Objectives To audit the quality of treatment of lower respiratory tract infections (LRTIs) and urinary tract infections (UTIs) and to identify targets for antibiotic stewardship.
Methods The audit involved collecting data on admitted patients, who were diagnosed with LRTIs or UTIs and subsequently received antibiotic treatment (January 2009–April 2009).
Key findings The percentage adherence rate for hospital antibiotic policy was 68.6% (24/35). Documentation of the CURB-65 score was found in 80% (16/20) of the patients’ clinical notes, for which 46.2% (6/13) of patients were treated according to their CURB-65 score. The percentages of delayed and missed doses for all antibiotics were 21.7% (254/1171) and 8.6% (101/1171), respectively. The percentage of patients switched from intravenous to oral antibiotics in accordance with the policy was 58.5% (31/53). The mean length of stay for patients switched in line with the guidelines was 6.9 days (range: 2–18 days) compared with 13.2 days (range: 4–28 days) for patients treated with intravenous antibiotics >24 h after the intravenous to oral switch criteria were fulfilled; this equates to on average an extra 6.3 days of hospitalisation (p=0.01).
Conclusions The study identified a number of targets for quality improvement including adherence to antibiotic policy, documentation of the CURB-65 score in patients’ notes and treating patients accordingly, addressing the issue of missed and delayed doses, and maintaining adherence to the hospital intravenous-to-oral antibiotic switch policy. The findings suggest that the quality of antibiotic prescribing could be improved by measuring and addressing such performance indicators.
- Clinical Pharmacy
- Infectious Diseases
Statistics from Altmetric.com
The continued emergence of antibiotic resistant pathogens poses major consequences for the individual and healthcare institutions including increased patient mortality, longer hospital stays and increased healthcare costs.1 ,2 It has been estimated that hospital-acquired infections cost the UK National Health Service £1 billion per year.3 As antimicrobial use is considered a major determinant in the evolution of resistance,4 imprudent antimicrobial use has significant implications for the selection and spread of drug resistant micro-organisms.5–7 Recent studies have shown that almost 50% antibiotics are believed to be inappropriately prescribed, which can lead to increased selection of resistant organisms, which in turn impacts upon patient morbidity and mortality.8 Antibiotic stewardship aims to improve patient care while reducing adverse effects associated with antibiotic overuse or misuse, including antimicrobial resistance, development of secondary infection, adverse drug reactions, increased length of hospital stay and additional healthcare costs.8–10 As evidence suggests that good antimicrobial stewardship can lead to less overall and inappropriate antimicrobial use, lower drug related costs and less emergence of antimicrobial resistance,8–11 ensuring full compliance with such measures is deemed necessary.
Several studies have shown the benefits of implementing an antibiotic intravenous-to-oral switch policy, while giving the same cure rate as a full intravenous course, including decreasing the length of hospital stay, providing economic savings, and decreasing medical and pharmacy time.12–14 Infections described as suitable for intravenous-to-oral switch include the following: community-acquired pneumonia (CAP) and urinary tract infections (UTIs).
Policies on the use of antibiotics (empirical treatment and surgical prophylaxis) have been in place in the audit site since 1995. A user guide to intravenous antibiotics was introduced in 1999 and all antibiotic policies were regularly reviewed and updated, and were made available via the intranet and by placing a hard copy on all wards.
The objective of this audit was to assess the quality of treatment of lower respiratory tract infections (LRTIs) and UTIs in Antrim Hospital and to identify targets for quality improvement, in relation to; adherence to the hospital antibiotic policy, documentation of the CURB-65 score in the medical notes of patients diagnosed with CAP, clinical cure on first-line treatment, the number of delayed and missed doses received by patients, and appropriateness of intravenous-to-oral switch, with its subsequent cost savings.
The audit was registered with The Trust Governance Department as a clinical audit.
Setting and audit period
The audit was carried out in Antrim Area Hospital in Northern Ireland, UK, a 426-bed district general teaching hospital. The present prospective investigation involved collecting data on admitted patients, who were diagnosed with LRTIs and/or UTIs and subsequently received antibiotic treatment, over a 3-month period (January–April 2009). The latter study period was accepted as an appropriate length of time to capture infections, especially given that the months included covered winter months, and to provide a convenient sample size. Data was collected from each patient's prescription record, medical notes and observation chart.
