Hierarchical Random-effects Model Research Articles

COMPARISON OF TRANSRADIAL AND FEMORAL APPROACHES FOR PERCUTANEOUS CORONARY INTERVENTIONS: A HIERARCHICAL BAYESIAN META-ANALYSIS

BACKGROUND: Transradial approach (TRA) for percutaneous coronary interventions (PCI) has been performed since 1994. Despite lower risks of access-site related complications compared to the femoral approach (FA), the clinical benefit of TRA is uncertain. We undertook a systematic review and meta-analysis comparing TRA and FA for PCI. METHODS: A literature search aimed at identifying all clinical trials and observational studies comparing TRA with FA for PCI with reports of clinical outcomes was performed. Odds ratios (OR) were estimated by a hierarchical Bayesian randomeffects model with prespecified stratification for observational and randomized designs. The primary outcomes examined were rates of death, combined incidence of death or myocardial infarction (MI), bleeding and transfusions, early ( 30 days) and late after PCI. RESULTS: We collected data from 73 studies (14 randomized, 59 observational) involving a total of 1,022,123 patients (prior to RIVAL Trial). There was a strong association between TRA and mortality early after intervention (OR 0.52, 95% credible interval (CrI) 0.43-0.62), although the effect was mainly due to observational studies (OR 0.49, 95% CrI 0.39-0.59), with an OR 0.74 (95% CrI 0.38-1.40) in randomized trials. An association between TRA and death or MI was observed over shortterm follow-up (OR 0.66, 95% CrI 0.51-0.82), with an OR of 0.72 (95% CrI 0.41-1.14) for long-term follow-up. Compared with FA, TRA was associated with a major reduction in bleeding (OR 0.22, 95% CrI 0.16-0.29) and in transfusions (OR 0.20, 95% CrI 0.10-0.31). These findings were consistent in both randomized and observational studies. CONCLUSIONS: Our results suggest that PCI performed by TRA is associated with lower risks of death, death or MI, bleeding and transfusions when compared to FA. Our results suggest that greater adoption of TRA has the potential to substantially improve outcomes among patients undergoing PCI. This meta-analysis will be updated with RIVAL Trial results for presentation at the meeting. Fonds de la Recherche en Sante du Quebec

Response to Letter Regarding Article, “Drug-Eluting or Bare Metal Stents for the Treatment of Saphenous Vein Graft Disease: A Bayesian Meta-Analysis”

HomeCirculation: Cardiovascular InterventionsVol. 4, No. 2Response to Letter Regarding Article, “Drug-Eluting or Bare Metal Stents for the Treatment of Saphenous Vein Graft Disease: A Bayesian Meta-Analysis” Free AccessReplyPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessReplyPDF/EPUBResponse to Letter Regarding Article, “Drug-Eluting or Bare Metal Stents for the Treatment of Saphenous Vein Graft Disease: A Bayesian Meta-Analysis” Jean-Michel Paradis, MD, Olivier F. Bertrand, MD, PhD, Robert DeLarochellière, MD, Jean-Pierre Déry, MD, MHS, Éric Larose, DVM, MD, Josep Rodés-Cabau, MD and Stéphane Rinfret, MD, SM Patrick Bélisle, MSc and Lawrence Joseph, PhD Jean-Michel ParadisJean-Michel Paradis Search for more papers by this author , Olivier F. BertrandOlivier F. Bertrand Search for more papers by this author , Robert DeLarochellièreRobert DeLarochellière Search for more papers by this author , Jean-Pierre DéryJean-Pierre Déry Search for more papers by this author , Éric LaroseÉric Larose Search for more papers by this author , Josep Rodés-CabauJosep Rodés-Cabau Search for more papers by this author and Stéphane RinfretStéphane Rinfret Search for more papers by this author Patrick BélislePatrick Bélisle Search for more papers by this author and Lawrence JosephLawrence Joseph Search for more papers by this author Originally published1 Apr 2011https://doi.org/10.1161/CIRCINTERVENTIONS.111.961581Circulation: Cardiovascular Interventions. 2011;4:e15As pointed out by Hakeem et al, other meta-analyses have been published recently on the use of drug-eluting versus bare metal stents in saphenous vein grafts. Because the data on which they are founded are available to researchers worldwide, meta-analyses are certainly more prone to redundancy than original research. However, we still believe that our meta-analysis differed slightly from other reviews because it was the first to use a bayesian hierarchical random-effects model. Moreover, at the time of our literature search, the number of patients collected in our meta-analysis was the highest ever reported. Finally, our work included new data from contemporary studies, including, for example, the STENT registry.1We agree with Hakeem et al on the fact that there is a need for a facility that will assume a prospective registration of protocols for systematic reviews and meta-analyses of studies evaluating health care interventions. This kind of centralized open registry would consequently provide more transparency throughout the publication process, from protocol redaction to revelation of potential bias. This would also avoid superfluous duplication. In fact, a simple, free, web-based, easily searchable platform would be beneficial to inform medical researcher on ongoing reviews and meta-analyses. Thus, we support initiatives like the PRISMA statement,2 which will probably lead to a more efficient use of the finite research funding and to an improvement of the overall quality of systematic reviews and meta-analyses.Jean-Michel Paradis, MDOlivier F. Bertrand, MD, PhDRobert DeLarochellière, MDJean-Pierre Déry, MD, MHSÉric Larose, DVM, MDJosep Rodés-Cabau, MDStéphane Rinfret, MD, SM Institut universitaire de cardiologie et de pneumologie de Québec (Quebec Heart and Lung Institute) Quebec City, CanadaPatrick Bélisle, MScLawrence Joseph, PhD McGill University Health Center Montreal, CanadaDisclosuresNone.

