Cookies on this website
We use cookies to ensure that we give you the best experience on our website. If you click 'Continue' we'll assume that you are happy to receive all cookies and you won't see this message again. Click 'Find out more' for information on how to change your cookie settings.
  • Measurement error and timing of predictor values for multivariable risk prediction models are poorly reported.

    27 June 2018

    OBJECTIVE: Measurement error in predictor variables may threaten the validity of clinical prediction models. We sought to evaluate the possible extent of the problem. A secondary objective was to examine whether predictors are measured at the intended moment of model use. METHODS: A systematic search of Medline was used to identify a sample of articles reporting the development of a clinical prediction model published in 2015. After screening according to a predefined inclusion criteria, information on predictors, strategies to control for measurement error and intended moment of model use were extracted. Susceptibility to measurement error for each predictor was classified into low and high risk. RESULTS: Thirty-three studies were reviewed, including 151 different predictors in the final prediction models. Fifty-one (33.7%) predictors were categorised as high risk of error, however this was not accounted for in the model development. Only 8 (24.2%) studies explicitly stated the intended moment of model use and when the predictors were measured. CONCLUSION: Reporting of measurement error and intended moment of model use is poor in prediction model studies. There is a need to identify circumstances where ignoring measurement error in prediction models is consequential and whether accounting for the error will improve the predictions.

  • Predictors of falls and fractures leading to hospitalization in people with schizophrenia spectrum disorder: A large representative cohort study.

    3 July 2018

    AIM: To investigate predictors of falls/fractures leading to hospitalisation in people with schizophrenia-spectrum disorders. METHODS: A historical cohort of people with schizophrenia-spectrum disorders (ICD F20-29) from 01/2006-12/2012 was assembled using data from the South London and Maudsley NHS Biomedical Research Centre Case Register. Falls/fractures were ascertained from a linkage to national hospitalisation data. Separate multivariate Cox regression analyses were employed to identify predictors of falls and fractures. RESULTS: Of 11,567 people with schizophrenia-spectrum disorders (mean age 42.6 years, 43% female), 579 (incidence rate 12.79 per 1000 person-years) and 528 (11.65 per 1000 person-years) had at least one reported hospital admission due to a fall or fracture respectively and 822 patients had at least either a recorded fall or a fracture during this period (i.e. 7.1% of sample). Overall, 6.69 and 10.74 years of inpatient hospital stay per 1000-person years of follow-up occurred due to a fall and fracture respectively. 14(0.12%) and 28(0.24%) died due to a fall and fracture respectively. In Multivariable analysis, increasing age, white ethnicity, analgesics, cardiovascular disease, hypertension, diseases of the genitourinary system, visual disturbance and syncope were significant risk factor for both falls and fractures. A previous fracture (HR 2.05, 95% CI 1.53-2.73) and osteoporosis (HR 6.79, 95% CI 4.71-9.78) were strong risk factors for consequent fractures. CONCLUSION: Comorbid physical health conditions and analgesic medication prescription were associated with higher risk of falls and fractures. Osteoporosis and previous fracture were strong predictors for subsequent fractures. Interventions targeting bone health and falls/fractures need to be developed and evaluated in these populations.

  • Letter to the editor - round table unites to tackle culture change in an effort to improve animal research reporting.

    3 July 2018

    A round table discussion was held during the LAVA-ESLAV-ECLAM conference on Reproducibility of Animal Studies on the 25th of September 2017 in Edinburgh. The aim of the round table was to discuss how to enhance the rate at which the quality of reporting animal research can be improved. This signed statement acknowledges the efforts that participant organizations have made towards improving the reporting of animal studies and confirms an ongoing commitment to drive further improvements, calling upon both academics and laboratory animal veterinarians to help make this cultural change.

  • Trajectories and predictors of the long-term course of low back pain: cohort study with 5-year follow-up.

    3 July 2018

    Low back pain (LBP) is a major health challenge globally. Research has identified common trajectories of pain over time. We aimed to investigate whether trajectories described in 1 primary care cohort can be confirmed in another, and to determine the prognostic value of factors collected 5 years prior to the identification of the trajectory. The study was conducted on 281 patients who had consulted primary care for LBP, at that point completed a baseline questionnaire, and then returned a questionnaire at 5-year follow-up plus at least 3 (of 6) subsequent monthly questionnaires. Baseline factors were measured using validated tools. Pain intensity scores from the 5-year follow-up and monthly questionnaires were used to assign participants into 4 previously derived pain trajectories (no or occasional mild, persistent mild, fluctuating, and persistent severe), using latent class analysis. Posterior probabilities of belonging to each cluster were estimated for each participant. The posterior probabilities for the assigned clusters were very high (>0.90) for each cluster except for the smallest "fluctuating" cluster (0.74). Lower social class and higher pain intensity were significantly associated with a more severe trajectory 5 years later, as were patients' perceptions of the greater consequences and longer duration of pain, and greater passive behavioural coping. Low back pain trajectories identified previously appear generalizable. These allow better understanding of the long-term course of LBP, and effective management tailored to individual trajectories needs to be identified.This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • Rheumatoid factor testing in Spanish primary care: A population-based cohort study including 4.8 million subjects and almost half a million measurements.

    3 July 2018

    OBJECTIVE: Rheumatoid factor (RF) testing is used in primary care in the diagnosis of rheumatoid arthritis (RA); however a positive RF may occur without RA. Incorrect use of RF testing may lead to increased costs and delayed diagnoses. The aim was to assess the performance of RF as a test for RA and to estimate the costs associated with its use in a primary care setting. MATERIAL AND METHODS: A retrospective cohort study using the Information System for the Development of Research in Primary Care database (contains primary care records and laboratory results of >80% of the Catalonian population, Spain). Participants were patients ≥18 years with ≥1 RF test performed between 01/01/2006 and 31/12/2011, without a pre-existing diagnosis of RA. Outcome measures were an incident diagnosis of RA within 1 year of testing, and the cost of testing per case of RA. RESULTS: 495,434/4,796,498 (10.3%) patients were tested at least once. 107,362 (21.7%) of those tested were sero-positive of which 2768 (2.6%) were diagnosed with RA within 1 year as were 1141/388,072 (0.3%) sero-negative participants. The sensitivity of RF was 70.8% (95% CI 69.4-72.2), specificity 78.7% (78.6-78.8), and positive and negative predictive values 2.6% (2.5-2.7) and 99.7% (99.6-99.7) respectively. Approximately €3,963,472 was spent, with a cost of €1432 per true positive case. CONCLUSIONS: Although 10% of patients were tested for RF, most did not have RA. Limiting testing to patients with a higher pre-test probability would significantly reduce the cost of testing.