Skip Navigation


JAC Advance Access originally published online on October 24, 2006
Journal of Antimicrobial Chemotherapy 2006 58(6):1303-1306; doi:10.1093/jac/dkl432
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
58/6/1303    most recent
dkl432v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hillier, S.
Right arrow Articles by Butler, C. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hillier, S.
Right arrow Articles by Butler, C. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

When do general practitioners request urine specimens for microbiology analysis? The applicability of antibiotic resistance surveillance based on routinely collected data

Sharon Hillier1,*, Joanna Bell2, Margaret Heginbothom2, Zoë Roberts1, Frank Dunstan1, Anthony Howard3, Brendan Mason2 and Christopher C. Butler4

1 Department of Epidemiology, Statistics and Public Health Centre for Health Sciences Research, Cardiff University, Neuadd Meirionnydd, Heath Park, Cardiff CF14 4XN, UK 2 National Public Health Service for Wales, CDSC Wales Abton House, Wedal Road, Cardiff CF14 3QX, UK 3 ICDS Corporate Office, Temple of Peace & Health, Cathays Park Cardiff CF10 3NW, UK 4 Department of General Practice, Centre for Health Sciences Research, Cardiff University Neuadd Meirionnydd, Heath Park, Cardiff CF14 4XN, UK

*Corresponding author: Tel: +44-29-20-687228; Fax: +44-29-20-687236; E-mail: sharon.hillier{at}nphs.wales.nhs.uk

Received 12 June 2006; returned 2 August 2006; revised 25 September 2006; accepted 28 September 2006


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Transparency declarations
 References
 
Objectives: We do not know how representative reported levels of resistance to antibiotics in urinary tract infections (UTIs) are as there is wide variation in the rate of urine specimens submitted to microbiology laboratories by general practices. We used a questionnaire to investigate variation in sampling for patients with suspected UTI to explore any systematic bias that may influence interpretation of surveillance data based on routine data.

Methods: We sent a questionnaire to a stratified random sample of general practitioners (GPs) in Wales for self-completion. The GPs were presented with six clinical scenarios and asked about their proposed clinical management.

Results: We found that nearly all of the GPs indicated they would request a specimen for scenarios representing a probable UTI in a female child and a probable asymptomatic UTI in pregnancy. There was some variation between the GPs about sampling in a situation of treatment failure in an older woman and recurrent UTI in a male diabetic, with 90% and 81%, respectively, indicating they would request a specimen for these scenarios. The greatest variation was in relation to scenarios concerning the management of a probable uncomplicated UTI, and early patient symptoms with pressure to prescribe, with 56% and 33% of GPs, respectively, indicating they would request a urine specimen for laboratory analysis.

Conclusions: In the light of this reported sampling behaviour, it is likely that there is a systematic bias in surveillance data based on routinely collected data, with samples from cases of uncomplicated UTI being under represented, potentially leading to an overestimation of true resistance rates.

Keywords: urinary tract infections , UTIs , community, survey , questionnaire


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Transparency declarations
 References
 
There is a wide variation in the rate of urine specimens submitted by general practices to microbiology laboratories. The range for Wales is from 0.6 to 237.2 specimens per year, per 1000 registered patients.1 Reported levels of resistance to antibiotics are generally based on this routine data but we do not know how representative these data are of the general population of patients consulting with symptoms of urinary tract infections (UTIs). This routine data may contain high numbers of specimens from patients with recurrent UTIs and treatment failures, possibly with higher levels of resistant organisms compared to data derived from systematic sampling.2,3 We therefore designed a questionnaire to investigate variation in sampling for patients with suspected UTI to explore any systematic bias that may influence interpretation of surveillance data based on routine data.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Transparency declarations
 References
 
In 2002, 208 general practitioner (GP) practices were randomly selected from the list of all practices in Wales stratified by sampling rates and Health Authority. To produce generalizable data the aim was to receive questionnaire responses from 20% of practices in Wales. Each practice was contacted by telephone and a questionnaire pack was sent to those agreeing to participate. The previously piloted pack comprised a general questionnaire on practice demographics for the practice manager to complete and a questionnaire on clinical information for each GP in the practice to complete. The practices were sent financial compensation for time spent on the questionnaires once their packs were returned completed.

The clinical questionnaire presented six clinical scenarios which were a range of possible presentations of UTI (Table 1) and the GPs were asked whether or not they would request the patient to submit a urine specimen.