Changes of occupational class differences in physical functioning: a panel study among employees (2000–2007)

BackgroundPanel studies on changes of occupational class differences in health have given varying results. The aim of this study was to examine changes of occupational class differences in physical functioning...

Longitudinal Association of HIV Conspiracy Beliefs with Sexual Risk Among Black Males Living with HIV

Research is needed to identify culturally relevant factors that may contribute to sexual risk among African Americans. We investigated HIV-specific medical mistrust as one such cultural factor, often exhibited as conspiracy beliefs about HIV (e.g., "AIDS was produced in a government laboratory"), which may be indicative of general suspicion of HIV treatment and prevention messages. Over a 6-month time-period, we measured endorsement of HIV conspiracy beliefs three times and frequency of condom use monthly among 181 HIV-positive African American males. A hierarchical multivariate repeated-measures logistic random effects model indicated that greater belief in HIV conspiracies was associated with a higher likelihood of reporting unprotected intercourse across all time-points. An average of 54% of participants who endorsed conspiracies reported unprotected intercourse, versus 39% who did not endorse conspiracies. Secondary prevention interventions may need to address medical mistrust as a contributor to sexual risk among African Americans living with HIV.

The effect of load magnitude on three-dimensional patellar kinematics in vivo

Studies of three-dimensional patellar kinematics done with little or no applied load may not accurately reflect kinematics at physiological load levels, and may provide different results to those acquired with greater applied loads or in physiologic weightbearing. We report the effect of load magnitude on three-dimensional patellar kinematics (flexion, spin and tilt; proximal, lateral and anterior translation) using a validated, sequential static, MRI-based method. Ten healthy subjects loaded their study knee to 0% (no load), 15% and 30% bodyweight (BW) using a custom designed loading rig. Differences between loading levels were determined as a function of knee flexion for each kinematic parameter using linear hierarchical random-effects models. Quadratic and random slope terms were included in the models when significant. We found that the patellae flexed less with knee flexion at 30% BW load compared to 0% BW load ( p<0.001) and 15% BW ( p=0.004) load. The patellae showed a slight medial tilt with knee flexion at 30% BW load which was significantly less than the medial tilt seen at 0% BW load ( p=0.017) and 15% BW load ( p=0.043) with knee flexion. Small but statistically significant differences were also observed for proximal and anterior translation; the patellae were in a more proximal and posterior position at 30% BW load than at 0% BW load ( p=0.010 and p=0.005, respectively) and 15% BW load ( p<0.001 and p=0.029, respectively). Since differences in three-dimensional patellar kinematics were observed between loading levels, magnitudes of prescribed loads must be considered when designing studies and comparing results between studies.

