Development of clinical guidelines
1999; Elsevier BV; Volume: 353; Issue: 9150 Linguagem: Inglês
10.1016/s0140-6736(05)74990-4
ISSN1474-547X
Autores Tópico(s)Cervical Cancer and HPV Research
ResumoIn their Dec 12 commentary, Suzanne and Robert Fletcher1Fletcher SW Fletcher RH Development of clinical guidelines.Lancet. 1998; 352: 1876Summary Full Text Full Text PDF PubMed Scopus (17) Google Scholar note the difficulty of monitoring clinicians' use of guidelines. Their point is illustrated by the difficulties we encountered in an inconclusive randomised controlled trial of the introduction of guidelines for the management of cervical Chlamydia trachomatis infection in general practices, in UK inner cities.In Spring, 1994, 28 practices in south London were recruited and put into matched pairs on the basis of number of principles (≤2, >2), Jarman underprivileged area rating (≥25, < 25), and cervical smear target reached (< 80%, ≥80%). One practice in each pair was randomly allocated to receive practicebased education and national guidelines, 2Centers for Disease Control.Recommendations for the prevention and management of Chlamydia trachomatis infections.MMWR. 1993; 42: 1-39PubMed Google Scholar modified for local use, on the management of women with chlamydial infection. Compliance with the guidelines was assessed under masked conditions from photocopies of the relevant section of general practitioners' medical records of all women diagnosed with chlamydial infection between May, 1994, and October, 1995. Outcome measures were: recognised antibiotic regimen, partner notification, and referral to a genitourinary clinic (GUM) for follow-up. We also assessed actual clinic attendance.The percentage of women in intervention and control practices who were managed according to the guidelines, and the differences in percentages were as follows: recognised antibiotic regimen 78 versus 56, difference 21 (95% CI −17 to 60); advised partner needs treatment 59 versus 40, difference 19 (−17 to 55); referred to GUM for follow-up 33 versus 16, difference 18 (−16 to 52); attended GUM 44 versus 25 (unable to analyse by practice because of patient's confidentiality). The estimated intracluster correlation coefficient for recognised antibiotic regimen was 0·29.Although there was a roughly 40% improvement in management in intervention practices, it was not a significant increase. There are several possible reasons for this finding. The main one was a much lower than expected prevalence of chlamydia infection,3Oakeshott P Kerry S Hay S Hay P Opportunistic screening for chlamydial infection at time of cervical smear testing in general practice: prevalence study.BMJ. 1998; 316: 351-352Crossref PubMed Scopus (22) Google Scholar partly due to the low sensitivity of the test used-enzyme immunoassay.4Grun L Tassano-Smith J Carder C et al.Comparison of two methods of screening for genital chlamydial infection in women attending in general practice: cross sectional survey.BMJ. 1997; 315: 226-230Crossref PubMed Scopus (61) Google Scholar In addition, this population is highly mobile and 12% of women positive for chlamydia were lost to follow-up. Thus, data on only 59 patients from eight pairs of practices were available for our analysis. Finally, management in control practices was much better than predicted from the pilot studies, probably as a result of the Hawthorne effect.Process measures are important, but what matters most is whether compliance with guidelines actually improves patients' health. This point is generally ignored. In our study, a harder outcome would have been the prevalence of chlamydia infection at test of cure 6 months after diagnosis. Allowing for clustering and attrition, we estimate that this would require screening 7000 women aged 16–24 years in 50 general practices with a more sensitive non-invasive test, such as ligase chain reaction on first-pass urines. Such a study would cost about £250 000, compared with £35 000 for the one we report here. The Fletchers correctly emphasise the importance of evaluating implementation of guidelines, especially in general practice.Our study was funded by South Thames R and D Project Grant Scheme. In their Dec 12 commentary, Suzanne and Robert Fletcher1Fletcher SW Fletcher RH Development of clinical guidelines.Lancet. 1998; 352: 1876Summary Full Text Full Text PDF PubMed Scopus (17) Google Scholar note the difficulty of monitoring clinicians' use of guidelines. Their point is illustrated by the difficulties we encountered in an inconclusive randomised controlled trial of the introduction of guidelines for the management of cervical Chlamydia trachomatis infection in general practices, in UK inner cities. In Spring, 1994, 28 practices in south London were recruited and put into matched pairs on the basis of number of principles (≤2, >2), Jarman underprivileged area rating (≥25, < 25), and cervical smear target reached (< 80%, ≥80%). One practice in each pair was randomly allocated to receive practicebased education and national guidelines, 2Centers for Disease Control.Recommendations for the prevention and management of Chlamydia trachomatis infections.MMWR. 1993; 42: 1-39PubMed Google Scholar modified for local use, on the management of women with chlamydial infection. Compliance with the guidelines was assessed under masked conditions from photocopies of the relevant section of general practitioners' medical records of all women diagnosed with chlamydial infection between May, 1994, and October, 1995. Outcome measures were: recognised antibiotic regimen, partner notification, and referral to a genitourinary clinic (GUM) for follow-up. We also assessed actual clinic attendance. The percentage of women in intervention and control practices who were managed according to the guidelines, and the differences in percentages were as follows: recognised antibiotic regimen 78 versus 56, difference 21 (95% CI −17 to 60); advised partner needs treatment 59 versus 40, difference 19 (−17 to 55); referred to GUM for follow-up 33 versus 16, difference 18 (−16 to 52); attended GUM 44 versus 25 (unable to analyse by practice because of patient's confidentiality). The estimated intracluster correlation coefficient for recognised antibiotic regimen was 0·29. Although there was a roughly 40% improvement in management in intervention practices, it was not a significant increase. There are several possible reasons for this finding. The main one was a much lower than expected prevalence of chlamydia infection,3Oakeshott P Kerry S Hay S Hay P Opportunistic screening for chlamydial infection at time of cervical smear testing in general practice: prevalence study.BMJ. 1998; 316: 351-352Crossref PubMed Scopus (22) Google Scholar partly due to the low sensitivity of the test used-enzyme immunoassay.4Grun L Tassano-Smith J Carder C et al.Comparison of two methods of screening for genital chlamydial infection in women attending in general practice: cross sectional survey.BMJ. 1997; 315: 226-230Crossref PubMed Scopus (61) Google Scholar In addition, this population is highly mobile and 12% of women positive for chlamydia were lost to follow-up. Thus, data on only 59 patients from eight pairs of practices were available for our analysis. Finally, management in control practices was much better than predicted from the pilot studies, probably as a result of the Hawthorne effect. Process measures are important, but what matters most is whether compliance with guidelines actually improves patients' health. This point is generally ignored. In our study, a harder outcome would have been the prevalence of chlamydia infection at test of cure 6 months after diagnosis. Allowing for clustering and attrition, we estimate that this would require screening 7000 women aged 16–24 years in 50 general practices with a more sensitive non-invasive test, such as ligase chain reaction on first-pass urines. Such a study would cost about £250 000, compared with £35 000 for the one we report here. The Fletchers correctly emphasise the importance of evaluating implementation of guidelines, especially in general practice. Our study was funded by South Thames R and D Project Grant Scheme.
Referência(s)