The World Health Organization, in its document on global strategy for containment of antimicrobial resistance, noted inappropriate prescribing and use of antibiotics as factors that both compromise treatment outcomes and contribute to the development of pathogen resistance to antibacterial agents.1 It further stressed the need for research towards filling gaps in existing knowledge in antibiotic prescribing.
Although prescribers are often assumed to be prescribing antibiotics appropriately, making clinical decisions to initiate antibiotic therapy and selecting antibiotics appropriately in cases of established infections can be challenging. In principle, this calls for the prescriber to establish the presence of an infection to justify his or her decision to prescribe an antibiotic. In the cases of persons in long-term-care facilities, for example, it can be difficult to establish a diagnosis of infection in some patients.2 In other cases, diagnosis of bacterial infections may be difficult because of the resemblance of clinical symptoms classically identified with certain bacterial infections with symptoms of other types of infections or disease conditions that do not have pathogenic bacteria as etiologies. Classic examples are the similarities of symptoms of viral and bacterial infections of the respiratory tract.3 Descriptively, symptoms of protozoal infections of the vagina manifesting as discharges similarly bear resemblance with symptoms of bacterial infections caused by Neisseria gonorrhea or Chlamydia trachomatis.4 Even in cases where bacteria pathogens may be etiologies of a diagnosed infection, an antibiotic prescribed for such an infection may be considered appropriate only when it targets, in terms of its activity, the exact pathogen or pathogens causing that infection.5 In this respect, and for it to be judged appropriately prescribed, factors that determine the therapeutic efficacy of the antibiotic would have to be considered also. For example, these include the antibiotic’s spectrum of activity, its physicochemical, pharmacokinetic and pharmacodynamic properties and its compatibility with other antibiotics in situations of multiple antibiotic prescribing.6,7 Taking these into account, prescribing antibiotics appropriately can be expected to be a challenging duty and designing procedures for assessing the appropriateness of antibiotic prescriptions a difficult and complicated task.
Many studies that investigate appropriateness of antibiotic prescriptions audit antibiotic prescribing by determining extents to which antibiotics are prescribed in conformity to treatment guidelines or some criteria of appropriateness based on opinions of antibiotic experts. In clinical environments where neither elaborate antibiotic treatment guidelines nor services of antibiotic experts are available, it may become necessary to assess the appropriateness of antibiotic prescriptions using alternative but equally versatile procedures. This article presents a methodology of assessing antibiotic prescriptions using a computer assisted procedure in determining the appropriateness of these prescriptions based on their conformities to sets of criteria formulated from basic principles of antibiotic prescribing. It also provides details of the methodology we used to develop an antibiotic prescription classification system as reported in our earlier published work on the impact of appropriate antibiotic prescribing on treatment evaluation parameters.8 Our methodology also demonstrated an added advantage of providing data that could be analyzed to make information available on sources of inappropriate prescribing in respect to specified criteria. We consider such information useful in formulating antibiotic prescribing policies aimed at promoting appropriate prescribing.
Materials and Methods
The primary objective of this article is to present details of procedures used to develop an antibiotic prescription classification system that avoids dependence on expert opinions or treatment guidelines in determining appropriateness of antibiotic prescriptions.
Relevant data on procedures of infection diagnosis and prescribed antibiotics were collected from patients’ case notes within a one-month period in five public hospitals in Lesotho. The prescriptions were assessed for their appropriateness on the basis of their conformities to sets of criteria developed from principles of antibiotic prescribing as detailed below. In all, 307 inpatient and 865 outpatient prescription records were assessed. Ethical permissions in accordance with the Helsinki Declaration were received from both the Ministry of Health of Lesotho through its ethics committee for public hospitals and individual Christian Health Association of Lesotho (CHAL) hospitals, as well as the ethics committee of the North-West University (South Africa).
