Articles |
The use of antibacterials in children: a report of the Specialist Advisory Committee on Antimicrobial Resistance (SACAR) Paediatric Subgroup
Paediatric Infectious Diseases Unit, 5th Floor Lanesborough Wing, St George's Hospital, Blackshaw Road, London SW17 0QT, UK
* Corresponding author. Tel: +44-208-725-3262; Fax: +44-208-725-1208; E-mail: mike.sharland{at}stgeorges.nhs.uk
 |
Abstract |
|---|
 Top Abstract IntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
The Paediatric Subgroup of SACAR has identified major changes in antibacterial use for children. The number of prescriptions for antibacterials for children in primary care in 2000 was half that in 1996, with amoxicillin still accounting for around 50% of all antibacterial prescribing. There is a marked regional variation in prescribing rates. The UK now has paediatric antibacterial prescribing rates lower than many other European countries at around 700 prescriptions/1000 children/year. The group has been using the General Practice Research Database and Hospital Episode Statistics data to identify trends in clinical disease patterns associated with this change in prescribing practice. Data are also now available from the Health Protection Agency for all paediatric bacteraemias reported over the last decade. In 1992, around 50% of bacteraemias were due to Gram-positive bacteria, but this has increased to over 75% in 2003 with Staphylococcus aureus now becoming the leading significant bacteraemia reported in children. It may now be possible to link existing childhood data sets to develop collaborative prospective antimicrobial surveillance networks, monitoring antimicrobial prescribing, bacterial isolates and clinical outcome data, and to provide updated evidence-based prescribing advice to all prescribers. Adherence to national guidelines and clinical outcome data can then be monitored in turn to provide a continuous audit process aimed at improving prescribing advice and reducing antimicrobial resistance.
Keywords: paediatrics , prescribing , surveillance
|
Introduction |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
The Paediatric Subgroup of SACAR (hereafter referred to as Paediatric SACAR) was formed because it was recognized that there is a particular problem with antimicrobial prescribing in childhood. An unpublished analysis of data from the early 1990s undertaken by the Prescribing Support Unit (PSU) showed that out of the 40 million antibacterial prescriptions per year in primary care, around 12 million were for children. Children have high rates of minor infection but because of their increased susceptibility to serious bacterial infection are frequently prescribed antibiotics. There was a concern that there may be an increasing bacterial resistance in childhood and that changes in childcare practices, particularly the marked increase of day care in pre-school groups, may be leading to increasing transfer of antibiotic-resistant organisms within these environments. There was also recognition that there were very few data obtained on changes in antimicrobial prescribing, antimicrobial resistance patterns and clinical disease incidence in children in the UK over the last decade. This report summarizes the progress made towards obtaining this information by Paediatric SACAR over the last 3 years.
|
Original work plan |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
Paediatric SACAR developed a multidisciplinary group that incorporates expertise from many different clinical disciplines. This includes representation from academic general practice and hospital paediatrics, pharmacy representation from both hospital and academia and the Information Centre for Health and Social Care (IC). The group also has infection control nursing representation and clinical epidemiology representation. The Health Protection Agency (HPA) provides critical microbiological data and analysis. The primary aim of the group has been to combine microbiological, clinical and prescribing data to create an integrated surveillance system that can help to monitor the impact of antibiotic prescribing guidance.
The work of the group can be divided into four main areas.
Monitoring antimicrobial prescribing
National prescribing data. One of the first priorities for the group was to identify trends in paediatric antibacterial prescribing in the community. For this work, Paediatric SACAR is very grateful to the PSU based in the IC. Unless otherwise stated, all the data have been obtained by the PSU from Electronic Prescribing Analysis and Cost Tool (ePACT). The PACT system, provided by the Prescription Pricing Division of the Business Services Agency, covers prescriptions from general practices (GPs) in England.
It was not possible to identify which prescriptions were for children, as the data available did not include any patient or clinical information. The group therefore undertook a validation of the assumption that the majority of prescriptions for liquid formulations of antibacterials are for children. Our direct check of prescriptions confirmed that virtually all oral liquid formulations of antibacterials are for children. Older children are likely to be given solid oral formulations, so reporting on liquid oral formulations only will underestimate the total use of antibacterials in children. Another factor that may lead to an underestimate of use of antibacterials in children is the direct supply (without a prescription) by out of hours service providers. It is known from discussions that amoxicillin liquid formulations are among the most frequently supplied medicines by this route.