The evaluation of adherence to the hospital antibiotic policy
For the purposes of this study, the following definitions were used: (1) the term LRTI is used for all LRTIs (including pneumonia); (2) patients with no signs of pneumonia, whereby no consolidation was observed on the chest X-ray, are classified as having a non-pneumonic LRTI (LRTI (n.s.p)); (3) patients with infective exacerbations of chronic obstructive pulmonary disease (COPD) are classified as LRTI (inf. ex. of COPD); (4) where LRTI diagnosis was uncertain, this was defined as an LRTI (undefined); and (5) patients diagnosed with pneumonia less than 48 h after admission to hospital were classified as CAP, and patients who were diagnosed with pneumonia greater than 48 h after hospital admission were diagnosed as hospital acquired pneumonia. The antibiotic treatment was compared with the policy with respect to type of antibiotic, regime, dose and frequency. The antibiotic policy in use during the study period is outlined in table 1. Adherence is defined as prescribing of an antibiotic at the dose, route and frequency that is recommended in the antibiotic policy. Non-adherence is defined as the prescribing of a different antibiotic to that recommended in the policy or prescribing of the recommended antibiotic at the incorrect dose, route or frequency, without any justifiable reason. Patients who received therapy off policy for appropriate reasons, for example, as per sensitivities, intolerance of recommended therapy, were not included in this part of the audit. The severity assessment of CAP was assessed using the British Thoracic Society CURB-65 assessment tool which is outlined in box 1. 15 This is a risk stratification tool for patients with CAP, whereby the mortality associated with CAP increases with the CURB-65 score. Patients with a CURB-65 score of 0 to 1 have a less than 3% mortality rate, compared with patients with a score of 3 to 5, who have a mortality rate of 15–40%. 16 Patients with CAP with no CURB-65 score could not be assessed on adherence and were therefore considered as indeterminate. Where there was ambiguity regarding reasons for non-adherence to the antibiotic policy, the ward pharmacist was consulted for clarification.
Criteria for calculating CURB-65 Score
The patient scores one point for each of the following criteria met
Urea >7 mmol/L,
Respiratory rate >30/min
Blood pressure <90 systolic or ≤60 diastolic
Age ≥65 years
The proportion of inadvertently missed and delayed doses was recorded for all intravenous treatments. Intentional delays in administering an antibiotic, and delayed doses on the 1st day of treatment were not included.
Empirical treatment of LRTI (n.s.p) and LRTI/UTI
Several patients included in this audit were diagnosed with LRTI (n.s.p) or LRTI/UTI. There were no recommendations in the antibiotic policy for the treatment of these infections; therefore adherence to the policy could not be assessed. However, antibiotics used for the empirical treatment of these infections were recorded in order to determine current practice in treating such infections. Patients who were on antibiotics prior to admission and patients for whom sensitivities were available were excluded as empirical therapy is not applicable in these circumstances.
Clinical cure on first-line treatment
The proportion of patients who achieved clinical cure on first-line treatment was defined as the proportion of patients treated successfully without needing additions or alterations to their antibiotic therapy, other than step-down therapy. Patients stepped up from oral to intravenous antibiotics were classified as an indeterminate group.
Appropriateness of intravenous-to-oral switch
Patients treated with intravenous antibiotic therapy were assessed if they were switched from intravenous-to-oral treatment in accordance with the hospital intravenous-to-oral switch policy. The criteria for intravenous-to-oral switch are outlined in box 2. All patients were reviewed after 72 h (3 days) of intravenous treatment and daily thereafter during the entire intravenous antibiotic course. Intravenous-to-oral switch was considered appropriate if performed within 24 h after all of the criteria were fulfilled. Patients that were switched less than 72 h after the start of intravenous treatment were considered appropriately switched if all criteria had been fulfilled for 24 h. The patients were categorised into one of the following groups. (1) Switched in accordance with the intravenous-to-oral switch policy. (2) Not switched in accordance with the intravenous-to-oral switch policy, that is, intravenous treatment continued for more than 24 h after the criteria were fulfilled, however, the patient was eventually switched to oral antibiotics; or intravenous treatment continued more than 24 h after the criteria were fulfilled and the patient was never switched to oral antibiotics; (3) Not applicable, that is, discontinuation of antibiotic treatment less than 24 h after the criteria were fulfilled; or antibiotic treatment discontinued or switched to oral before criteria were fulfilled; or patients who were not suitable for switch.
intravenous to oral switch criteria
Criteria for intravenous-to-oral switch
LRTI: Mean pulse rate<100 beats/min
UTI: Mean pulse rate<90 beats/min
Mean respiratory rate <20 breaths/min
Resolution or improvement of hypotension
Absence of hypoxia
White cell count 4–12 × 109
Ability to take oral medication
Availability of suitable oral formulations
LRTI, lower respiratory tract infection; UTI, urinary tract infections.
