Hayley E Jones

The Prevalence of Non-Alcoholic Fatty Liver Disease in Children and Adolescents: A Systematic Review and Meta-Analysis

Background & Aims Narrative reviews of paediatric NAFLD quote prevalences in the general population that range from 9% to 37%; however, no systematic review of the prevalence of NAFLD in children/adolescents has been conducted. We aimed to estimate prevalence of non-alcoholic fatty liver disease (NAFLD) in young people and to determine whether this varies by BMI category, gender, age, diagnostic method, geographical region and study sample size. Methods We conducted a systematic review and meta-analysis of all studies reporting a prevalence of NAFLD based on any diagnostic method in participants 1–19 years old, regardless of whether assessing NAFLD prevalence was the main aim of the study. Results The pooled mean prevalence of NAFLD in children from general population studies was 7.6% (95%CI: 5.5% to 10.3%) and 34.2% (95% CI: 27.8% to 41.2%) in studies based on child obesity clinics. In both populations there was marked heterogeneity between studies (I2 = 98%). There was evidence that prevalence was generally higher in males compared with females and increased incrementally with greater BMI. There was evidence for differences between regions in clinical population studies, with estimated prevalence being highest in Asia. There was no evidence that prevalence changed over time. Prevalence estimates in studies of children/adolescents attending obesity clinics and in obese children/adolescents from the general population were substantially lower when elevated alanine aminotransferase (ALT) was used to assess NAFLD compared with biopsies, ultrasound scan (USS) or magnetic resonance imaging (MRI). Conclusions Our review suggests the prevalence of NAFLD in young people is high, particularly in those who are obese and in males.

The Prevalence of Non-Alcoholic Fatty Liver Disease in Children and Adolescents

Background & Aims Narrative reviews of paediatric NAFLD quote prevalences in the general population that range from 9% to 37%; however, no systematic review of the prevalence of NAFLD in children/adolescents has been conducted. We aimed to estimate prevalence of non-alcoholic fatty liver disease (NAFLD) in young people and to determine whether this varies by BMI category, gender, age, diagnostic method, geographical region and study sample size. Methods We conducted a systematic review and meta-analysis of all studies reporting a prevalence of NAFLD based on any diagnostic method in participants 1–19 years old, regardless of whether assessing NAFLD prevalence was the main aim of the study. Results The pooled mean prevalence of NAFLD in children from general population studies was 7.6% (95%CI: 5.5% to 10.3%) and 34.2% (95% CI: 27.8% to 41.2%) in studies based on child obesity clinics. In both populations there was marked heterogeneity between studies (I2 = 98%). There was evidence that prevalence was generally higher in males compared with females and increased incrementally with greater BMI. There was evidence for differences between regions in clinical population studies, with estimated prevalence being highest in Asia. There was no evidence that prevalence changed over time. Prevalence estimates in studies of children/adolescents attending obesity clinics and in obese children/adolescents from the general population were substantially lower when elevated alanine aminotransferase (ALT) was used to assess NAFLD compared with biopsies, ultrasound scan (USS) or magnetic resonance imaging (MRI). Conclusions Our review suggests the prevalence of NAFLD in young people is high, particularly in those who are obese and in males.

The identification of "unusual" health-care providers from a hierarchical model

It has become common to adopt a hierarchical model structure when comparing the performance of multiple health-care providers. This structure allows some variation in such measures, beyond that explained by sampling variation, to be “normal,” in recognition of the fact that risk-adjustment is never perfect. The shrinkage estimates arising from such a model structure also have appealing properties. It is not immediately clear, however, how “unusual” providers, that is, any with particularly high or low rates, can be identified based on such a model. Given that some variation in underlying rates is assumed to be the norm, we argue that it is not generally appropriate to identify a provider as interesting based only on evidence of it lying above or below the population mean. We note with concern, however, that this practice is not uncommon. We examine in detail three possible strategies for identifying unusual providers, carefully distinguishing between statistical “outliers” and “extremes.” A two-level normal model is used for mathematical simplicity, but we note that much of the discussion also applies to alternative data structures. Further, we emphasize throughout that each approach can be viewed as resulting from a Bayesian or a classical perspective. Three worked examples provide additional insight.