Chapter 5 of the British National Formulary (BNF), the UK national guidance for prescribing, covers infections. Figure 1 shows the annual data for each calendar year for oral liquid formulations of antibacterials (BNF 5.1) for the last 10 years (source: Prescription Cost Analysis database, PSU). There was a marked decline in the prescribing of these items between 1996 and 2004. The number of prescription items for 2000 is 64% of that for 1996 and for 2005 is 55% of that for 1996. The increase for 2005 over 2004 is 0.92%. The net increase in cost for 2005 over 2004 is due to changes in central pricing policy introduced during 2005. There had therefore been around a 50% reduction in paediatric antimicrobial community prescribing over the last decade that we were unaware of.
|
The subgroup also questioned which antibacterials are given to children. Figure 2 shows the national prescribing from selected sections of BNF 5.1 (antibacterials). The data are presented as prescription items per month over a 6 year period. Penicillins are the most frequently prescribed antibacterials. There is a significant seasonal variation with peak use over the winter months. The vast majority of these items are oral formulations, with a small number of injections. Topical formulations are not included. The most commonly prescribed antibacterials are amoxicillin, phenoxymethylpenicillin, flucloxacillin, erythromycin and trimethoprim.
|
Figure 3 shows the monthly prescription items for liquid formulations only for penicillins (BNF 5.1.1) and illustrates the very marked seasonal variation with the peak occurring in December and the trough in August each year. This is more prominent in children, with winter prescribing well over double the summer antibacterial prescribing rates. Interestingly, this is not true for flucloxacillin where the highest rate of prescribing occurs in August each year (when flucloxacillin accounts for
20% of all liquid formulations of penicillins), whereas in December flucloxacillin accounts for < 8% of these prescriptions (Figure 4).
|
|
Worryingly, there is some evidence of a recent slight increase in prescribing of amoxicillin. Figure 5 shows the annual (financial year) number of prescription items of oral liquid formulations for each of amoxicillin, flucloxacillin and phenoxymethylpenicillin for the last 5 years. Amoxicillin accounts for
50% of all oral liquid antibacterials (BNF 5.1) and 75% of all oral liquid penicillins (BNF 5.1.1) prescribed in primary care in England. The number of items for amoxicillin shows a steady decline year on year, until 2005–06 over 2004–05 where there is a 3.1% increase in items. A similar small increase has also been seen for erythromycin.
|
The group looked in more detail at amoxicillin prescribing as a marker antibiotic for investigating variation in usage and the potential for monitoring a reduction in prescribing rates. Figure 6 shows the number of prescription items for oral liquid formulations of amoxicillin per person aged under 18 for each Strategic Health Authority (SHA) over the last 5 years. Clearly, older children within this age range are unlikely to receive oral liquid formulations of medicines, but this does allow the group to apply some population weighting to the SHA utilization figures. It is also possible to make some estimates about the variation in the rate of prescribing of amoxicillin in different areas of England and look at trends over the last 5 years.
|
These data suggest that there is a marked variation in prescribing rates. For example, in 2001–02, Birmingham and the Black Country SHA had the highest rate of prescribing of oral liquid formulations of amoxicillin, and that this rate declined substantially over the next few years. The majority of the SHAs show an increase in the rate of prescribing for 2005–06 over 2004–05. Paediatric SACAR would now like to look further at variation in prescribing by primary care trust and practice. It is likely that this will show wide variation and possible further potential scope for reduction in prescribing. Modelling the variation in prescribing by practice and identifying the factors associated with high prescribing for antimicrobials also require further work. Future research in this area includes quantifying the difference between high and low prescribers down to practice level and determining the drivers for this variation.