Statements on guiding principles of antibiotic prescribing were developed into criteria (Table 1) and formulated into a prescription assessment tool. Responses of Yes, No or Not applicable to questions in the assessment tool were assessed. A table indicating the diagnoses, symptoms and etiologies of infections as reviewed from the literature was compiled. This was used as a reference source in the data compilation procedures to make decisions on whether or not diagnosed infections for which antibiotics were prescribed were of bacterial causes.
Using Statistical Analysis Systems® (SAS) for Windows 9.1®, prescriptions’ conformities to developed criteria were combined into conditions. The conditions defined both the extent to which the presence of infections were established prior to antibiotic prescribing and the extent to which antibiotics were prescribed appropriately (Tables 2 and 3). The conditions were combined further into seven categories of prescription appropriateness (Tables 4 and 5). The use of the procedures enabled the categorization of all prescriptions studied into the different categories of appropriateness.
All 307 inpatient and 865 outpatient antibiotic prescriptions were categorized into the seven predefined categories of prescription appropriateness. A total of 28.2% of inpatient and 78.4% of outpatient prescriptions were written appropriately for the treatment of infections (categories A1 and A2). This is in comparison with 47.9% of inpatient and 18.7% of outpatient prescriptions that were prescribed inappropriately for the treatment of infections or prescribed inappropriately for diagnosed clinical conditions for which use of antibiotics were considered not justified (categories B and F) (Table 6).
Methodologies of assessing appropriateness of antibiotic prescriptions
In studies on the assessment of appropriateness of antibiotic prescriptions, instruments were based most commonly on judgments of panels of antibiotic experts,9-12 or on prescribing recommendations in infection treatment guidelines.13-15 Assessment procedures based on judgments of experts were mostly based on direct judgments on the appropriateness of prescriptions at panel sittings where prescriptions were examined.9,10 In some cases such judgments were based on prescription conformities to criteria previously developed by an expert or panel of experts which in themselves may be based on provisions of recommendations in national guidelines.11,12
Though undeniably a popular method of assessing antibiotic prescriptions, the use of panels of experts in assessing antibiotic prescriptions is not without shortcomings. Typically, difficulties may be encountered in obtaining agreements and consistencies in scorings of such panels.10 Engaging the services of antibiotic experts may also increase research costs and limit the use of the methodology in resource-limited countries. It may also be inapplicable in clinical environments where these experts are not available. Under these circumstances, an alternative procedure could be to base prescription assessment on the use of instruments developed from recommendations in antibiotic treatment guidelines. In clinical environments where elaborate antibiotic treatment guidelines are not available, the development of such instruments may be limited. These situations exist in Lesotho, and may also be prevailing in a number of resource-limited countries. They make alternative methods of assessing the appropriateness of antibiotic prescriptions worthwhile.
The methodology of assessing the appropriateness of antibiotic prescriptions in this study employed an instrument developed from principles of antibiotic prescribing as documented in literature. These included essentially the principles of: i) establishing presence and sites of infections prior to antibiotic prescribing;7,16,17 ii) establishing potential sources of infection or comorbid conditions predisposing patients to certain infections prior to prescribing antibiotics for prophylactic reasons;7,18,19 iii) making appropriate antibiotic selections for empiric or definitive treatment of infections including establishing the need for and ensuring the effectiveness of prescribed antibiotics;6,7,16,17,20,21 and iv) appropriately selecting and initiating antibiotic therapy in clinically ill or hospitalized patients.16,17,22 The method generally assessed antibiotic prescriptions against the establishment of the need for antibiotic use in circumstances for which the drugs were prescribed. It also assessed the appropriate prescribing of the agents with respect to their doses, compatibility with co-prescribed antibiotics as well as their effectiveness in treating infections for which they have been prescribed. By avoiding the use of either the services of antibiotic experts or criteria compiled solely from elaborate antibiotic treatment guidelines, our method provides a novel approach to assessing antibiotic prescriptions in clinical environments that lack these amenities.