Bayesian models for subgroup analysis in clinical trials

Background In a pharmaceutical drug development setting, possible interactions between the treatment and particular baseline clinical or demographic factors are often of interest. However, the subgroup analysis required to investigate such associations remains controversial. Concerns with classical hypothesis testing approaches to the problem include low power, multiple testing, and the possibility of data dredging. Purpose As an alternative to hypothesis testing, the use of shrinkage estimation techniques is investigated in the context of an exploratory post hoc subgroup analysis. A range of models that have been suggested in the literature are reviewed. Building on this, we explore a general modeling strategy, considering various options for shrinkage of effect estimates. This is applied to a case-study, in which evidence was available from seven-phase II–III clinical trials examining a novel therapy, and also to two artificial datasets with the same structure. Methods Emphasis is placed on hierarchical modeling techniques, adopted within a Bayesian framework using freely available software. A range of possible subgroup model structures are applied, each incorporating shrinkage estimation techniques. Results The investigation of the case-study showed little evidence of subgroup effects. Because inferences appeared to be consistent across a range of well-supported models, and model diagnostic checks showed no obvious problems, it seemed this conclusion was robust. It is reassuring that the structured shrinkage techniques appeared to work well in a situation where deeper inspection of the data suggested little evidence of subgroup effects. Limitations The post hoc examination of subgroups should be seen as an exploratory analysis, used to help make better informed decisions regarding potential future studies examining specific subgroups. To a certain extent, the degree of understanding provided by such assessments will be limited by the quality and quantity of available data. Conclusions In light of recent interest by health authorities into the use of subgroup analysis in the context of drug development, it appears that Bayesian approaches involving shrinkage techniques could play an important role in this area. Hopefully, the developments outlined here provide useful methodology for tackling such a problem, in-turn leading to better informed decisions regarding subgroups.

Indications for red blood cell transfusion in cardiac surgery: a systematic review and meta-analysis

BACKGROUND Good blood management is an important determinant of outcome in cardiac surgery. Guidelines recommend restrictive red blood cell transfusion. Our objective was to systematically review the evidence from randomised controlled trials and observational studies that are used to inform transfusion decisions in adult cardiac surgery. METHODS We did a systematic review by searching PubMed, Embase, Cochrane Library, and DARE, from inception to May 1, 2015, databases from specialist societies, and bibliographies of included studies and recent relevant review articles. We included randomised controlled trials that assessed the effect of liberal versus restrictive red blood cell transfusion in patients undergoing cardiac and non-cardiac surgery, and observational studies that assessed the effect of red blood cell transfusion compared with no transfusion on outcomes in adult cardiac patients after surgery. We pooled adjusted odds ratios using fixed-effects and random-effects meta-analyses. The primary outcome was 30-day mortality. FINDINGS We included data from six cardiac surgical randomised controlled trials (3352 patients), 19 non-cardiac surgical trials (8361 patients), and 39 observational studies (232,806 patients). The pooled fixed effects mortality odds ratios comparing liberal versus restrictive transfusion thresholds was 0.70 (95% CI 0.49-1.02; p=0.060) for cardiac surgical trials and 1.10 (95% CI 0.96-1.27; p=0.16) for trials in settings other than cardiac surgery. By contrast, observational cohort studies in cardiac surgery showed that red blood cell transfusion compared with no transfusion was associated with substantially higher mortality (random effects odds ratio 2.72, 95% CI 2.11-3.49; p<0.0001) and other morbidity, although with substantial heterogeneity and small study effects. INTERPRETATION Evidence from randomised controlled trials in cardiac surgery refutes findings from observational studies that liberal thresholds for red blood cell transfusion are associated with a substantially increased risk of mortality and morbidity. Observational studies and trials in non-cardiac surgery should not be used to inform treatment decisions or guidelines for patients having cardiac surgery. FUNDING None.

