This week in the BMJ
Volume 330, Number 7501, Issue of 21 May 2005Try the new optimal strategy for searching Medline
Try the new optimal strategy for searching Medline
Using Medline to retrieve high quality clinical studies on prevention or treatment of health disorders will be easier and better with improved search strategies. Haynes and colleagues (p 1179) tested the sensitivity, specificity, precision, and accuracy of strategies developed from 4862 unique terms in 18 404 combinations, attempting to retrieve quality studies on prevention or treatment published in 2000 in 161 clinical journals indexed in Medline. No one search strategy will perform perfectly, but clinicians can use the most specific search to retrieve a few sound articles, and researchers can use the most sensitive search to be comprehensive.
|
![[To top]](/icons/back.gif){kind=icon}
Trial outcomes are applicable to all comparable patients
Patients who participate in randomised controlled trials do no better and no worse than patients receiving the same or similar treatment outside trials. In a systematic review, Vist and colleagues (p 1175) analysed comparisons of outcomes in five randomised controlled trials and 50 cohort studies that provided data for 31 140 patients treated within trials and 20 380 comparable patients treated outside trials. Overall, neither dichotomous nor continuous outcomes showed statistically significant differences. The results of randomised controlled trials are therefore applicable to comparable patients in usual clinical practice, say the authors.
Hip x rays help to predict worsening of osteoarthritis
A radiograph has strong additional value in predicting progression of osteoarthritis. In their population based cohort study, Reijman and colleagues (p 1183) included 1904 participants aged 55 and older who at baseline had radiographic osteoarthritic signs of grade 1 or greater as defined by Kellgren and Lawrence. Multivariate logistic regression models showed that the Kellgren and Lawrence score was the best predictor of progression, but x ray films were a strong additional predictor (odds ratio 5.8), especially for patients with hip pain at baseline (odds ratio 24.3).
|
1 in 8 Indian women report chronic fatigue
Chronic fatigue syndrome is fairly common among Indian women. A population based survey of 3000 randomly selected women aged 18-50 by Patel and colleagues (p 1190) showed that more than 12% of women report chronic fatigue, and the most prominent risk factors are older age, socioeconomic deprivation, gender disadvantage (marital sexual violence in particular), and symptoms of common mental disorders. Almost 20% of women who refused to participate said that the reason was that a family member had not given permission.
|
RCTs can't properly assess effectiveness of acupuncture
Randomised controlled trials (RCTs) may not be the best research design to evaluate complex non-pharmacological treatments such as acupuncture, say Paterson and Dieppe on page 1202. These trials were developed to test new drugs and are based on biomedical assumptions, while disregarding elements such as talking or listening as incidental (placebo) factors. Since incidental factors cannot always be separated from the specific effects of non-pharmacological treatments such as acupuncture, they should be taken into account in evaluating the efficacy of these treatments, say the authors.
|
Prevalence of asthma in Finland hasn't stabilised
Although the prevalence of atopic diseases has recently been levelling off in some European countries after several decades of increasing, this is not the case in Finland. Latvala and colleagues (p 1186) examined the trends for asthma, allergic rhinitis, and atopic eczema in almost 1.4 million Finnish men aged 18-19. Between 1966 and 2003 the prevalence of asthma increased 12-fold (from 0.29% to 3.45%) and showed a continuous rising trend throughout the period. In the past 13 years, though, asthma seems to have become milder and better controlled.
What's new
Keep updated
Services
Tools
Online poll
Resources
Print issues
