by Weston et al. [20] and the CDC [21] were used to estimate the likelihood of CRE infection, including a history of positive CRE culture, intubation, long hospitalization, exposure to carbapenems, patient transferred from an area known to have CRE outbreaks, frequent admission to the hospital or frequent exposure to antibiotics. We also found that recommendations from ID specialists heavily influenced the selection of empirical CRE treatment by physicians. Specific treatments found in this study agreed with Daikos et al. [22] and included combination therapies that had colistin as a core with fosfomycin and aminoglycoside as supplements or the use of at least 2 drugs that the bacteria were susceptible to. However, while Daikos et al. [22] found the use of double carbapenem therapy in their study, it was not reported in ours. This might be because using double carbapenem is not practical when the MIC is more than 8 mg/L [23].

The selection of empirical therapy for CRE Infection was influenced by the specialty of the physicians. Most internal medicine physicians were able to prescribe consistently with the antimicrobial treatment guidelines [6, 24, 25]. On the contrary, surgeons, and orthopedists reported that they usually waited for recommendations from ID specialists. The physician attributes such as their specialty, clinical experience, and practice decisions may have influenced the physicians’ decision-making regarding CRE infection. This issue is consistent with other studies [8, 26]. Other factors involved in the selection of antibiotics for specific treatments that were also reported in the literature were the culture reports, site of infection, MIC, pharmacokinetics and pharmacodynamics, disease condition, and clinical manifestations of the patient [7, 23]. In addition, antibiogram and antibiotic control policies in the hospitals may have influenced the determination of a prescribing pattern, agreeing with the current operational guidelines of Thailand [27].

The limitation of this research is that the interviews did not include all types of specialists available in the hospital. We believe that this problem did not affect the quality of this study because the inpatient care system operating in this hospital has non-ID specialists as the primary medical care providers. ID specialists with fellowships usually provide supervision for the residents or care for more critical patients. Next, the first author is a registered pharmacist and member of the committee for antimicrobial-resistant management both at the hospital level and health district level. Therefore, she is already familiar with policies, systems, and personnel involved with antimicrobial resistance in the hospital. In this sense, she has already submerged in the environment and her experiences on the topics can be used to triangulate with the findings from this study. Her involvement also allowed personal reflexivity to be conducted. However, this can lead to a lack of new perspectives and selective reporting bias. The second author who had limited experience on the topic acted as a fresh eye and help with thematic analysis to reduce this limitation. We also did group discussions to minimize the biases and maintain the phenomenology approach in this study. Another limitation is that antibiotics and the pattern of antibiotics used for the treatment of CRE infection in this research were influenced heavily by the National Drug List and the experience of physicians working in the hospital. Hence, other drugs that may have been used in Thailand were not mentioned. This limitation should not significantly affect the data since the treatment of CRE in Thailand was in general homogenous [25, 26, 28].

In conclusion, physicians realized that CRE resistance was an important problem. However, physicians still did not comply with the CRE treatment guidelines because of the lack of confidence and knowledge in treating CRE infection; insufficient necessary information; and unstandardized ICP practice. Providing accessible necessary information, enough ID specialists, and DUE were reported as solutions for the CRE challenges.