Family Practice Vol. 19, No. 1, 12-17
© Oxford University Press 2002
Original Paper |
Influencing antibiotic prescribing by prescriber feedback and management guidelines: a 5-year follow-up
Department of General Practice, University of New South Wales,
a Family Medicine Research Unit, University of Sydney and
b The Royal Australian College of General Practitioners Training Program, New South Wales, Australia.
Zwar N, Henderson J, Britt H, McGeechan K and Yeo G. Influencing antibiotic prescribing by prescriber feedback and management guidelines: a 5-year follow-up. Family Practice 2002; 19: 12–17.
Received 1 May 2001; Accepted 3 September 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. The extent of use of antibiotics for upper respiratory tract infection (URTI) prompted a previous study of an educational intervention based on prescriber feedback and management guidelines. This study demonstrated a reduction in antibiotic prescribing for URTI and a more appropriate choice of antibiotic for tonsillitis/streptococcal pharyngitis. There are few long-term follow-up studies of educational programmes of this kind.
Objectives. This follow-up study aimed to examine if the reduction in antibiotic prescribing observed in the intervention group of the original study remained present after 5 years, and how the prescribing behaviour of the GPs involved in the follow-up differed from a large national survey of GP prescribing.
Methods. Attempts were made to contact the 157 GPs involved in the original study. Of these, 121 were both located and currently working in general practice. Ninety-six consented to take part and, of these, 79 completed a morbidity and treatment survey of 100 patient encounters (response rate 65.3%).
Results. The intervention group (n = 37) maintained their pattern of prescribing of antibiotics for URTI and choice of antibiotic for tonsillitis/streptococcal pharyngitis, with no significant change between the completion of the original study and the 5-year follow-up. The control group (n = 42) showed a downward trend in antibiotic prescribing for URTI, with the effect that no significant differences remained between groups at the 5-year follow-up. At the 5-year follow-up, both groups prescribed significantly fewer antibiotics for URTI and showed greater adherence to prescribing guidelines for tonsillitis/streptococcal pharyngitis than participants in a large national GP survey (n = 984).
Conclusion. This study demonstrated maintenance of prescribing behaviour in the intervention group in the long term. However, the changes in prescribing observed in the control group and the power limitations of the study make it uncertain whether this was the result of a sustained effect of the educational intervention. The differences in both groups from the large national GP survey suggest that other influences on prescribing (such as participation in vocational training for general practice) were also having an important effect.
Keywords. Antibiotics, education, follow-up, prescribing, respiratory tract infections.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The extent of use of antibiotics to treat upper respiratory tract infections (URTIs) in general practice and the need for effective educational programmes on this topic prompted the original study.1 This study involved 157 general practice trainees randomized to a treatment group (n = 78) that received an intervention on antibiotic use and a control (n = 79) that received an intervention on an unrelated topic (benzodiazepine use). The intervention on antibiotic use involved prescriber feedback and management guidelines delivered first in a written form to the entire group then as an educational outreach visit to a subgroup of high prescribers. This intervention was effective in reducing the rate of prescribing of antibiotics for undifferentiated URTI and improving the choice of antibiotic for tonsillitis/streptococcal pharyngitis.
Though recent Australian data show evidence of a small decrease in use of some specific types of antibiotics,2 antibiotics remain the group of drugs most commonly prescribed by GPs in Australia, representing 17.8% of all prescriptions.3 The problems of increasing bacterial resistance to antibiotics continue and are attributed to the high level of antibiotic usage in the last few decades.4,5 Froom et al.6 have observed that levels of antibiotic resistance vary between countries, with the highest rates occurring in countries where common respiratory infections and otitis media are treated routinely with antibiotics.
There are only a small number of published studies on the long-term effects of educational programmes on prescribing. Molstad et al.7 found a persistence of effect on antibiotic use in a 5-year follow-up of an educational programme based on group discussion and agreed indications for antibiotic treatment of respiratory infections. This study was not randomized but compared prescribing for the patients of one practice with prescribing data from the geographic region. May et al.8 studied the ongoing effect of an educational outreach programme for doctors in the community and found evidence of sustained changes in prescribing of non-steroidals over a 5-year period.