Recapture or Precapture? Fallibility of Standard Capture-Recapture Methods in the Presence of Referrals Between Sources

Capture-recapture methods, largely developed in ecology, are now commonly used in epidemiology to adjust for incomplete registries and to estimate the size of difficult-to-reach populations such as problem drug users. Overlapping lists of individuals in the target population, taken from administrative data sources, are considered analogous to overlapping “captures” of animals. Log-linear models, incorporating interaction terms to account for dependencies between sources, are used to predict the number of unobserved individuals and, hence, the total population size. A standard assumption to ensure parameter identifiability is that the highest-order interaction term is 0. We demonstrate that, when individuals are referred directly between sources, this assumption will often be violated, and the standard modeling approach may lead to seriously biased estimates. We refer to such individuals as having been “precaptured,” rather than truly recaptured. Although sometimes an alternative identifiable log-linear model could accommodate the referral structure, this will not always be the case. Further, multiple plausible models may fit the data equally well but provide widely varying estimates of the population size. We demonstrate an alternative modeling approach, based on an interpretable parameterization and driven by careful consideration of the relationships between the sources, and we make recommendations for capture-recapture in practice.

Illicit and pharmaceutical drug consumption estimated via wastewater analysis. Part B: Placing back-calculations in a formal statistical framework

Concentrations of metabolites of illicit drugs in sewage water can be measured with great accuracy and precision, thanks to the development of sensitive and robust analytical methods. Based on assumptions about factors including the excretion profile of the parent drug, routes of administration and the number of individuals using the wastewater system, the level of consumption of a drug can be estimated from such measured concentrations. When presenting results from these ‘back-calculations’, the multiple sources of uncertainty are often discussed, but are not usually explicitly taken into account in the estimation process. In this paper we demonstrate how these calculations can be placed in a more formal statistical framework by assuming a distribution for each parameter involved, based on a review of the evidence underpinning it. Using a Monte Carlo simulations approach, it is then straightforward to propagate uncertainty in each parameter through the back-calculations, producing a distribution for instead of a single estimate of daily or average consumption. This can be summarised for example by a median and credible interval. To demonstrate this approach, we estimate cocaine consumption in a large urban UK population, using measured concentrations of two of its metabolites, benzoylecgonine and norbenzoylecgonine. We also demonstrate a more sophisticated analysis, implemented within a Bayesian statistical framework using Markov chain Monte Carlo simulation. Our model allows the two metabolites to simultaneously inform estimates of daily cocaine consumption and explicitly allows for variability between days. After accounting for this variability, the resulting credible interval for average daily consumption is appropriately wider, representing additional uncertainty. We discuss possibilities for extensions to the model, and whether analysis of wastewater samples has potential to contribute to a prevalence model for illicit drug use.

Development of a Transparent Interactive Decision Interrogator to Facilitate the Decision-Making Process in Health Care

Background Decisions about the use of new technologies in health care are often based on complex economic models. Decision makers frequently make informal judgments about evidence, uncertainty, and the assumptions that underpin these models. Objectives Transparent interactive decision interrogator (TIDI) facilitates more formal critique of decision models by decision makers such as members of appraisal committees of the National Institute for Health and Clinical Excellence in the UK. By allowing them to run advanced statistical models under different scenarios in real time, TIDI can make the decision process more efficient and transparent, while avoiding limitations on pre-prepared analysis. Methods TIDI, programmed in Visual Basic for applications within Excel, provides an interface for controlling all components of a decision model developed in the appropriate software (e.g., meta-analysis in WinBUGS and the decision model in R) by linking software packages using RExcel and R2WinBUGS. TIDIs graphical controls allow the user to modify assumptions and to run the decision model, and results are returned to an Excel spreadsheet. A tool displaying tornado plots helps to evaluate the influence of individual parameters on the model outcomes, and an interactive meta-analysis module allows the user to select any combination of available studies, explore the impact of bias adjustment, and view results using forest plots. We demonstrate TIDI using an example of a decision model in antenatal care. Conclusion Use of TIDI during the NICE appraisal of tumor necrosis factor-alpha inhibitors (in psoriatic arthritis) successfully demonstrated its ability to facilitate critiques of the decision models by decision makers.