Statistics and economic evaluation
Descriptive statistics and frequency analyses were undertaken. Due to the small sample size in this study, Fisher's exact test and non-parametric tests (ie, Mann-Whitney U test) were used to assess the effect of adherence to the antibiotic policy on clinical cure rate and the effect of appropriate intravenous-to-oral switch on length of stay. All analyses were performed using SPSS (V.15) for Windows and a p<0.05 was considered statistically significant. In order to assess the economic impact of the intravenous-to-oral switch, the price of the inappropriate intravenous treatment was compared with the price of the equivalent step-down treatment. The prices were obtained from the hospital pharmacy information system and all cost calculations were based on the price of the antibiotic as at April 2009. The cost was then extrapolated for a 1 year period. The latter annual cost was calculated based on the assumption that the incidence of LRTI and UTI cases was consistent throughout the year.
Over the 3-month study period, 135 patients were included, 22 of these patients were excluded due to death during antibiotic treatment, transfer to another hospital during antibiotic treatment, missing data and uncertainty of diagnosis. As such, data was obtained for the 113 eligible patients, 55.8% (63/113) female and 44.2% (50/113) male, who were all treated with antibiotics for LRTIs (65.5%) and/or UTI (34.5%). Details of the frequency of diagnosis are presented in table 2. The median age was 75 years (IQR: 26 years). Two patients were less than 18 years, 29 patients were aged between 19 years and 64 years and 72 patients are greater than or equal to 65 years. Twenty-three per cent (26/113) of the patients were admitted from nursing homes or other healthcare institutions. Out of the 113 patients studied, 35 (31%) were eligible for assessment in terms of their adherence to the hospital antibiotic policy. The percentage adherence rate to the hospital antibiotic policy was 68.6% (24/35); 31.4% (11/35) of patients were not treated in accordance with the hospital policy. Different trends in adherence rates were observed for each diagnosis (table 2). Assessment of the documentation of the CURB-65 score was carried out in 20 of the 21 patients diagnosed with CAP as one patient was excluded due to an indeterminate CURB-65 score. Documentation of the CURB-65 score was found in 80% (16/20) of the cases. Twenty-five per cent (4/16) of patients had a CURB-65 score of 0 to 1, 25% (4/16) patients had a score of 2 and 8, 50% (8/16) patients had a score of 3 to 5. The percentage of patients treated in line with the CURB-65 score was 46.2% (6/13); three patients were excluded since they were being treated with antibiotics at admission.
Clinical cure on first-line treatment was found in 64.6% (73/113) of all patients, 28.3% (32/113) were assessed as not clinically cured, 5.3% (6/113) were stepped up from oral to intravenous antibiotics (no other change or addition in antibiotic treatment) and 1.8% (2/113) were not assessed as they died during the study. Data about clinical cure for each diagnosis is presented in table 2. Clinical cure on first-line treatment was found in 70.8% (17/24) of patients treated in line with the antimicrobial policy and 72.7% (8/11) of patients not treated in line with the antimicrobial policy. The use of Fisher's exact test showed that there was no significant difference in cure rate between the two groups (p=0.99).
Data was collected for 173 courses of intravenous antibiotic treatment (107 patients) in order to determine the proportion of missed and delayed doses. Forty-eight courses were excluded due to inadequately recorded data on antibiotic administration (28 courses in 23 patients) and once-only prescriptions (20 doses in 7 patients). The percentages of inadvertently delayed and missed doses for all antibiotics were 21.7% (254/1171) and 8.6% (101/1171), respectively. Trends in antibiotic prescribing, and missed and delayed doses are presented in table 3.