This 5-year follow-up study aimed to examine if the reduction in antibiotic prescribing observed in the intervention group of the original study remained present after 5 years. The study also examined if the prescribing behaviour of the GPs involved in the follow-up differed from a large national survey of GP prescribing (n = 984).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Attempts were made to contact the 157 GPs from the original study from address details held by the Royal Australian College of General Practitioners and the New South Wales Medical Registration Board. Sixteen could not be located and a further 20 were not available because they were overseas, on leave or no longer in general practice. This left 121 available, of whom 96 consented to take part. Of these, 79 completed a morbidity and treatment survey of 100 patient encounters (response rate 65.3%).
The survey instrument, data entry and coding methods used were identical to those used in a continuous national study of general practice activity in Australia known as the Bettering the Evaluation and Care of Health (BEACH) survey.9 This enabled comparisons to be made with national data.
The representativeness of the doctors who took part in the follow-up compared with the original study was examined. Participants in the intervention and control groups at follow-up were compared on the basis of sex and age with the groups at the time of the completion of the final survey (survey 3) of the original study.
The antibiotic intervention and control groups from survey 3 were divided on the basis of their participation in the follow-up for comparison of prescribing behaviour. For the 78 members of the antibiotic intervention group at survey 3, only the data from the 37 who participated in the follow-up study were used to compare their prescribing at survey 3 with that at follow-up. For the 79 members of the control group at survey 3, only the data for the 42 members who participated in the follow-up were used.
Rates were then determined for encounters, selected diagnoses and prescribing in accordance with those examined at survey 3. These were: diagnosis rates per 100 encounters for undifferentiated URTI, tonsillitis/ streptococcal pharyngitis; antibiotic prescribing for all indications, mean rate per 100 encounters; antibiotic prescribing rates for each of the above diagnostic labels, mean rate per 100 diagnoses; and choice of antibiotic prescribed for tonsillitis/streptococcal pharyngitis, mean rate of agreement with Australian Antibiotic Guidelines10 per 100 diagnoses where any antibiotic was prescribed.
All of the above rates were determined for both the intervention and control groups. Comparisons were then made to detect any differences between those at survey 3 and those from the follow-up study. The rates were also compared with those from the national BEACH Survey 1998–99 data.
The rates and proportions were determined and differences assessed using SAS.11 The data elements reported are the rates as listed above and the 95% confidence intervals (CI). As with the BEACH data, calculations in this study have incorporated the single-stage clustered study design according to Kish's formula12 to determine the amount of both within-cluster and between-cluster variance.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Comparisons to examine representativeness of the follow-up participants to the original groups
For both intervention and control groups, no significant differences were found in the comparison of sex and age distribution of participants in the follow-up with non-participants. Within the sample involved in the follow-up study, there were no significant differences between the intervention and control groups in sex or age distribution. To assess further the representativeness of the follow-up participants, comparisons of prescribing behaviour at survey 3 between follow-up participants and non-participants were undertaken. For both the antibiotic intervention group and the control group, there were no statistically significant differences in the prescribing behaviour at survey 3 between the follow-up participants and non-participants.
Comparisons of prescribing behaviour
No statistically significant differences were evident in the prescribing behaviour of follow-up participants in the antibiotic intervention group between survey 3 and at follow-up for any of the conditions of interest (Table 1
).
|
As shown in Table 2
, no statistically significant differences were evident in the prescribing behaviour of follow-up participants in the control group between survey 3 and at follow-up. A downward trend was noticed in the antibiotic prescribing rate for URTI between survey 3 (35.4 per 100 URTI problems, 95% CI 25.4–45.5) and follow-up (17.7 per 100 URTI problems, 95% CI 9.6–25.7), but the 95% confidence intervals rendered this change not statistically significant.
|
No statistically significant differences in antibiotic prescribing were found between the antibiotic intervention group and the control at follow-up (Table 3
). There was a slight downward trend in antibiotic prescribing for URTI by the control group (17.7 per 100 URTI problems, 95% CI 9.6–25.7) compared with the antibiotic intervention group (26.7, 95% CI 15.2–38.2), but the confidence intervals rendered this difference of no statistical significance.
|
Comparisons with BEACH participants
The characteristics of BEACH participants reflect those of the general practice population in Australia. The follow-up participants were more likely to be female (65.7% in the intervention group; 62.5% in the control group) than the BEACH participants (30%). As expected, they were younger, with >90% of both groups aged <44 years, compared with 42.6% of BEACH participants. All follow-up participants had completed the Royal Australian College of General Practitioners Training Programme, while only 30% of BEACH participants had completed the Training Programme.