Risk adjustment of heroin treatment outcomes for comparative performance assessment in England

AIMS Variability in effectiveness of treatment for substance abuse disorder (SUD) is an important and understudied issue. This study aimed to quantify the extent of outcome variability in the English SUD treatment system after adjusting for potential confounding variables. DESIGN Prospective cohort study using data from the English national drug treatment outcome monitoring database. SETTING All 149 administrative areas delivering publicly funded SUD services in the National Health Service and non-governmental sector. PARTICIPANTS New adult admissions between January 2008 and October 2010 with illicit heroin-related problems in all administrative areas, with an in-treatment review conducted between 5 and 26 weeks (mean = 129.5 days; SD = 40.0) up to 30 April 2011 (n = 65 223; 75.6% of eligible clients). Individuals were divided randomly to form model developmental and internal validation samples. These were contrasted with an independent (external) sample of the same population admitted to treatment between November 2010 and April 2011 and followed to 31 October 2011 (n = 13 797; 81.4% of those eligible). MEASUREMENTS AND ANALYSIS The outcome measure was self-reported illicit heroin use, categorized as abstinent or deteriorated (the latter by Reliable Change Index), each risk-adjusted by person-level (demographics, clinical severity and treatment complexity) and area-level (SUD prevalence, social deprivation and severity averages) covariates by multivariable logistic regression using multiply imputed outcome and covariate data. Risk-adjusted models were assessed by information criteria and discrimination (c-index). Standardized outcome rates were compared by funnel plot with 95% and 99% control limits. FINDINGS Models of heroin abstinence (48.4%) and deterioration (3.2%) were comparable across the developmental and validation samples (c-index = 0.70-0.71 and 0.82-0.87), with 79.2 and 94.0%, respectively, of the 149 treatment areas falling within 95% control limits. At the 99% limit, seven areas (4.7%) achieved abstinence rates above the national average, and eight had relatively poor abstinence rates (5.4%). At the 99% control limit, one area achieved very low deterioration outcomes and two (1.3%) were worse that the average. Risk adjustment served to increase abstinence rates in good performing areas by 0.63% and reduce abstinence rates by 0.37% in poor performing areas, and by 0.12% and 0.18%, respectively, for deterioration. CONCLUSION There is some exceptional variability in the apparent effectiveness of the English treatment system for substance use disorders. It is important to determine the source of this variability in order to inform drug treatment delivery and its evaluation both in England and overseas.

Early life influences on the risk of injecting drug use: case control study based on the Edinburgh Addiction Cohort

AIMS To investigate childhood influences on onset of injection drug use. DESIGN Matched case-control study. SETTING Edinburgh, Scotland. PARTICIPANTS A total of 432 individuals presenting at a community health facility with injection drug use and 432 age- and sex-matched non-injecting controls recruited through the same facility. MEASUREMENTS Main exposures considered were family structure and experience of public care, carer substance use, physical and sexual victimization and conduct problems, all measured at personal interview. The outcome was history of adult injection drug use recorded in medical records corroborated at personal interview. FINDINGS Compared to two-parent families all other family structures were associated with increased risk of injection drug use, the greatest increased risk being associated with public care. Violence, criminality and financial problems in the family were also associated with increased risk, as were all types of carer substance use. The greatest increased risk was associated with markers of early conduct problems, particularly school exclusion and childhood contact with the criminal justice system. In multivariable analyses the strongest risk factors for later injecting were always having lived with a relative or family friend (not always a parent) and in care/adopted/foster home at any point [odds ratio (OR) = 2.66, 95% confidence interval (CI): 1.02-6.92 and OR = 2.17, 95% CI: 0.91-5.17, respectively], experienced violence from parent or carer (OR = 2.06, 95% CI: 1.26, 3.38) and early evidence of conduct problems [ever excluded from school (OR = 2.73, 95% CI: 1.68, 4.45); childhood criminality (ever arrested by police pre-adult OR = 3.05, 95% CI: 1.90, 4.89, ever been in borstal/young offenders/list D school OR = 4.70, 95% CI: 2.02, 10.94)]. After adjustment for family structure and conduct problems, sexual victimization was associated weakly with injecting onset (OR = 1.29, 95% CI: 0.76-2.19). More than 70% of injection drug use onset appeared attributable to the risk factors identified. CONCLUSIONS Injection drug use in adults is associated strongly with prior childhood adversity, in particular not living with both parents and early conduct problems. Prevention initiatives should also consider these risk factors.