Of the 21 patients who were classified as LRTI (n.s.p.), 7 patients were excluded since they were being treated with antibiotics at admission (n=6) or had a sensitivity result on the 1st day of antibiotic treatment (n=1), leaving a total number of 14 patients. The most commonly used antibiotic was co-amoxiclav, which was prescribed in 50% (7/14) of the LRTI (n.s.p.) patients, followed by the combination of co-amoxiclav and clarithromycin (14.3%; 2/14), and amoxicillin (14.3%; 2/14). Of the 17 patients included with a diagnosis of UTI/LRTI or with both, 9 patients were excluded due to antibiotics at admission or sensitivity results on the 1st day of antibiotic treatment. In the remaining 8 patients the most commonly prescribed antibiotic was co-amoxiclav which was the first-line treatment for 50% (4/8) of all patients treated empirically for LRTI/UTI, followed by a combination of co-amoxiclav and clarithromycin (25%; 2/8).
Appropriateness of intravenous-to-oral switch and additional costs:
The switch from intravenous-to-oral antibiotics was evaluated in 53 patients. The percentage of patients switched in accordance with the policy was 58.5% (31/53) and the percentage of patients not switched in line with the policy was 41.5% (22/53). Trends in adherence to the intravenous-to-oral switch policy are shown in table 4. For patients not switched in line with the policy, 54.5% (12/22) of patients were switched >24 h after the criteria were fulfilled and 45.5% (10/22) were never switched to oral therapy. For the 22 patients treated with intravenous antibiotics >24 h after the criteria were fulfilled, the mean number of days with inappropriate intravenous treatment was 2.36 days (range: 1–5 days). This yielded a total number of 172 extra intravenous doses at a cost of £675.50; the cost of the corresponding oral step-down antibiotics was £31.30. Thus, the total cost incurred due to not switching patients from intravenous to oral was £644.20 (£31.30–675.50). Taking into account that this audit was carried out over a 3-month period, extrapolation of this to 1 year would generate an additional cost of £2576.80 (£644.20×4). The mean length of stay for patients switched in line with the policy was 6.9 days (range: 2–18 days) compared with 13.2 days (range: 4–28 days) for patients treated with intravenous antibiotics >24 h after fulfilled criteria; this equates to on average an extra 6.3 days of hospitalisation, which was shown to be statistically significant (p=0.01). Considering that the cost for a hospital bed (in medical wards) in Antrim Area Hospital is £350 per night the opportunity cost due to prolonged hospital stay for patients not switched in line with the policy was calculated at £2205 (£350×6.3) per patient or £48510 for all 22 patients; the latter could be extrapolated to an annual opportunity cost of £194 040 (£48 510×4).
Recently, much attention has been paid to the importance of antibiotic resistance, with ongoing scientific debate on the best way that healthcare organisations can devote their efforts to prevent the emergence and spread of multidrug resistant pathogens. However, it is important to note that antibiotic use in humans represents a small proportion of total antibiotic use. Antibiotics are used widely in animal health to treat and prevent disease and are incorporated into animal feed to improve growth rate. 17 Some steps to reduce this use have already been taken whereby from 1 January 2006 there was an EU-wide ban on the use of antibiotics as growth promoters. 18 Therefore controlling antibiotic use in humans is only one step in addressing the ubiquitous problem of antibiotic resistance. Nevertheless, experience has shown that controlling resistance in hospitals, while difficult, is achievable. The objective of the present audit was to study the quality of treatment of LRTI and UTI and to identify targets for quality improvement. The rate of clinical cure on first-line treatment and cost savings associated with intravenous-to-oral switch were also assessed. The results showed that 68.6% of the prescribed antibiotics were chosen according to the hospital antibiotic policy. Other studies investigating adherence to hospital antibiotic policies have demonstrated similar findings.19 ,20 As the importance of antibiotic policies in curbing inappropriate antibiotic use is proven,8–11 and considering that antibiotic resistance is a problem,21 full adherence with hospital antibiotic policies should be targeted. In addition, the CURB-65 score was documented in notes of 80% of the patients with CAP; however 46% of the latter patients were not treated in line with the CURB-65 score. Again, full documentation and treatment adherence in relation to CURB-65 should be sought.
