Claire Keogh

Validation of the CHADS2 clinical prediction rule to predict ischaemic stroke. A systematic review and meta-analysis.

The CHADS2 predicts annual risk of ischaemic stroke in non-valvular atrial fibrillation. This systematic review and meta-analysis aims to determine the predictive value of CHADS2. The literature was systematically searched from 2001 to October 2010. Data was pooled and analysed using discrimination and calibration statistical measures, using a random effects model. Eight data sets (n = 2815) were included. The diagnostic accuracy suggested a cut-point of ≥ 1 has higher sensitivity (92%) than specificity (12%) and a cut-point of ≥ 4 has higher specificity (96%) than sensitivity (33%). Lower summary estimates were observed for cut-points ≥ 2 (sensitivity 79%, specificity 42%) and ≥ 3 (specificity 77%, sensitivity 50%). There was insufficient data to analyse cut-points ≥ 5 or ≥ 6. Moderate pooled c statistic values were identified for the classic (0.63, 95% CI 0.52-0.75) and revised (0.60, 95% CI 0.43-0.72) view of stratification of the CHADS2. Calibration analysis indicated no significant difference between the predicted and observed strokes across the three risk strata for the classic or revised view. All results were associated with high heterogeneity, and conclusions should be made cautiously. In conclusion, the pooled c statistic and calibration analysis suggests minimal clinical utility of both the classic and revised view of the CHADS2 in predicting ischaemic stroke across all risk strata. Due to high heterogeneity across studies and low event rates across all risk strata, the results should be interpreted cautiously. Further validation of CHADS2 should perhaps be undertaken, given the methodological differences between many of the available validation studies and the original CHADS2 derivation study.

Developing an International Register of Clinical Prediction Rules for Use in Primary Care: A Descriptive Analysis

PURPOSE We describe the methodology used to create a register of clinical prediction rules relevant to primary care. We also summarize the rules included in the register according to various characteristics. METHODS To identify relevant articles, we searched the MEDLINE database (PubMed) for the years 1980 to 2009 and supplemented the results with searches of secondary sources (books on clinical prediction rules) and personal resources (eg, experts in the field). The rules described in relevant articles were classified according to their clinical domain, the stage of development, and the clinical setting in which they were studied. RESULTS Our search identified clinical prediction rules reported between 1965 and 2009. The largest share of rules (37.2%) were retrieved from PubMed. The number of published rules increased substantially over the study decades. We included 745 articles in the register; many contained more than 1 clinical prediction rule study (eg, both a derivation study and a validation study), resulting in 989 individual studies. In all, 434 unique rules had gone through derivation; however, only 54.8% had been validated and merely 2.8% had undergone analysis of their impact on either the process or outcome of clinical care. The rules most commonly pertained to cardiovascular disease, respiratory, and musculoskeletal conditions. They had most often been studied in the primary care or emergency department settings. CONCLUSIONS Many clinical prediction rules have been derived, but only about half have been validated and few have been assessed for clinical impact. This lack of thorough evaluation for many rules makes it difficult to retrieve and identify those that are ready for use at the point of patient care. We plan to develop an international web-based register of clinical prediction rules and computer-based clinical decision support systems.

Clinical prediction rules in practice: review of clinical guidelines and survey of GPs

BACKGROUND The publication of clinical prediction rules (CPRs) studies has risen significantly. It is unclear if this reflects increasing usage of these tools in clinical practice or how this may vary across clinical areas. AIM To review clinical guidelines in selected areas and survey GPs in order to explore CPR usefulness in the opinion of experts and use at the point of care. DESIGN AND SETTING A review of clinical guidelines and survey of UK GPs. METHOD Clinical guidelines in eight clinical domains with published CPRs were reviewed for recommendations to use CPRs including primary prevention of cardiovascular disease, transient ischaemic attack (TIA) and stroke, diabetes mellitus, fracture risk assessment in osteoporosis, lower limb fractures, breast cancer, depression, and acute infections in childhood. An online survey of 401 UK GPs was also conducted. RESULTS Guideline review: Of 7637 records screened by title and/or abstract, 243 clinical guidelines met inclusion criteria. CPRs were most commonly recommended in guidelines regarding primary prevention of cardiovascular disease (67%) and depression (67%). There was little consensus across various clinical guidelines as to which CPR to use preferentially. SURVEY Of 401 responders to the GP survey, most were aware of and applied named CPRs in the clinical areas of cardiovascular disease and depression. The commonest reasons for using CPRs were to guide management and conform to local policy requirements. CONCLUSION GPs use CPRs to guide management but also to comply with local policy requirements. Future research could focus on which clinical areas clinicians would most benefit from CPRs and promoting the use of robust, externally validated CPRs.