View this table:
[in this window]
[in a new window]

  		Table 1. Clinical scenarios

 
The questionnaire then asked the GPs about their general management of UTIs with questions on whether or not they had an agreed practice policy; their estimate of the proportion of patients with clinically suspected UTI that they would routinely request a urine specimen from; their management of treatment failure; and how their laboratory antibiotic susceptibility results influenced their choice of empirical therapy and management of treatment failure.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Transparency declarations
 References
 
Response rate

Of the 208 practices initially contacted, 176 agreed to take part in the study. Completed questionnaires were received from 103 practices (49.5% of total contacted initially) and 293 individual GPs. Questionnaires were available from 16% of GPs (293/1783) and from 20% of the practices in Wales (103/508). There was no response given on some of the individual questions and therefore the denominator is less than 293 in some of the results. The age distribution of the 293 GPs that replied were 11% aged under 35, 40% were aged 35–45 years, 37% were aged 46–55 and 12% were aged over 55; and 63% (186/293) of the GPs were male. There were no statistical differences with respect to age or gender between respondents and all GPs in Wales. However, fewer GPs ‘over 55 years of age’ responded [36 (12.4%) versus 324 (18.2%)]. Twenty-seven per cent of responding GPs (78/293) had been qualified for <10 years, 43% were qualified between 10 and 19 years, and 30% had been qualified at least 20 years. Seventy-three per cent (215/293) had achieved post-graduation diplomas or degrees, with 19% stating that they had received some post-graduate or continuing medical education on UTI management and 22% on antibiotic resistance.

Investigating the reasons why a GP requests a specimen

Table 2 shows the percentage of GPs who indicated that they would request a urine specimen for each clinical scenario.


View this table:
[in this window]
[in a new window]

  		Table 2. Percentage of GPs who indicated that they would request a urine specimen for each clinical scenario

 
Fifty-three per cent of GPs indicated that they would have a low threshold and send urine specimens for at least five of these clinical scenarios, 35% were classified as having a moderate threshold indicating that they would send for four of the scenarios and 12% of GPs had a high threshold as they would request a specimen for less than four scenarios.

Table 3 shows the GPs own estimate of the proportion of patients with clinically suspected UTIs from whom they requested a urine specimen.


View this table:
[in this window]
[in a new window]

  		Table 3. GPs own estimate of the proportion of patients with clinically suspected UTIs from whom they requested a urine specimen

 
Thirty-eight (13%) of the GPs stated that there was a practice policy for requesting urine specimens.

How GPs thought that the laboratory results influenced their choice of antibiotics

Two hundred and twenty-four (76%) of the GPs indicated that laboratory test results ‘always or often’ influenced their choice of antibiotic for general empirical therapy of UTIs. However, 113 GPs (41%) would treat Case 2—‘the probable uncomplicated UTI’ without requesting a specimen.

Two hundred and eighty-five (97%) of the GPs indicated that results ‘always or often’ influenced their change in therapy in treatment failure, although 29 (10%) indicated they would not send in a specimen but would prescribe for Case 3—‘treatment failure in an older woman’.

Two hundred and eighty-six (98%) of the GPs stated that laboratory results ‘always or often’ influenced their change in therapy when resistance was reported to the initial agent. However, 56 (19%) would not send in a specimen but would prescribe for Case 4—‘recurrent UTI in a male diabetic’.


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Transparency declarations
 References
 
For the purpose of surveillance of antibiotic-resistant bacteria, it would be ideal if GPs could request urine specimens for all patients presenting with suspected UTI. Cost, organizational factors and potentially unnecessary burden on patients makes this approach unfeasible. However, understanding the sampling variation between GPs will aid interpreting the applicability of routinely collected data to everyday practice.

We presented six clinical scenarios to a stratified random sample of GPs, and found that nearly all of the GPs indicated they would request a specimen for scenarios representing a probable UTI in a female child (Case 1) and a probable asymptomatic UTI in pregnancy (Case 5). Children do have slightly higher antibiotic resistance rates than young adults1 but it is unlikely that pregnant women have higher rates. There was some variation between the GPs about sampling in a situation of treatment failure in an older woman (Case 3) and recurrent UTI in a male diabetic, with 90% and 81%, respectively, indicating they would request a specimen for these scenarios. However, the high percentage suggests a low probability of surveillance data being biased regarding this type of patient. The greatest variation was in relation to the management of a probable uncomplicated UTI (Case 2), and early patient symptoms with pressure to prescribe (Case 6), with 56% and 33% of GPs respectively indicating they would request a specimen. If this were extrapolated to surveillance data it is likely therefore that samples from patients with suspected uncomplicated UTIs are underrepresented as might be expected taking into account current practice of empirical treatment of uncomplicated UTIs.