Effect of Physical Inactivity on Cognitive Performance after 2.5 Years of Follow-Up

Objective: To examine the association between physical activity and cognitive performance in a longitudinal study. Methods: We analyzed data from 17.333 noninstitutionalized persons aged 50 years or older in 11 European countries who participated in Wave 1 (2004/2005) and Wave 2 (2006/2007) of the Survey of Heath, Ageing, and Retirement in Europe (SHARE). Physical activity at baseline was measured as self-reported frequency of sports or activities requiring vigorous activity, and frequency of activities demanding a moderate level of activity. Cognitive function was measured at baseline and after a mean of 2.5 years of follow-up by delayed word recall and verbal fluency tests. The effects of physical activity at baseline on cognitive performance at follow-up were assessed in hierarchical multilevel random effects models adjusted for sociodemographic variables (age, education), somatic comorbidities, functional impairment (basic and instrumental activities of daily living, maximum grip strength), depressive symptoms, and body mass index. Results: After adjusting for potential confounders, we found physical inactivity (neither moderate nor vigorous) to be associated with a higher rate of cognitive decline over a mean follow-up of 2.5 years (β = –1.79 (SE = 0.17) for verbal fluency; β = –0.35 (SE = 0.04) for delayed word recall). Further analyses showed that vigorous activities more than once a week were especially related to change in cognition over time. Conclusion: Engagement in moderate and vigorous physical activities protects against cognitive decline in older age. Participation in physical activities may be of particular importance when other risk factors for cognitive decline are present.

Estimation of Bayes Factors in a Class of Hierarchical Random Effects Models Using a Geometrically Ergodic MCMC Algorithm

We consider a Ba yesian random effects model that is commonly used in meta-analysis, in which the random effects have a t distribution, with degrees of freedom parameter to be estimated. We develop a Markov chain Monte Carlo algorithm for estimating the posterior distribution in this model, and establish geometric convergence of the algorithm. The geometric convergence rate has important theoretical and practical ramifications. Indeed, it implies that, under standard moment conditions, the ergodic averages used to estimate posterior quantities of interest satisfy central limit theorems. Moreover, it guarantees the consistency of a batch means estimate of the asymptotic variance in the CLT, which in turn allows for the construction of asymptotically valid standard errors. We show how our Markov chain can be used, in conjunction with an importance sampling method, to carry out an empirical Bayes approach for estimating the degrees of freedom parameter. To illustrate our methodology we consider a meta-analysis of studies that link intake of nonsteroidal anti-inflammatory drugs to a reduction in colon cancer risk, in which some of the studies are outliers. To model the distribution of the study effects we consider the family of t distributions, as well as a family of mixtures of Dirichlet process priors centered at the t distributions, and show how our methodology can be used to make a choice of model. Supplemental materials are available online.

An optical coherence tomography study of a biodegradable vs. durable polymer-coated limus-eluting stent: a LEADERS trial sub-study

Incomplete endothelialization has been found to be associated with late stent thrombosis, a rare but devastating phenomenon, more frequent after drug-eluting stent implantation. Optical coherence tomography (OCT) has 10 times greater resolution than intravascular ultrasound and thus appears to be a valuable modality for the assessment of stent strut coverage. The LEADERS trial was a multi-centre, randomized comparison of a biolimus-eluting stent (BES) with biodegradable polymer with a sirolimus-eluting stent (SES) using a durable polymer. This study sought to evaluate tissue coverage and apposition of stents using OCT in a group of patients from the randomized LEADERS trial. Fifty-six consecutive patients underwent OCT during angiographic follow-up at 9 months. OCT images were acquired using a non-occlusive technique at a pullback speed of 3 mm/s. Data were analysed using a Bayesian hierarchical random-effects model, which accounted for the correlation of lesion characteristics within patients and implicitly assigned analytical weights to each lesion depending on the number of struts observed per lesion. Primary outcome was the difference in percentage of uncovered struts between BESs and SESs. Twenty patients were included in the analysis in the BES group (29 lesions with 4592 struts) and 26 patients in the SES group (35 lesions with 6476 struts). A total of 83 struts were uncovered in the BES group and 407 out of 6476 struts were uncovered in the SES group [weighted difference -1.4%, 95% confidence interval (CI) -3.7 to 0.0, P = 0.04]. Results were similar after adjustment for pre-procedure lesion length, reference vessel diameter, number of implanted study stents, and presence of stent overlap. There were three lesions in the BES group and 15 lesions in the SES group that had > or =5% of all struts uncovered (difference -33.1%, 95% CI -61.7 to -10.3, P < 0.01). Strut coverage at an average follow-up of 9 months appears to be more complete in patients allocated to BESs when compared with SESs. The impact of this difference on clinical outcome and, in particular, on the risk of late stent thrombosis is yet to be determined.