Prescription assessment results
Antibiotic prescribing in outpatient settings appeared to be more appropriate than in inpatient settings. However, 55.8% (378 of 677) were cases for which the agents were prescribed for infections with suspected bacterial etiologies (category A2). To be adjudged appropriately prescribed, antibiotics need to be both prescribed first and foremost for infections with bacterial etiologies and also selected to provide antimicrobial coverage for the diagnosed infection.1,23 Categorizing antibiotic prescriptions written for suspected bacterial infections points to a problem area of antibiotic prescribing that needs redress. We consider this an advantage of the methodology. Based on the types of infections encountered most at the setting, prescribers could establish the presence of infections from sites and symptoms of infections to justify their decisions to prescribe antibiotics. For such infections, they may lack the means or the expertise to differentially diagnose these infections and confirm bacterial pathogens as their etiologies before prescribing antibiotics. The prescription of antibiotics for infections of upper respiratory tract infections (URTIs) is a standard case of reference in such instances. These infections as seen in outpatient settings have been shown to be mainly of viral etiologies with less than 2% being of bacterial causes.24 In the absence of differential diagnosis, a majority of antibiotic prescriptions classified as category A2 may actually be given for the treatment of URTIs or other types of infections which may not have bacterial causes.
A total of 30.0% of inpatient prescriptions were inappropriately prescribed for the empiric treatment of infections (Category B) and 17.9% were prescribed for clinical conditions for which use of antibacterial agents were deemed not justified (Category F). The results highlight inabilities of prescribers to select appropriate antibiotic treatments for diagnosed infections as the more prominent of the causes of inappropriate prescribing of antibiotics in inpatient settings of study site hospitals. The reverse was true for outpatient departments where most inappropriately prescribed antibiotics were indicated for clinical conditions not requiring antibiotic treatments.
Antibiotic prescribing for infection prophylaxis was done mostly in inpatient settings. This was expected because such modes of antibiotic usage are encountered more in surgical wards where the agents are commonly prescribed for preventing post-surgical wound infections than in medical wards where the agents are most used for treating rather than preventing infections. An almost equal percentage of prescriptions were written appropriately (category D) and inappropriately (category E) for purposes of preventing infections in inpatient settings. This result may be a problem of inappropriate prescribing of the antibiotics for prophylactic reasons in the inpatient setting. The 1.3% of inpatient prescriptions with antibiotics prescribed according to results of culture sensitivity tests and with none such prescriptions identified from outpatient departments, documents essentially empiric prescribing of antibiotics as a mainstay of treating infections at study site hospitals. This is a significant finding as it accentuates the importance of prescribers’ appropriate selection and prescribing of the agents to ensure the effective treatment of infections.
The greatest challenge of the use of the method is the ability of a researcher to correctly decide whether or not a prescription being assessed conforms to set criteria against which it is evaluated. Inability to correctly decide on the conformity of a prescription to such set criteria compromises results. To address this challenge in the assessment procedure, informed decisions on the conformity of prescriptions to the assessment criteria were made by using data collection tools that served as reference source on infections and their causative agents as well as the therapeutic and physico-chemical properties of antibiotics.
A methodology employing an antibiotic prescription assessment instrument based on criteria formulated from principles of antibiotic prescribing has been developed and used successfully in assessing appropriateness of antibiotic prescriptions. It capably classified each prescription into either of two major categories of appropriately and inappropriately prescribed antibiotics and also according to respective purposes and/or reasons for which antibiotics may be prescribed. Considered an advantage of the method, this enabled areas of antibiotic usage with associated problems and the reasons for such problems to be to be identified. In Lesotho and in line with the objectives of this study, employment of the methodology established high rates of inappropriate prescribing of antibiotics in the empiric treatment or prophylaxis of infections in inpatient settings. In contrast to results obtained for inpatient settings, high rates of appropriate prescribing of antibiotics according to principles appeared to prevail in outpatient settings. The majority of such prescriptions, however, were identified as prescriptions of antibiotics prescribed for infections with unconfirmed bacterial etiologies.