A systematic review by Hummers-Pradier et al.4 concluded that for most people managed in general practice with uncomplicated UTI, urine culture was not necessary. They based this conclusion on Grade B evidence (clinical trial with non-randomized controlled design). However, a prospective cohort study of 160 patients consulting their GP with symptoms of UTI in eight general practices in Avon, UK, concluded that antibiotic treatment decisions should not be based on UTI symptoms alone.5 The authors proposed that current clinical practice resulted in a large proportion of patients receiving unnecessary antibiotic treatment.

It is worth noting that over half of the GPs responding to our questionnaire indicated they would request specimens for the young woman with suspected uncomplicated UTI (Case 2). This is probably the most common presentation of suspected UTI although good evidence on consultation rates by clinical subgroups is lacking. A sentinel practice study in England and Wales found that the mean weekly incidence of patients consulting with UTI in the community was ~65 per 100 000 patients.3

Fifty-three per cent of GPs indicated that they would send urine specimens for at least five of these clinical scenarios which would provide a good basis for surveillance. Eighty-eight per cent of GPs had a moderate or low threshold for requesting urine microbiology analysis, indicating they would do so for at least four of the clinical cases.

Interestingly GPs estimates of the percentage of their own patients with clinically suspected UTI from whom they would request a specimen were more broadly distributed than the response to the clinical cases suggested. It is therefore helpful to ask specific clinical questions to focus response as opposed to asking respondents about their impressions of overall sampling behaviour. GPs estimates of the proportion of patients for whom they requested a urine specimen varied considerably within practices and only 38 (13%) GPs stated that they had a practice policy.

Of interest was GPs perceptions of the influence of laboratory results on their choice of antibiotics. Many GPs stated that laboratory results influenced their choice of antibiotic, but that they often treated patients without requesting a urine specimen. Presumably, the laboratory results for patients for whom they did request a sample influenced their choice of antibiotic for those patients who were managed empirically.

We were disappointed that we received responses from only 49.5% of practices. Indeed, this response rate was reached after initially phoning the GP practices to recruit them, offering a financial incentive and following up with additional telephone calls. Reassuringly, however, the age and gender of responding GPs were broadly similar to all GPs in Wales, and we did receive responses from 20% of GP practices in Wales.

Acquiescence bias, where respondents provide answers which they assume the questioner will approve of, may have led to an overestimation of specimen submissions. However, clinical scenarios were used to help to combat this bias and to focus the responder to their usual clinical practice.

We were not able to consider actual presentation rates of suspected UTI at the level of practices. It is possible that some practice populations are more likely to self-treat while others will present to their GPs, and some practice populations will have a higher incidence of UTI than others. Different rates of actual presentation may have influenced GPs responses.

There is a limited research on GPs decision to request urine specimens for culture and susceptibility. A Norwegian study presented a single case history of a 65-year-old woman with possible UTI to 1171 GPs and received 909 responses.6 Eighty-six per cent of those indicating they would request further investigations (n = 895) stated they would request a bacteriological culture.

Ludlam and colleagues compared the susceptibilities of uropathogens isolated from routine urine samples submitted by 80 general practices served by a single laboratory to susceptibilities submitted by five sentinel practices in the same locality that had agreed to submit urine samples on all patients presenting with a clinical diagnosis of UTI during a 3 month period.7 Resistance rates to cefalexin, norfloxacin and gentamicin were marginally lower and resistance to ampicillin, trimethoprim and co-amoxiclav was slightly higher in isolates from the sentinel practices. However, only the resistance rate for co-amoxiclav was significantly different between the two groups of isolates.

Richards undertook a sentinel study where clinicians from three practices were asked to submit all urine specimens from patients who they intended to treat for UTI over a 12 month period.8 He found that resistance to ampicillin and trimethoprim was lower during systematic sampling, although only trimethoprim reached statistical significance. However, whether this difference reached clinical significance is uncertain. Baerheim et al.9 reported statistically significantly lower resistance to all antibiotics tested in consecutive female patients with suspected lower UTI presenting in general practice compared with samples that were routinely sent to the laboratory. Once again the clinical significance of this difference is uncertain. Of more concern is interpreting the data and understanding the source of data. Batchelor et al.10 highlighted the problem when guideline changes to submission of urine specimens resulted in apparent large changes in resistance rate. Guidelines were disseminated advocating submitting urine specimens only for recurrent, relapsing and complicated UTIs which was followed by a 6.9% increase in ampicillin resistance and a 34.4% increase in trimethoprim resistance.