Bayesian Analysis of Evidence from Studies of Warfarin v Aspirin for Symptomatic Intracranial Stenosis

Bayesian analyses of symptomatic intracranial stenosis studies were conducted to compare the benefits of long-term therapy with warfarin to aspirin. The synthesis of evidence of effect from previous nonrandomized studies in monitoring a randomized clinical trial was of particular interest. Sequential Bayesian learning analysis was conducted and Bayesian hierarchical random effects models were used to incorporate variability between studies. The posterior point estimates for the risk rate ratio (RRR) were similar between analyses, although the interval estimates resulting from the hierarchical analyses are larger than the corresponding Bayesian learning analyses. This demonstrated the difference between these methods in accounting for between-study variability. This study suggests that Bayesian synthesis can be a useful supplement to futility analysis in the process of monitoring randomized clinical trials.

Improving phylogenetic analyses by incorporating additional information from genetic sequence databases

Statistical analyses of phylogenetic data culminate in uncertain estimates of underlying model parameters. Lack of additional data hinders the ability to reduce this uncertainty, as the original phylogenetic dataset is often complete, containing the entire gene or genome information available for the given set of taxa. Informative priors in a Bayesian analysis can reduce posterior uncertainty; however, publicly available phylogenetic software specifies vague priors for model parameters by default. We build objective and informative priors using hierarchical random effect models that combine additional datasets whose parameters are not of direct interest but are similar to the analysis of interest. We propose principled statistical methods that permit more precise parameter estimates in phylogenetic analyses by creating informative priors for parameters of interest. Using additional sequence datasets from our lab or public databases, we construct a fully Bayesian semiparametric hierarchical model to combine datasets. A dynamic iteratively reweighted Markov chain Monte Carlo algorithm conveniently recycles posterior samples from the individual analyses. We demonstrate the value of our approach by examining the insertion-deletion (indel) process in the enolase gene across the Tree of Life using the phylogenetic software BALI-PHY; we incorporate prior information about indels from 82 curated alignments downloaded from the BAliBASE database.

Improving the performance of predictive process modeling for large datasets

Advances in Geographical Information Systems (GIS) and Global Positioning Systems (GPS) enable accurate geocoding of locations where scientific data are collected. This has encouraged collection of large spatial datasets in many fields and has generated considerable interest in statistical modeling for location-referenced spatial data. The setting where the number of locations yielding observations is too large to fit the desired hierarchical spatial random effects models using Markov chain Monte Carlo methods is considered. This problem is exacerbated in spatial-temporal and multivariate settings where many observations occur at each location. The recently proposed predictive process, motivated by kriging ideas, aims to maintain the richness of desired hierarchical spatial modeling specifications in the presence of large datasets. A shortcoming of the original formulation of the predictive process is that it induces a positive bias in the non-spatial error term of the models. A modified predictive process is proposed to address this problem. The predictive process approach is knot-based leading to questions regarding knot design. An algorithm is designed to achieve approximately optimal spatial placement of knots. Detailed illustrations of the modified predictive process using multivariate spatial regression with both a simulated and a real dataset are offered.

Winning the Battle Against Childhood Physical Inactivity: The Key to Bone Strength?

Winning the Battle Against Childhood Physical Inactivity: The Key to Bone Strength?

Variations in hospitalisation costs for acute myocardial infarction – a comparison across Europe

The aim of this study was to determine whether between-country variations in hospital costs are larger than within-country variations and, furthermore, to explore reasons for this variability. For this purpose, we chose the primary treatment of patients with acute myocardial infarction (AMI) as an episode of care. We obtained hospitalisation costs and reimbursement rates from 45 hospitals in nine different EU member states (i.e. Denmark, England, France, Germany, Hungary, Italy, Netherlands, Poland, and Spain) for the year 2005. To further analyse the variations in hospital costs, we employed a hierarchical random effects model based on treatment and hospital characteristics and using purchasing power parities (PPPs) as a proxy for country-specific price levels. The between-country standard error was estimated at 2473 euros, whereas the within-country standard error was estimated at 1242 euros. Our regression analysis showed that percutaneous coronary intervention was associated with significantly increased hospitals costs compared to other treatment strategies. We were able to distinguish between three groups of countries with different cost levels based on the number of hospitals that were able to provide these services (i.e. percutaneous transluminal coronary angioplasty (PTCA) with intracoronary stenting). Excluding Hungary, Poland, and Spain, where none of the participating hospitals were able to provide these procedures, the between-country standard error decreased to 1632 euros, whereas the within-country standard error increased to 1416 euros. Finally, we observed exogenous price-level effects between countries and within countries for hospitals located in urban areas.