International trends. When looking at the fall in antibacterial prescribing in the UK, the group noted from the literature that similar, though smaller falls, had been seen in the USA and Europe.1,2 However it was also clear that there remained marked differences in prescribing rates across Europe. Intercontinental Medical Statistics (IMS) data for antibiotic prescriptions given to 0–17-year-olds during October 2002–September 2005 were extracted for France, Germany, Italy, Spain, UK and USA, and population data for 0–17-year-olds was obtained from UNICEF (http://www.unicef.org). Paediatric antibiotic prescribing decreased in 2004/05 when compared with 2002/03 for all countries included in this study (Figure 7). Of these countries, Spain had the highest average rate of prescribing (1581.0, 95% CI 1467.2–1694.9 prescriptions per 1000 persons), compared with the UK which had the lowest rate (608.67, 95% CI 605.0–612.3 prescriptions per 1000 persons).
|
Seasonal variation for paediatric antibiotic prescribing was observed for all countries throughout the study period. Interestingly, prescribing was lowest in September and peaked in March each year for all countries, except the UK, for which the peak was observed in December as discussed above. France displayed the greatest degree of seasonal fluctuation, whereas UK varied the least. It should be noted that Dutch prescribing is still < 50% of UK prescribing rates.
Paediatric SACAR is now working with the European Society of Paediatric Infectious Diseases to determine the available data on antimicrobial prescribing, and resistance data that are available from different EU countries.
General Practice Research Database and IMS prescribing trends. Paediatric SACAR has been working with the Department of Paediatric Pharmacy Research at the London School of Pharmacy and the Institute of Child Health in London to use General Practice Research Database (GPRD) and IMS data to look in more detail for the reasons antibiotics are prescribed in the community. It can be been seen from these data that there are now around 3.3 million prescriptions for liquid formulations of amoxicillin every year for children in the UK, representing around half of all antibacterial prescribing in children, as there are now around 6.6 million prescriptions each year. The group has focused on amoxicillin prescribing for otitis media (OM), as amoxicillin for children with OM probably has significant potential for further reduction in prescribing.
The group initially examined the prescribing trends for amoxicillin for OM, looking at the GPRD codes for all OM, but excluding the chronic OM conditions. These codes showed a 56% decline in OM over the study period (1990–2005), yet antibacterial prescribing for OM in GPRD remained relatively stable. However, if the analysis is performed using acute OM (AOM) codes, the trend is more stable showing no reduction in antibiotic prescribing for AOM (Figure 8).
|
The group therefore tried to determine what amoxicillin is prescribed for in general practice. Figure 9 shows the results of an analysis in GPRD of a random sample of paediatric amoxicillin prescriptions to identify the clinical indication for that prescription. It clearly demonstrates that a high proportion of antibacterials in the community are still prescribed for minor respiratory tract infection (RTI).
|
This work has shown how complicated it is to determine the reasons for changes in prescribing patterns from clinical databases. It is most likely that the reasons for primary care doctors prescribing amoxicillin have changed over time, and that clinicians may also alter the coding of their clinical diagnosis depending on whether they have decided to prescribe or not prescribe. The recent publication3 of a meta-analysis of antibiotic prescribing for OM has suggested that prescribing can be safely focussed on children aged under 2 years with more severe disease. The group is currently identifying what proportion of all AOM amoxicillin prescribing would fall within these limits. If new guidance, from, for example, the Children's BNF, in the future suggests this is the only appropriate group to prescribe for then modelling can be undertaken to determine the impact of this guidance.
Further work needs to be performed on the modelling of different prescribing practice for OM and other clinical disease including RTI. It is clear that both upper and lower RTIs are associated with high antibacterial prescribing in primary care. Following the introduction of Prevenar the conjugate pneumococcal vaccine into the routine primary immunization schedule, with a catch-up programme for those under 2 years of age, further modelling can be done on the impact of the vaccine on RTIs in childhood and the consequent reduction in antibiotic prescribing that could be expected.
The other major antibacterial where the prescribing rate has not fallen, but has remained stable over time is trimethoprim. This is a concern as Haemophilus influenzae resistance has fallen for ampicillin and remained stable for trimethoprim. Recent guidance from the National Institute of Health and Clinical Excellence (NICE; http://www.nice.org.uk) has suggested that trimethoprim should not be prescribed routinely for antibiotic prophylaxis for urinary tract infections in children, suggesting that a further reduction could be possible in this clinical area.