The antibiotic intervention group and BEACH participants
There were no significant differences in the overall antibiotic prescribing rates for all combined respiratory conditions between the antibiotic intervention group and the BEACH participants. Table 4 demonstrates that there were some differences in the prescribing behaviour for individual conditions of the antibiotic intervention group at follow-up and the national BEACH participants. The prescribing rate of antibiotics for URTI was significantly lower for the intervention group (26.7 per 100 URTI problems, 95% CI 15.2–38.2) than for BEACH GPs (42.3, 95% CI 39.2–45.4). The intervention group showed a significantly higher rate of agreement with prescribing guidelines for tonsillitis/ streptococcal pharyngitis (78.4 per 100 tonsillitis/ streptococcal pharyngitis problems, 95% CI 62.9–94.0) compared with BEACH GPs (46.8, 95% CI 41.9–51.6).
|
The control group and BEACH participants
There were no significant differences in the overall antibiotic prescribing rates for total respiratory conditions between the control group and the BEACH GPs (Table 5
). The prescribing rate of antibiotics for URTI was significantly lower for the control group follow-up participants (17.7 per 100 URTI problems, 95% CI 9.6–25.7) than for BEACH participants (42.3, 95% CI 39.2–45.4). The control group follow-up participants had a significantly higher rate of agreement with the recommended treatment guidelines for tonsillitis/streptococcal pharyngitis (76.1 per 100 tonsillitis/streptococcal pharyngitis problems, 95% CI 63.5–88.7) compared with BEACH participants (46.8, 95% CI 41.9–51.6).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The aim of the study was to evaluate the long-term effects, at 5-years follow-up, of educational interventions intended to improve the prescribing behaviour of general practice trainees. It is one of only a small number of studies to examine for long-term effects of education on prescribing.
In order to compare the two intervention groups over time, it was first necessary to establish the extent to which participants at follow-up represented the original intervention groups at the completion of the original study (survey 3, which represented the baseline measure for this study). The results demonstrate that, even with the 65% response rate, the follow-up participants were representative of the original intervention groups at survey 3, in terms of age and gender distribution and of prescribing behaviour.
No statistically significant differences were found between the prescribing rates of the antibiotic intervention group at survey 3 and at follow-up, suggesting that the effect of the intervention had been maintained over time. However, it is not possible to attribute this effect with certainty to the intervention, as there were no longer any significant differences between the prescribing rates of the intervention and control groups. The differences observed at survey 3 no longer existed at the 5-year follow-up due to an unexpected decrease in prescribing rates of antibiotics for URTI by the control group. Another unexpected change occurring in the control group was a high rate of agreement with prescribing guidelines for tonsillitis/streptococcal pharyngitis. The rate of agreement with these guidelines at follow-up was similar to that of the antibiotic intervention group.
Both intervention and control groups were prescribing antibiotics for URTI at significantly lower rates than the BEACH participants and significantly more often in agreement with prescribing guidelines for tonsillitis/ streptococcal pharyngitis. This finding suggests that some common factor between the groups was a more powerful influence on prescribing than the antibiotic intervention.
The obvious common factor which may have influenced prescribing is participation in vocational training for general practice. All participants for both follow-up groups had completed the Training Programme, all within the previous 5 years, compared with only 30% of the BEACH participants, and these may have completed training at any time within the past 25 years. This would support the findings of the RACGP Report13 that by 3–5 years post-training, there were fewer differences between Training Programme and non-Training Programme GPs in terms of their consultation patterns, but their prescribing behaviour continued to show real differences.
Another possible explanation for the differences in prescribing behaviour observed are the sex and age differences between the two samples. The gender distribution of BEACH doctors, with two-thirds male and one-third female, while representative of general practice, is almost opposite to that of the two follow-up groups. There are also large differences in the age distribution between BEACH and follow-up participants. Over 97% of follow-up participants were <45 years old while only 43% of BEACH doctors were in this age group. The differences in both groups when compared with BEACH doctors may be a result of participation in the Training Programme, but it is difficult to determine if the differences are not as likely to be related to the differences in age or gender.
