Antibiotic prescribing among patients with severe infectious diseases in two private sector hospitals in Central India – a time series analysis over 10 years

Antibiotic resistance is an emerging global threat, as it causes significant morbidity and mortality worldwide [1, 2]. In an antibiotic surveillance report between 2016 and 2017 that focused on 22 countries, bacteria resistant to at least one of the most commonly used antibiotics were identified in patients diagnosed with bloodstream infections [3]. According to the World Health Organization (WHO), there is a need for global action against antibiotic resistance to ensure the effectiveness of antibiotic treatment in the future [4].

In low- and middle-income countries (LMICs), the infectious disease burden is often high, and antibiotic resistance is one of the common causes of mortality among patients with infectious diseases [2, 5]. Tackling antibiotic resistance requires costly equipment for microbiological analyses to determine the susceptibility of bacteria, and these methods must be implemented and reported [5, 6]. In LMICs, access to diagnostic methods is often limited. Consequently, antibiotics, frequently broad-spectrum antibiotics or fixed-dose combinations (FDCs), are commonly prescribed empirically based on a clinical suspicion of infection [7‐10]. However, antibiotic resistance is not confined to LMICs.

Improper use of antibiotics contributes to the development of antibiotic resistance. As reported previously, antibiotics should be prescribed only for relevant indications [11]. Furthermore, they should be as targeted as possible and administered at correct doses for appropriate treatment durations and by a suitable route of administration [11]. Mapping of antibiotic prescribing practices can identify areas for improvement. Long-term studies on antibiotic prescribing in LMICs, especially studies comparing practices among various health care providers and in various settings, are scarce [6].

The study population was inpatients admitted to the study hospitals, with severe infections for which empirical antibiotic treatment is recommended [12]. The hospitals selected for this study were representative of the majority of healthcare facilities in LMICs. The primary aim was to present and compare antibiotic prescribing practices over a 10-years period in two private-sector hospitals in Ujjain, Madhya Pradesh, India. A secondary aim was to compare antibiotic prescribing practices for selected indications with global recommendations for antibiotic treatment.

In total, 134,666 patients were admitted to the NTH, and 109,108 patients were admitted to the TH between 2008 and 2017. Data from 3766 patients were included in the analysis (NTH: n = 2504; TH: n = 1262, Table 2, Fig. 1). Overall, the patients from the TH were younger than the patients from the NTH (mean age NTH: 49.2, TH: 47.1 years, P < 0.01). At both hospitals, there were smaller proportions of women admitted, compared to men (percentage of admitted women NTH: 36%, TH: 24%) (Table 3). Antibiotics were commonly prescribed in both hospitals, although a significantly higher percentage of patients admitted to the NTH were prescribed antibiotics as compared with those admitted to the TH (89% at the TH, 92% at the NTH, P < 0.05) (Table 3). Among the different diagnostic groups, there were no differences in antibiotic prescribing practices between the hospitals, except for meningitis, where 90% of the patients in the NTH were prescribed antibiotics as compared with 70% in the TH (P < 0.05). The number of antibiotic prescriptions per patient in the TH was higher than that in the NTH. The average number of antibiotic prescriptions, i.e. prescription of one specified antibiotic with stated dose, frequency and duration in days, per patient was 22 in the TH and 8 in the NTH (Table 3). The duration of the hospital stay of the patients in the TH was higher than that of the patients in the NTH (mean 10.1 days at the TH and 4.4 days at the NTH, P < 0.05).

Prescribing of antibiotics for all indications increased between 2008 and 2017 in the NTH (P < 0.01), whereas prescribing practices did not change significantly during this period in the TH (P = 0.07, Fig. 2, Table 4). Antibiotics included in the access category comprised 61% of the total antibiotics prescribed in the TH and 40% of the total prescribed in the NTH (P < 0.01, Fig. 3, Table 4). Prescribing of access antibiotics increased in the NTH between 2008 and 2017. Prescribing of antibiotics categorized as watch antibiotics comprised 29% of the total antibiotics prescribed in the TH and 40% of the total prescribed in the NTH. Prescribing of watch antibiotics rose in both hospitals between 2008 and 2017 (P < 0.01 for both hospitals, Fig. 3, Table 4). Reserve antibiotics comprised less than 1 % antibiotics prescribed in both hospitals. However, prescribing of reserve antibiotics increased between 2008 and 2017 in the TH (P < 0.01, Fig. 3, Table 4).