There is a need to undertake further work to identify more clearly the reasons for the overall fall in antimicrobial prescribing. There has been a continued fall in reports to the HPA of viral infections such as respiratory syncytial virus.4 The overall key question is to attempt to determine a safe and appropriate level of antimicrobial prescribing. This may change over time as the burden of disease changes due to changes in vaccination policies and predisposing viral infections in childhood.
The group noted early on that there is no systematic use of NHS clinical disease databases to follow trends in clinical disease patterns for children. For common childhood diseases this appears to be the formal responsibility of no particular organization outside of a research setting. Having noted a sharp fall in antimicrobial prescribing we were concerned that this may have been associated with changes in clinical disease pattern. There had previously been reports of a possible higher incidence of mastoiditis in the Netherlands associated with their very low prescribing rates for OM. With the help of the Department of Primary HealthCare in Oxford an increase was noted in the incidence of mastoiditis from Hospital Episode Statistics (HES) data that was temporally associated with the fall in antibiotic prescribing (Figure 10).5
|
Paediatric SACAR is now conducting a detailed case–control study in GPRD of 800 cases of childhood mastoiditis to determine whether childhood mastoiditis is associated with failure to prescribe antibiotics in primary care. We think this is unlikely as we have previously noted that prescribing for AOM appears not to have fallen over time.
There have also been reports suggesting an increase in empyema in childhood over the last few years. Similar methodology can be undertaken to look at whether this is a true increase using a combination of HES data and GPRD and to undertake case–control studies to determine the association with prescribing or not prescribing antibiotics. This methodology could be utilized in a number of clinical areas, and the group is looking at HES and GPRD trends of other common childhood bacterial infections at present.
There is clearly a need when the Department of Health (DH) is actively promoting policies and guidelines aimed at reducing antibiotic prescribing to develop a formal mechanism to determine whether these policies could be associated with harm. The group is also currently working with the HPA to study in more detail the trends in all childhood death certificates that include any infection in the notification. This is obviously the most extreme form of complication that could be associated with changes in antimicrobial prescribing, and although hopefully there are unlikely to be any major linkages identified, it is important to determine this at least retrospectively.
Paediatric SACAR noted that there had been no detailed summary of microbiology isolates from the HPA specifically focusing on children and has therefore been looking at the trends in aetiology of bacteraemia in children. All paediatric bacteraemia data from 1992 to 2005 have been extracted from LabBase 2 and are undergoing analysis, but this report has focused initially on children aged < 5 years.
The total number of reports of bacteraemia in children aged < 5 increased by 39% from 3717 in 1992 to 5347 in 2005. Trends in the number of reported bacteraemias due to different species of bacteria are shown in Figure 11, and they show that 65% of the cases involved children aged < 1 year, with 18% of all cases aged < 7 days. Sex was known for 98% of reported cases, and in each year, there were more cases in males than in females, either in total or in stratified age groups. It should be emphasized that there are no data available on the number of blood cultures taken overall during the study period, and it is therefore not possible to determine true rates of bacteraemia from the HPA data.
|
Gram-positive versus Gram-negative bacteria. Throughout the entire study period, there were more bacteraemias caused by Gram-positive bacteria than by Gram-negative bacteria. Moreover, from 2001, the proportion of cases involving Gram-positive bacteria increased markedly (Figure 12). Although half of the cases of bacteraemia in children aged 6 months to 4 years were due to Gram-positive bacteria in 1992, this increased to
75% in 2005 (Figures 13 and 14). The largest proportional increase for a particular organism was seen in Group A streptococci which increased from 46 in 1992 to 116 in 2005 (an increase of 152%). Nonetheless, Group A streptococci were still relatively uncommon, comprising only 2% of the total number of cases in 2005.
|
|
|
Of particular note was the marked increase in bacteraemia caused by coagulase-negative staphylococci (CoNS) which increased by 112%, from 632 in 1992 to 1339 in 2005. The observed increase was most pronounced between 2000 and 2004. In 2005, 25% (1339/5347) of all reported bacteraemias were due to CoNS; 398 (47%) of these were seen in infants aged between 1 and 3 weeks (Figure 15). Although some of these data are likely to represent contaminants, the rise in all Gram-positive infections may be related to the increased use of central intravascular catheters in paediatric practice.
|
Staphylococcus aureus reports increased by 56% (622 cases in 2005), whereas Streptococcus pneumoniae increased by 23% (546 cases in 2005). Although Escherichia coli, Klebsiella spp. and Enterobacter spp showed proportional increases over the duration of the study (34, 91 and 99%, respectively), they still accounted for only 13% of the total number of cases in 2005 (Figure 11).