The findings of the study demonstrated that clinical cure on first-line treatment was similar in both groups of patients whether or not treated in line with antimicrobial policy. These results are not surprising since several different effective antimicrobial regimens are available. However, the importance of these findings are reflected in the fact that physicians are primarily concerned with the effects of antimicrobial therapy on patient outcomes in the short term and have a limited perception of the problem of antimicrobial resistance or length of stay, a possible result of their prescribing practice in their own institutions.21–23 In a study about identifying the preferred patterns of antibiotic prescribing for patients, physicians rated the issue of contributing to antibiotic resistance lowest among seven factors influencing their choices.24 Analysis of data showed that percentages of delayed and missed doses for all antibiotics were 21.8% and 8.6%, respectively. The right timing of antibiotic doses is essential to maintain therapeutic levels of the drug in the blood. Delays in commencing antibiotics and low levels of the antibiotic in the blood due to delayed or missed doses may make these drugs less effective and may increase the chances of development of antibiotic resistance. In relation to the most frequently used antibiotics in LRTI (n.s.p) and LRTI/UTI patients, co-amoxiclav and the combination of co-amoxiclav and clarithromycin were among the most prescribed antibiotics. Several publications have reported on the contribution of co-amoxiclav and macrolides to the emergence and spread of resistance in hospitals.25–27 The findings suggest the need to optimise their use through establishing an antibiotic policy for the treatment of LRTI (n.s.p) and LRTI/UTI patients.
The findings of the study showed that appropriate switching from intravenous to oral therapy, according to the hospital antibiotic switch policy, was relatively low (58.5%). Compliance with intravenous-to-oral antibiotic switch policies may be beneficial in that oral therapy is more convenient for the patient and hospital staff, and it reduces the occurrence of vasculitis associated with peripheral vascular cannulae. In addition, the costs associated with oral treatment are much reduced when compared with intravenous therapy. Surveys have shown that approximately 40% of hospitalised patients who are treated with antibiotics receive an intravenous agent,28 and as such intravenous antibiotics account for a considerable portion of a hospital's medication budget. This audit has shown the opportunity cost of prolonged treatment with intravenous antibiotics and in addition has highlighted the increased length of stay associated with excess intravenous antibiotic treatment. The benefits of compliance to a hospital intravenous-to-oral antibiotic switch policy, that is, a reduction in drug costs, length of stay and patient-related morbidity, have been shown in several studies.29–33 This study has the limitation of the small sample size used to assess antibiotic policy adherence. Although the initial sample of patients was 113, many of these patients were excluded as they were diagnosed with LRTI (n.s.p) of LRTI/UTI for which there were no treatment recommendations in the antibiotic policy. Other patients were excluded as they had received previous empirical antibiotic therapy or treatment was based on the results of sputum or urine cultures. This audit would have benefited from a longer study period that would allow the assessment of the rate adherence to the antibiotic policy in a larger patient population. In addition it would be useful to investigate the effect of inappropriate use of antibiotics on local resistance patterns.
In conclusion, the value of antibiotic stewardship in controlling resistance and improving patient health outcomes has been proven and documented in numerous studies; however, more work on its actual implementation in hospitals is needed. The present audit identified a number of targets for quality improvement including adherence to antibiotic policy, documentation of the CURB-65 score in patients’ notes and treating patients accordingly, improving the prescribers’ perception of the problem of antimicrobial resistance, addressing the problem of missed and delayed doses, optimising the use of co-amoxiclav and macrolides, and maintaining adherence to the hospital intravenous-to-oral antibiotic switch policy. One possible solution for the successful implementation of such recommendations is via regular monitoring using simple audit tools.34 This audit suggests that the quality of antibiotic prescribing could be improved by auditing current prescribing practices, identifying performance indicators, and as such informing more focused interventions aimed at improving treatment protocols and the containment of antibiotic resistance.
What is already known on this subject
The continued emergence of antibiotic resistant pathogens poses major consequences for the individual and healthcare institutions.
Adequate antimicrobial stewardship can lead to less overall and inappropriate antimicrobial use, and less emergence of antimicrobial resistance.
Ensuring successful implementation and compliance with antibiotic stewardship is critical.
What this study adds
The study identified a number of targets for antibiotic stewardship quality improvement, and suggests that the quality of antibiotic prescribing could be improved by measuring and addressing such performance indicators.
More work on robust antibiotic stewardship implementation and sustainability mechanisms in hospitals is needed.
Contributors HM, MAA, JCM and MGS contributed to the design of this work, the analysis plan of data, and drafted the work, revised it, and approved the publication. MPK, GC and FAM contributed to the acquisition of data, analysing the data with interpretation of the findings, drafted the work, revised it and approved the publication.
Competing interests None.
Ethics approval The audit was registered with The Trust Governance Department as a clinical audit.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.