Hierarchical random effect models for coastal erosion of cliffs in the Holderness coast

Prediction of possible cliff erosion at some future date is fundamental to coastal planning and shoreline management, for example to avoid development in vulnerable areas. Historically, to predict cliff recession rates deterministic methods were used. More recently, recession predictions have been expressed in probabilistic terms. However, to date, only simplistic models have been developed. We consider the cliff erosion along the Holderness Coast. Since 1951 a monitoring program has been started in 118 stations along the coast, providing an invaluable, but often missing, source of information. We build hierarchical random effect models, taking account of the known dynamics of the process and including the missing information.

Statistical Performance Evaluation of Biometric Authentication Systems Using Random Effects Models

As biometric authentication systems become more prevalent, it is becoming increasingly important to evaluate their performance. This paper introduces a novel statistical method of performance evaluation for these systems. Given a database of authentication results from an existing system, the method uses a hierarchical random effects model, along with Bayesian inference techniques yielding posterior predictive distributions, to predict performance in terms of error rates using various explanatory variables. By incorporating explanatory variables as well as random effects, the method allows for prediction of error rates when the authentication system is applied to potentially larger and/or different groups of subjects than those originally documented in the database. We also extend the model to allow for prediction of the probability of a false alarm on a "watch-list" as a function of the list size. We consider application of our methodology to three different face authentication systems: a filter-based system, a Gaussian Mixture Model (GMM)-based system, and a system based on frequency domain representation of facial asymmetry.

Reproducibility of Renal Function Measurements in Adult Men with Diabetic Nephropathy: Research and Clinical Implications

Background/Aim: Measurement of the renal function is critical to follow progression of kidney disease. Short-term and long-term variabilities in these measurements have significant impacts on clinical decision making and clinical trials. The goal of this study was to describe the variability in these measurements and to calculate minimum sample size estimates over varying time frames for clinical trials. Methods: We studied 44 elderly men with diabetic nephropathy who participated in a clinical trial. Glomerular filtration rate and renal plasma flow were measured by continuous infusion technique with five urine collection periods on two occasions 4 months apart. Protein and creatinine excretion rates were measured in the same specimens. In addition, two consecutive 24-hour specimens every month for 4 months were collected to analyze urine protein, creatinine, urea nitrogen, and electrolytes. A hierarchical random effects model was used to analyze the reproducibility from hour to hour, from day to day, and from month to month. Results: A total of 824 urine specimens were analyzed, of which 412 constituted specimens collected in the short term and 412 were 24-hour urine collections. Hour-to-hour variation accounted for 45% for urinary clearance of iothalamate, but for only 0.5% of the variability in plasma clearance of iothalamate. Day-to-day variability in 24-hour urinary excretion rates for creatinine was 46% and for protein 10%. Month-to-month variability in 24-hour excretion rates for creatinine was 11% and for protein 19%. The urine protein/creatinine ratio had a day-to-day variability of 2% and a month-to-month variability of 19%. Sample size requirements can be reduced by correcting for urine creatinine for some but not all urinary analytes. Conclusions: In nephrotic men with diabetic nephropathy, the coefficient of variation in the month-to-month protein excretion rate is 36%. Approximately 28 patients in each arm of two groups are needed to detect a difference in protein excretion rate of 28% (1 g/day in this study). The coefficient of variation in plasma iothalamate clearance over 4 months is 16%. To detect a 10% change in glomerular filtration rate between two groups, 44 patients per group are needed. To be deemed statistically significant, a change in daily protein excretion rate of at least 72% over month(s) is needed in individual patients.

Automatic approximation of the marginal likelihood in non-Gaussian hierarchical models

Fitting of non-Gaussian hierarchical random effects models by approximate maximum likelihood can be made automatic to the same extent that Bayesian model fitting can be automated by the program BUGS. The word “automatic” means that the technical details of computation are made transparent to the user. This is achieved by combining a technique from computer science known as “automatic differentiation” with the Laplace approximation for calculating the marginal likelihood. Automatic differentiation, which should not be confused with symbolic differentiation, is mostly unknown to statisticians, and hence basic ideas and results are reviewed. The computational performance of the approach is compared to that of existing mixed-model software on a suite of datasets selected from the mixed-model literature.