The proportion of cases due to particular genera or species of bacteria varied with the age of the patients. For example, CoNS were the most common cause of bacteraemia in children aged 1–3 weeks, whereas pneumococci were most common in the 6–11 months and 1–4 years age categories (Figure 15).
Since the group started work, the first Children's BNF (BNF-C) was produced in 2005. In effect, this has given for the first time UK national guidelines for antimicrobial prescribing for children. The group is currently using the paediatric bacteraemia data, therefore, to determine what proportion of bacteraemias would be treated effectively by the BNF-C empirical guidance for the septic child. This is clearly problematic as many of these bacteraemias are from children with hospital-acquired infection or with other medical problems. However, it is possible to make a reasonable estimate of which organisms are more likely to be community acquired and which hospital acquired and determine the resistance patterns in each. We are also aiming to produce a summary paper on the overall trends of paediatric bacteraemia resistance. The data will now be monitored annually prospectively.
We are also now working with the paediatric intensive care unit research network (PICANET) and the neonatal intensive care unit research network (NeoNIN) to see if we can relate clinical patient data from these networks with the microbiology results from the HPA dataset. This will enable the group to determine both the relationship between reporting bias and under-reporting in LabBase 2 in comparison to the smaller, more active reporting systems within PICANET and NeoNIN. The work on resistance patterns is of particular importance to neonatal infections with a particular focus of the group being on the increasing medical, social and financial problems of multidrug-resistant outbreaks on neonatal units.
The group also noted that there was a particular problem with staphylococcal disease in children. The incidence of both methicillin-susceptible S. aureus and methicillin-resistant S. aureus (MRSA) were rising with very limited data about the clinical associations, management, outcome and possible strategies for containment. With the help of the DH and using British Paediatric Surveillance Unit (BPSU) methodology, the group has been conducting a national study on paediatric MRSA. The interim data confirm that the predominant strain is EMRSA-15 and that the strongest clinical association is very low birth weight babies in neonatal intensive care units. This opens the possibility of an important control measure which is to enhance infection control with neonatal units (NNU).
The recent reports6 of severe disease in children caused by strains of S. aureus producing the Panton-Valentine leucocidin toxin is concerning and emphasizes the urgent need to obtain further information on the epidemiology of such infections in paediatric disease, possibly using an extension of the BPSU methodology.
It is clear that healthcare-associated infection is closely integrated into the work-plan of paediatric SACAR. The development of the DH Committee on Antimicrobial Resistance and Healthcare Associated Infection is important in this area. Bacteraemia data confirm that there is an increase in multidrug resistance in high-risk populations. The group is now working closely with other specialty groups to determine the clinical associations and management from high-risk areas including paediatric oncology.
Antimicrobial prescribing guidance
The BNF-C now gives unified prescribing advice for children in primary, secondary and tertiary care. The Chair of Paediatric SACAR is the lead clinical advisor for the BNF-C Chapter 5 (Infections), but we are concerned that BNF-C guidance needs to have a more solid foundation. There is a need to move forward with developing a formal system of evidence-based prescribing for the BNF-C and perhaps other guidance mechanisms based on a systematic annual review of all the available evidence. The BNF format has many advantages but also has some limitations in its ability to provide short evidence-based summaries of common and rarer clinical infectious problems in childhood.
The Royal College of Paediatrics and Child Health (RCPCH) produce a Manual of Childhood Infections—the Blue Book. In the USA, the definitive guidance for all antimicrobial prescribing for children is given by the Red Book. The RCPCH is considering developing the Blue Book as a web UK evidence-based resource for paediatricians integrating the different guidance available. This could be particularly beneficial as the range of prescribers increases outside the traditional area. This would need to include an annual review of newly published evidence including any new meta-analyses performed or systematic reviews, a review of any changes in microbiology patterns of resistance and any changes in clinical disease presentation. These different mechanisms could feed into formal guidance and be disseminated through the RCPCH, expanding the evidence base behind the BNF-C and acting as part of the National electronic Library for Infection as a National electronic Library of Childhood Infection. This is the area that Paediatric SACAR has least developed so far, but is at the core of future work.