The overall antibiotic prescribing among all included patients increased from 2008 to 2017 at the NTH (P < 0.01) but did not significantly change at the TH (P = 0.07, Fig. 1, Table 4). Antibiotics included in the access category comprised 61% of the total antibiotic prescribing at the TH and 40% at the NTH. Prescribing of access antibiotics increased at the NTH from 2008 to 2017 (P < 0.01, Fig. 3, Table 4). Prescribing of antibiotics categorized as watch antibiotics comprised 29% of the total antibiotic prescribing at the TH and 40% at the NTH. Prescribing of watch antibiotics increased at both hospitals from 2008 to 2017 (P < 0.01 for both hospitals, Fig. 3, Table 4). Reserve antibiotics comprised less than 1 % of the antibiotic prescribing at both hospitals, however prescribing of reserve antibiotics increased from 2008 to 2017 at the TH (P < 0.01, Fig. 3, Table 4). Prescribing of FDCs of antibiotics (J01R) comprised 8% of the antibiotic prescribing at the TH and 18% at the NTH. Prescribing of FDCs increased at both hospitals from 2008 to 2017 (P < 0.01 for both hospitals, Fig. 3, Table 4).

Patients diagnosed of cellulitis, peritonitis, pneumonia and sepsis accounted for 88% of the admissions to the NTH and 93% of the admissions in the TH. Table 5 shows the antibiotic groups (first-choice or second-choice) prescribed for each of these four diagnoses, which accounted for 75% of all antibiotics prescribed. Total antibiotic prescribing among patients with cellulitis, peritonitis and pneumonia increased between 2008 and 2017 in both the hospitals (P < 0.01 for both hospitals, Fig. 2, Table 4). Antibiotic prescribing among patients with sepsis increased between 2008 and 2017 in the NTH (P = 0.03) but decreased in the TH (P < 0.01, Fig. 2, Table 4). Sepsis was the only one of the four diagnoses where antibiotic consumption decreased at one of the hospitals.

According to our knowledge, this is the first study that compares antibiotic prescribing practices for selected infectious diagnoses over a 10 years period in two Indian private-sector hospitals. In the NTH, prescribing of antibiotics for all indications, including antibiotics specifically for cellulitis, pneumonia, peritonitis and sepsis diagnoses, increased from 2008 to 2017. In the TH, although antibiotic prescribing practices did not change during the study period, antibiotic prescribing for cellulitis, pneumonia and peritonitis increased and decreased for sepsis. Between 2008 and 2017, prescribing of access, watch and FDC of antibiotics rose in the NTH, and prescribing of watch, reserve and FDCs increased in the TH.

In this study, we analysed adherence to various international recommendations for antibiotic prescribing: the WHO’s recommendations for empirical antibiotic treatment for cellulitis and pneumonia, the recommendations for empirical antibiotic treatment from the ‘Surviving sepsis campaign’ for sepsis and the recommendations of the World Society of Emergency Surgery for peritonitis [15, 18, 19]. Adherence to these international guidelines increased in the TH compared to the NTH during the study period, especially prescribing of antibiotics for peritonitis and pneumonia.

Guidelines on empirical antibiotic treatment are often based on whether an infection is healthcare associated or community acquired. As both bacterial flora and susceptibility patterns vary worldwide, the following factors are important to consider while selecting the most appropriate antibiotic: the bacterium most likely to be the cause of the infection, patient’s clinical status, allergies to specific antibiotics and current or previous antibiotic resistance and responses to antibiotic treatment [1, 18, 19]. In this study, most of the patients did not have cultures sent for analysis, as the use of microbiological analyses was limited at both hospitals. The lack of microbiological analyses makes it difficult to comment on the rationale underlying the antibiotic prescription practices in the hospitals. However, prescribing of broad-spectrum antibiotics in both hospitals was high, which is in line with previous reports on prescribing of broad-spectrum antibiotics in Indian hospitals [12, 20]. As all the patients in the NTH paid for the treatment they received, the patients might have put pressure on their physicians to prescribe broad-spectrum antibiotics. In a qualitative study including 36 Indian doctors, Kotwani et al. reported that doctors faced demands from their patients to prescribe ‘strong’ antibiotics and that they sometimes prescribed antibiotics because they did not have time to debate with patients due to time constraints in busy health care facilities [21]. The aforementioned factors, as well as the desire to avoid re-consultation, might have contributed to the prescribing of broad-spectrum antibiotics in the hospitals in the present study.