Though there are a number of possible explanations for the pattern of prescribing observed at the 5-year follow-up, the results of this study need to be viewed with caution. The wide, overlapping confidence intervals for most of the criteria examined indicate that the null hypotheses cannot be rejected. Although the response rates for the project (60% for the intervention group; 71% for the control group) were quite adequate for a follow-up study, they were not sufficient in total to provide conclusive results with sufficient power.
While the power limitations make it difficult to draw definite conclusions from the results of this study, research projects such as this which examine the long-term benefits of educational programmes need to be pursued. Such research is important in determining whether resources are being used appropriately for educational programmes which have long-term benefits. Soumerai et al.14 found concurrent reminders useful in influencing prescribing behaviour over a period of 1 year, and May et al.8 found ongoing educational outreach successful in sustaining changes to prescribing behaviour over a 5-year period. Munck et al.15 showed effects on choice of antibiotic over a 3-year period in GPs involved in an ongoing audit process. These studies suggest that methods of continual reinforcement may be the most effective in producing sustained changes in behaviour. Our study provides some support for this argument in that the involvement of both groups in an ongoing educational programme (vocational training) may have overwhelmed the effect of a short-term programme provided to one of the groups.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Zwar N, Wolk J, Sanson-Fisher R, Kehoe L. Influencing antibiotic prescribing in general practice: a trial of prescriber feedback and management guidelines. Fam Pract 1999; 16:495–500.
2 MacIssac P. Use antibiotics wisely. MediScene 1999; 9:2.
3 Britt H, Sayer GP, Miller GC et al. General Practice Activity in Australia 1998–99. AIHW Cat. No. GEP 2. Canberra: Australian Institute of Health and Welfare (General Practice Series no. 2).
4 Turnidge J, Zwar N. Minimising inappropriate prescribing of antibiotics. Med Today 2000; 1:70–81.
5
Bauchner H, Pelton SI, Klein JO. Parents, physicians and antibiotic use. Pediatrics 1999; 103:395–398.
6
Froom J et al. Antimicrobials for acute otitis media? A review from the International Primary Care Network. Br Med J 1997; 315:98–101.
7
Molstad S, Ekedahl A, Hovelius B, Thimansson H. Antibiotic prescription in primary care: a five year follow-up of an education program. Fam Pract 1994; 11:282–286.
8 May FW, Rowett DS, Gilbert AL, McNeece JI, Hurley E. Outcomes of an educational-outreach service for community medical practitioners: non-steroidal anti-inflammatory drugs. Med J Aust 1999; 170:471–474.[Web of Science][Medline]
9 Britt H, Sayer GP, Miller GC et al. BEACH Bettering the Evaluation and Care of Health. A Study of General Practice Activity. Six-month Interim Report. AIHW cat. No. GEP 1. Canberra: Australian Institute of Health and Welfare (General Practice Series no. 1).
10 Antibiotic Guidelines, 5th edn. Victorian Postgraduate Medical Foundation, 1990–91.
11 SAS Proprietary Software Release 6.12. SAS Institute Inc., Cary, NC, USA.
12 Kish L. Survey Sampling. New York: John Wiley & Sons, 1965.
13 Britt H, Miller G, Sayer G, Sedgwick D. Comparison of Practice Patterns Using Health Insurance Commission Data for GPs who Have Completed the RACGP Training Program and Those Who Have Not. Report by the Family Medicine Research Unit, Department of General Practice, University of Sydney. Sydney: Royal Australian College of General Practitioners.
14 Soumerai SB, Avorn J, Taylor WC, Wessels M, Maher D, Hawley SL. Improving choice of prescribed antibiotics through concurrent reminders in an educational order form. Med Care 1993; 31:552–558.[Web of Science][Medline]
15 Munck AP, Gahrn-Hansen B, Sogaard P, Sogaard J. Long-lasting improvement in general practitioners' prescribing of antibiotics by means of a medical audit. Scand J Primary Health Care 1999; 17:185–190.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. H. Gaur, M. E. Hare, and R. I. Shorr Provider and Practice Characteristics Associated With Antibiotic Use in Children With Presumed Viral Respiratory Tract Infections Pediatrics, March 1, 2005; 115(3): 635 - 641. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||