In summary, our study shows that when GPs were presented with clinical scenarios, there was excellent agreement in reported sampling practice for some scenarios (e.g. a child with suspected UTI and a pregnant woman with suspected UTIs). The responses were also in broad agreement on the sampling of recurrent UTI and treatment failure. There was most disagreement for the scenario representing uncomplicated UTI. Since this is the commonest type of UTI managed in general practice, it is likely that there is a systematic bias in surveillance data based on routinely collected data, with samples from cases of uncomplicated UTI being under represented, potentially leading to an overestimation of true resistance rates.

Currently, we are not able to link accurate clinical information to microbiology results for large numbers of patients managed in routine care in general practice. If this could be done, then antibiotic susceptibility data could be more meaningfully interpreted. Attempting this, perhaps initially in sentinel practices, could improve our understanding and interpretation of the antibiotic susceptibility data.


   Transparency declarations
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Transparency declarations
References
 
None to declare.


   Acknowledgements
 
Many thanks to Dr John Magee for his input into establishing the project. We would like to thank the general practitioners and practice managers who took part in the study. This study was funded by a Wales office for research in health and social care grant (Project R00/1/027) as part of a larger study on the link between antibiotic prescribing and resistance in the community.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Transparency declarations
 References
 
1 Howard AJ, Magee JT, Fitzgerald KA, et al. (2001) Factors associated with antibiotic resistance in coliform organisms from community urinary tract infection in Wales. J Antimicrob Chemother 47:305–13.[Abstract/Free Full Text]

2 Oppenheim BA. (2002) Surveillance of antibiotic resistance in urinary tract infections—are we misleading prescribers? Commun Dis Public Health 5:181–2.[Medline]

3 Fleming DM, Ross AM, Cross KW, et al. (2003) The reducing incidence of respiratory tract infection and its relation to antibiotic prescribing. Br J Gen Pract 53:778–83.[Web of Science][Medline]

4 Hummers-Pradier R and Kochen MM. (2002) Urinary tract infections in adult general practice patients. Br J Gen Pract 52:752–61.[Web of Science][Medline]

5 Fahey T, Webb E, Montgomery AA, et al. (2003) Clinical management of urinary tract infection in women: a prospective cohort study. Fam Pract 20:1–6.[Abstract/Free Full Text]

6 Bolann BJ, Thue G, Sandberg S. (2004) Urine analysis and decision-making on cystitis in general practice. Scan J Clin Lab Invest 64:729–36.[CrossRef][Web of Science][Medline]

7 Ludlam H, Sule O, Knapton M, et al. (2004) Can routine laboratory data guide empirical prescribing? J Clin Pathol 57:672.[Free Full Text]

8 Richards J. (2002) Monitoring antibiotic resistance in urinary isolates from the community—a spotter practice model. Commun Dis Public Health 5:226–9.[Medline]

9 Baerheim A, Digranes Ø, Hunskaar S. (1999) Are resistance patterns in uropathogens published by microbiological laboratories valid for general practice? APMIS 107:676–80.[Web of Science][Medline]

10 Batchelor BIF, Crook DWM, Jones T, et al. (2002) Impact of guidelines for the diagnosis of urinary tract infection on trimethoprim susceptibility of Escherichia coli. J Antimicrob Chemother 49:223.[Free Full Text]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J Antimicrob ChemotherHome page
D. De Backer, T. Christiaens, S. Heytens, A. De Sutter, E. E. Stobberingh, and G. Verschraegen
Evolution of bacterial susceptibility pattern of Escherichia coli in uncomplicated urinary tract infections in a country with high antibiotic consumption: a comparison of two surveys with a 10 year interval
J. Antimicrob. Chemother., August 1, 2008; 62(2): 364 - 368.
[Abstract] [Full Text] [PDF]

Home page
 Antimicrob. Agents Chemother.Home page
R. Colgan, J. R. Johnson, M. Kuskowski, and K. Gupta
Risk Factors for Trimethoprim-Sulfamethoxazole Resistance in Patients with Acute Uncomplicated Cystitis
Antimicrob. Agents Chemother., March 1, 2008; 52(3): 846 - 851.
[Abstract] [Full Text] [PDF]

Home page
 BMJHome page
D. Mangin and L. Toop
Urinary tract infection in primary care
BMJ, March 24, 2007; 334(7594): 597 - 598.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
58/6/1303    most recent
dkl432v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hillier, S.
Right arrow Articles by Butler, C. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hillier, S.
Right arrow Articles by Butler, C. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?