In the NTH, the FDCs of antibiotics (J01R) were commonly prescribed to all patients, including those with cellulitis, peritonitis, pneumonia and sepsis. In contrast, the FDCs were less frequently prescribed in the TH. Prescribing of the FDCs for all indications increased in the TH during the study period but decreased for diagnoses of pneumonia and sepsis. Prescribing FDCs of antibiotics is not recommended, as they have been shown to drive antibiotic resistance, a common consequence of unnecessarily prescribed antibiotics, often in incorrect doses [22]. Appropriate prescribing of antibiotics requires that the dose be tailored for the individual patient, which is often not possible while prescribing the FDCs. A few FDCs that includes unapproved formulations, are known to be widely used in India [23‐26]. In March 2016, the Indian Government banned around 330 FDCs of drugs, of which 63 (19%) were FDCs of antibiotics [27]. However, there are still more than 118 FDCs of antibiotics available in the Indian market [27]. The presence of medical representatives and lack of local prescribing guidelines may have contributed to the higher prescribing of FDCs of antibiotics in the NTH as compared with that in the TH, where medical representatives are forbidden, and mainly generic medicines are procured by the management. Previous research demonstrated that pressure from pharmaceutical companies influence physicians’ prescribing practices in India [28].

Regardless of the country or setting, access antibiotics should primarily be used to save, whereas watch and reserve antibiotics should be used only for specific and limited indications in critically ill patients or patients with infections caused by bacteria with known antibiotic resistance [17]. In both hospitals, reserve antibiotics comprised less than 1% of the total antibiotics prescribed. In the TH, access antibiotics were most commonly prescribed (61% of antibiotics prescribed), followed by watch antibiotics (29%). However, in the NTH, access and watch antibiotics were prescribed in equal numbers (40% each). These results indicated that watch antibiotics accounted for a higher proportion of antibiotics prescribed in the NTH than in the TH. Furthermore, in the TH, prescribing of watch and reserve antibiotics for cellulitis or peritonitis increased in both hospitals during the study period, whereas prescribing of these antibiotics for sepsis decreased.

As noted earlier, the lack of microbiological analyses makes it difficult to comment on the rationale underlying the antibiotic prescription practices in the two hospitals. Nevertheless, the prescribing practices can be considered in terms of international recommendations on antibiotic aimed at reducing the emergence of antibiotic resistance. Based on the results of the present study, greater adherence to international guidelines for empirical antibiotic treatment is needed at both hospitals to optimize antibiotic use. The access, watch and reserve categorization of antibiotics provides a practical guide for proper antibiotic prescribing and can provide a basis for the development of local prescribing guidelines [17].

The relatively low incidence of some infectious diseases, such as infective endocarditis, among admissions to the NTH and the TH may be explained by underdiagnosing, which is a major problem in hospitals, as described in a previous study on infective endocarditis in India [29]. In many health care facilities in LMICs, microbiological tests and imaging methods are seldom used due to a lack of access to these diagnostic methods or a lack of time and money [8, 21]. In only a small number of cases in the present study, samples were sent for microbiological analyses, despite such analyses being readily available in both hospitals. Patient- and prescriber-related factors have been put forward to explain why culture tests are not routinely performed in the hospitals [12, 21, 30]. Patient-related factors include patients not being able to afford the tests or prefer to stay for short periods to pay less at the NTH. Prescriber-related factors might include doctors not having the time to wait for lab results due to overcrowding in the hospitals or an additional factor may be monetary driven, that doctors are paid for the number of patients they admit to the hospital so they might be wishing to see as many patients as possible in a given period [12, 21, 30]. Routine use of diagnostic methods, such as microbiological analysis and imaging methods, for patients with suspected infections might contribute to better management of and guidance on antibiotic treatment for infectious diseases to reduce antibiotic overuse.

Over the 10 years period, prescribing of all antibiotic categories, including access, watch and FDCs of antibiotics, increased in the NTH. In the TH, antibiotic prescribing practices did not change significantly, although prescribing of watch, reserve and FDCs of antibiotics increased between 2008 and 2017. Antibiotic prescribing for pneumonia, peritonitis and cellulitis increased in both hospitals between 2008 and 2017. In the TH, antibiotic prescribing for sepsis decreased but increased in the NTH. The results indicate that antibiotic prescribing practices need to be improved in both hospitals, although the TH generally prescribed more recommended antibiotics and fewer FDCs. Furthermore, the prescribing of recommended antibiotics improved over the study period in the TH. Factors contributing to extensive prescribing of watch antibiotics and FDCs in the NTH could be pressure from pharmaceutical companies and patients, as well as a lack of a local prescribing guidelines. The establishment of antibiotic stewardship programs based on the access, watch and reserve antibiotic categorization, as well as the implementation of locally adapted lists of essential medicines and prescribing guidelines could contribute to improved antibiotic prescribing practices and thus limit the development of antibiotic resistance. Furthermore, the implementation of routine diagnostic methods, such as microbiological analysis, could improve the management of infectious diseases and guide antibiotic therapy decisions for appropriate prescribing of antibiotics. The hospitals included in the present study are typical of those in similar settings in LMICs, and our results are in line with those of previously published studies in this area. The results highlight areas for improvement in antibiotic prescribing practices and same could be anticipated from other similar settings.