Hospital prescribing. The group has had very limited success in determining the rates of hospital prescribing for antibiotics for children, because of difficulties with access to the DH/IMS data set. The group has been collaborating with the CDSC Wales to look at this with their joint data set.
|
Collaborative links |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
Throughout the last 3 years, Paediatric SACAR has worked closely with other subgroups of SACAR where their work may overlap. These include the Surveillance Subgroup, Prescribing Subgroup and Professional Educational Subgroup. NICE has also published guidance for urinary tract infections and for the febrile child. Paediatric SACAR has responded formally to the consultations for both sets of guidance pointing out the low priority that both paediatric NICE guidance seem to give to the problems of antimicrobial prescribing for children.
|
Conclusion |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
The work of Paediatric SACAR has, in the subgroup's view, demonstrated a need to develop a longer-term collaborative paediatric antimicrobial surveillance network (Figure 16). We think it may be possible to develop a pilot scheme for formulating and updating prescribing guidelines. Prospective monitoring of its effectiveness in terms of adherence, safety and efficacy could be done with established clinical networks and databases at minimal cost. The National Programme for Information Technology (NPfIT)—Connecting for Health could provide even better linkage of clinical and prescribing information at a patient level in the future. Until then, much more can be done to determine more clearly the appropriate guidance on which children can safely not be given antibiotics. Improving antibiotic guidance both in primary care and in high-risk hospital areas is still the best option to reduce the rates of antimicrobial resistance in children in the UK.
|
|
Transparency declarations |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
None to declare.
|
References |
|---|
|
TopAbstractIntroductionOriginal work planCollaborative linksConclusionTransparency declarationsReferences |
|---|
1 Finkelstein JA, Stille C, Nordin J, et al. Reduction in antibiotic use among US children, 1996-2000. Pediatrics (2003) 112:620–7.
2
Otters HB, van der Wouden JC, Schellevis FG, et al. Trends in prescribing antibiotics for children in Dutch general practice. J Antimicrob Chemother (2004) 53:361–6.
3 Rovers MM, Glasziou P, Appelman CL, et al. Antibiotics for acute otitis media: a meta-analysis with individual patient data. Lancet (2006) 368:1429–35.[CrossRef][Medline]
4 Health Protection Agency. Laboratory Reports of RSV Received by CfI from NHS and Health Protection Agency Microbiology Laboratories, by Date of Specimen, 2006/07 and Recent Years. (4 April 2007, date last accessed).
5
Sharland M, Kendall H, Yeates D, et al. Antibiotic prescribing and hospital admissions for peritonsillar abscess, mastoiditis and rheumatic fever in children: a time trend analysis. BMJ (2005) 331:328–9.
6 Kaplan SL. Community-acquired methicillin-resistant Staphylococcus aureus infections in children. Semin Pediatr Infect Dis (2006) 17:113–9.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P L Thompson, N Spyridis, M Sharland, R E Gilbert, S Saxena, P F Long, A P Johnson, and I C K Wong Changes in clinical indications for community antibiotic prescribing for children in the UK from 1996 to 2006: will the new NICE prescribing guidance on upper respiratory tract infections just be ignored? Arch. Dis. Child., May 1, 2009; 94(5): 337 - 340. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Marra, C. A. Marra, K. Richardson, L. D. Lynd, A. Kozyrskyj, D. M. Patrick, W. R. Bowie, and J. M. FitzGerald Antibiotic Use in Children Is Associated With Increased Risk of Asthma Pediatrics, March 1, 2009; 123(3): 1003 - 1010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Edgar, S. D. Boyd, and M. J. Palame Sustainability for behaviour change in the fight against antibiotic resistance: a social marketing framework J. Antimicrob. Chemother., February 1, 2009; 63(2): 230 - 237. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||