Efficacy of inactivated hepatitis A vaccine in HIV-infected patients: a hierarchical bayesian meta-analysis

Patients with human immunodeficiency virus-1 (HIV) have elevated risk for hepatitis A virus (HAV) coinfection. Sequelae from coinfection include increased risk of liver-related morbidity and mortality. This study synthesizes the results of trials measuring response to HAV vaccine in HIV-infected patients using Bayesian meta-analysis methodologies. PubMed, OhioLINK/Medline, and other sources were used to search for studies analyzing response to HAV vaccine in HIV-infected patients. Studies were evaluated for quality. A Bayesian hierarchical random-effects model was used to estimate overall response to vaccine. Between-study variance was calculated. Sensitivity analyses were conducted to determine the potential for bias. The literature search yielded eight studies for inclusion in the meta-analysis, with a combined total of 458 patients. Observed proportions of response in individual studies ranged from 50 to 95%. The intention-to-treat meta-analysis estimated a combined proportion of HIV+ patients responding to vaccine of 64% (95% CI 52–75%). Heterogeneity between studies was significant. The overall response rate to HAV vaccine in HIV-infected patients is lower than typically cited. Up to 1/2 of HIV-infected patients may be nonresponders. Future research will be required to better understand the correlates of response.

Detecting Small-Area Similarities in the Epidemiology of Childhood Acute Lymphoblastic Leukemia and Diabetes Mellitus, Type 1: A Bayesian Approach

Childhood acute lymphoblastic leukemia and diabetes mellitus, type 1, have common epidemiologic and etiologic features, including correlated international incidence and associations with infections. The authors examined whether the diseases' similar large-scale distributions are reflected in small geographic areas while also examining the influence of sociodemographic characteristics. Details of 299 children (0-14 years) with acute lymphoblastic leukemia and 1,551 children with diabetes diagnosed between 1986 and 1998 were extracted from two registers in Yorkshire, United Kingdom. Standardized incidence ratios across 532 electoral wards were compared using Poisson regression, confirming significant associations between population mixing and the geographic heterogeneity of both conditions. Bayesian methods analysis of spatial correlation between diseases by modeling a bivariate outcome based on their standardized incidence ratios was applied; spatial and heterogeneity components were included within a hierarchical random effects model. A positive correlation between diseases of 0.33 (95% credible interval: -0.20, 0.74) was observed, and this was reduced after control for population mixing (r = 0.18), population density (r = 0.14), and deprivation (r = 0.06). The Bayesian approach showed a modest but nonsignificant joint spatial correlation between diseases, only partially suggesting that the risk of both was associated within some electoral wards. With Bayesian methodology, population mixing remained significantly associated with both diseases. The links between diabetes and acute lymphoblastic leukemia observed for large regions are weaker for small areas. More powerful replications are needed for confirmation of these findings.

Assessing the impacts of grazing levels on bird density in woodland habitat: a Bayesian approach using expert opinion

AbstractMany studies on birds focus on the collection of data through an experimental design, suitable for investigation in a classical analysis of variance (ANOVA) framework. Although many findings are confirmed by one or more experts, expert information is rarely used in conjunction with the survey data to enhance the explanatory and predictive power of the model.We explore this neglected aspect of ecological modelling through a study on Australian woodland birds, focusing on the potential impact of different intensities of commercial cattle grazing on bird density in woodland habitat.We examine a number of Bayesian hierarchical random effects models, which cater for overdispersion and a high frequency of zeros in the data using WinBUGS and explore the variation between and within different grazing regimes and species. The impact and value of expert information is investigated through the inclusion of priors that reflect the experience of 20 experts in the field of bird responses to disturbance.Results indicate that expert information moderates the survey data, especially in situations where there are little or no data. When experts agreed, credible intervals for predictions were tightened considerably. When experts failed to agree, results were similar to those evaluated in the absence of expert information. Overall, we found that without expert opinion our knowledge was quite weak. The fact that the survey data is quite consistent, in general, with expert opinion shows that we do know something about birds and grazing and we could learn a lot faster if we used this approach more in ecology, where data are scarce. Copyright © 2005 John Wiley &amp; Sons, Ltd.