Optimized retrieval of primary care clinical prediction rules from MEDLINE to establish a web-based register

OBJECTIVES Identifying clinical prediction rules (CPRs) for primary care from electronic databases is difficult. This study aims to identify a search filter to optimize retrieval of these to establish a register of CPRs for the Cochrane Primary Health Care field. STUDY DESIGN AND SETTING Thirty primary care journals were manually searched for CPRs. This was compared with electronic search filters using alternative methodologies: (1) textword searching; (2) proximity searching; (3) inclusion terms using specific phrases and truncation; (4) exclusion terms; and (5) combinations of methodologies. RESULTS We manually searched 6,344 articles, revealing 41 CPRs. Across the 45 search filters, sensitivities ranged from 12% to 98%, whereas specificities ranged from 43% to 100%. There was generally a trade-off between the sensitivity and specificity of each filter (i.e., the number of CPRs and total number of articles retrieved). Combining textword searching with the inclusion terms (using specific phrases) resulted in the highest sensitivity (98%) but lower specificity (59%) than other methods. The associated precision (2%) and accuracy (60%) were also low. CONCLUSION The novel use of combining textword searching with inclusion terms was considered the most appropriate for updating a register of primary care CPRs where sensitivity has to be optimized.

Impact analysis studies of clinical prediction rules relevant to primary care: a systematic review.

Objectives Following appropriate validation, clinical prediction rules (CPRs) should undergo impact analysis to evaluate their effect on patient care. The aim of this systematic review is to narratively review and critically appraise CPR impact analysis studies relevant to primary care. Setting Primary care. Participants Adults and children. Intervention Studies that implemented the CPR compared to usual care were included. Study design Randomised controlled trial (RCT), controlled before–after, and interrupted time series. Primary outcome Physician behaviour and/or patient outcomes. Results A total of 18 studies, incorporating 14 unique CPRs, were included. The main study design was RCT (n=13). Overall, 10 studies reported an improvement in primary outcome with CPR implementation. Of 6 musculoskeletal studies, 5 were effective in altering targeted physician behaviour in ordering imaging for patients presenting with ankle, knee and neck musculoskeletal injuries. Of 6 cardiovascular studies, 4 implemented cardiovascular risk scores, and 3 reported no impact on physician behaviour outcomes, such as prescribing and referral, or patient outcomes, such as reduction in serum lipid levels. 2 studies examined CPRs in decision-making for patients presenting with chest pain and reduced inappropriate admissions. Of 5 respiratory studies, 2 were effective in reducing antibiotic prescribing for sore throat following CPR implementation. Overall, study methodological quality was often unclear due to incomplete reporting. Conclusions Despite increasing interest in developing and validating CPRs relevant to primary care, relatively few have gone through impact analysis. To date, research has focused on a small number of CPRs across few clinical domains only.

Quantifying patient preferences for symptomatic breast clinic referral: a decision analysis study

Objectives Decision analysis study that incorporates patient preferences and probability estimates to investigate the impact of women’s preferences for referral or an alternative strategy of watchful waiting if faced with symptoms that could be due to breast cancer. Setting Community-based study. Participants Asymptomatic women aged 30–60 years. Interventions Participants were presented with 11 health scenarios that represent the possible consequences of symptomatic breast problems. Participants were asked the risk of death that they were willing to take in order to avoid the health scenario using the standard gamble utility method. This process was repeated for all 11 health scenarios. Formal decision analysis for the preferred individual decision was then estimated for each participant. Primary outcome measure The preferred diagnostic strategy was either watchful waiting or referral to a breast clinic. Sensitivity analysis was used to examine how each varied according to changes in the probabilities of the health scenarios. Results A total of 35 participants completed the interviews, with a median age 41 years (IQR 35–47 years). The majority of the study sample was employed (n=32, 91.4%), with a third-level (university) education (n=32, 91.4%) and with knowledge of someone with breast cancer (n=30, 85.7%). When individual preferences were accounted for, 25 (71.4%) patients preferred watchful waiting to referral for triple assessment as their preferred initial diagnostic strategy. Sensitivity analysis shows that referral for triple assessment becomes the dominant strategy at the upper probability estimate (18%) of breast cancer in the community. Conclusions Watchful waiting is an acceptable strategy for most women who present to their general practitioner (GP) with breast symptoms. These findings suggest that current referral guidelines should take more explicit account of women’s preferences in relation to their GPs initial management strategy.