Search

We investigate the behavior of the Lasso for selecting invalid instruments in linear instrumental variables models for estimating causal effects of exposures on outcomes, as proposed recently by Kang et al. Invalid instruments are such that they fail the exclusion restriction and enter the model as explanatory variables. We show that for this setup, the Lasso may not consistently select the invalid instruments if these are relatively strong. We propose a median estimator that is consistent when less than 50% of the instruments are invalid, and its consistency does not depend on the relative strength of the instruments, or their correlation structure. We show that this estimator can be used for adaptive Lasso estimation, with the resulting estimator having oracle properties. The methods are applied to a Mendelian randomization study to estimate the causal effect of body mass index (BMI) on diastolic blood pressure, using data on individuals from the UK Biobank, with 96 single nucleotide polymorphisms as potential instruments for BMI. Supplementary materials for this article are available online., (© 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.)

Aims: To investigate whether smokers prescribed varenicline had lower risks of serious ill-health during the 4 years following treatment compared with those prescribed nicotine replacement therapy (NRT)., Design: Observational cohort study of electronic medical records., Setting: A total of 370 UK general practices sampled from the Clinical Practice Research Datalink., Participants: A total of 126 718 patients aged 18 and over who were issued smoking cessation prescriptions between 1 September 2006 and 31 March 2014., Measurements: Our primary outcome was all-cause mortality within 2 years of first prescription as indicated by linked Office of National Statistics data. Our secondary outcomes were cause-specific mortality, all-cause, cause-specific hospitalization, primary care diagnosis of myocardial infarction or chronic obstructive pulmonary disease (COPD), body mass index and attendance rate to primary care within 2 years of first prescription. Risk differences and 95% confidence intervals were estimated by multivariable adjusted regression and propensity score matched regression. We used instrumental variable analysis to overcome residual confounding., Findings: People prescribed varenicline were healthier at baseline than those prescribed NRT in almost all characteristics, highlighting the potential for residual confounding. Our instrumental variable analysis results found that people prescribed varenicline had a similar risk of mortality at 2 years [risk difference per 100 patients treated = 0.67, 95% confidence interval (CI) = -0.11 to 1.46)] to those prescribed NRT, and there were similar rates of all-cause hospitalization, incident primary-care diagnoses of myocardial infarction and COPD. People prescribed varenicline subsequently attended primary care less frequently., Conclusions: Smokers prescribed varenicline in primary care in the United Kingdom do not appear to be less likely to die, be hospitalized or experience a myocardial infarction or chronic obstructive pulmonary disease during the following 2 years compared with smokers prescribed nicotine replacement therapy, but they gain more weight and attend primary care less frequently., (© 2018 The Authors. Addiction published by John Wiley & Sons Ltd on behalf of Society for the Study of Addiction.)

Competing Interests: Conflicts of interest We report no conflicts of interest.

Background: There is limited evidence about the effectiveness of varenicline and nicotine replacement therapy (NRT) for long-term smoking cessation in primary care, or whether the treatment effectiveness differs by socioeconomic position (SEP). Therefore, we estimated the long-term effectiveness of varenicline versus NRT (> 2 years) on smoking cessation, and investigated whether effectiveness differs by SEP., Methods: This is a prospective cohort study of electronic medical records from 654 general practices in England, within the Clinical Practice Research Datalink, using three different analytical methods: multivariable logistic regression, propensity score matching and instrumental variable analyses. Exposure was prescription of varenicline versus NRT, and the primary outcome was smoking cessation at 2 years' follow-up; outcome was also assessed at 3, 6, and 9 months, and at 1 and 4 years after exposure. SEP was defined using the Index of Multiple Deprivation., Results: At 2 years, 28.8% (N = 20 362/70 610) of participants prescribed varenicline and 24.3% (N = 36 268/149 526) of those prescribed NRT quit; adjusted odds ratio was 1.26 [95% confidence interval (CI): 1.23 to 1.29], P < 0.0001. The association persisted for up to 4 years and was consistent across all analyses. We found little evidence that the effectiveness of varenicline differed greatly by SEP. However, patients from areas of higher deprivation were less likely to be prescribed varenicline; adjusted odds ratio was 0.91 (95% CI: 0.90 to 0.92), P < 0.0001., Conclusions: Patients prescribed varenicline were more likely to be abstinent up to 4 years after first prescription than those prescribed NRT. In combination with other evidence, the results from this study may be used to update clinical guidelines on the use of varenicline for smoking cessation., (© The Author 2017. Published by Oxford University Press on behalf of the International Epidemiological Association)

There is increasing interest in the use of instrumental variable analysis to overcome unmeasured confounding in observational pharmacoepidemiological studies. This is partly because instrumental variable analyses are potentially less biased than conventional regression analyses. However, instrumental variable analyses are less precise, and regulators and clinicians find it difficult to interpret conflicting evidence from instrumental variable compared with conventional regression analyses. In this paper, we describe three techniques to assess which approach (instrumental variable versus conventional regression analyses) is least biased. These techniques are negative control outcomes, negative control populations and tests of covariate balance. We illustrate these methods using an analysis of the effects of smoking cessation therapies (varenicline) prescribed in primary care., (© The Author 2017. Published by Oxford University Press on behalf of the International Epidemiological Association.)

Background: Instrumental variable analysis, for example with physicians' prescribing preferences as an instrument for medications issued in primary care, is an increasingly popular method in the field of pharmacoepidemiology. Existing power calculators for studies using instrumental variable analysis, such as Mendelian randomization power calculators, do not allow for the structure of research questions in this field. This is because the analysis in pharmacoepidemiology will typically have stronger instruments and detect larger causal effects than in other fields. Consequently, there is a need for dedicated power calculators for pharmacoepidemiological research., Methods and Results: The formula for calculating the power of a study using instrumental variable analysis in the context of pharmacoepidemiology is derived before being validated by a simulation study. The formula is applicable for studies using a single binary instrument to analyse the causal effect of a binary exposure on a continuous outcome. An online calculator, as well as packages in both R and Stata, are provided for the implementation of the formula by others., Conclusions: The statistical power of instrumental variable analysis in pharmacoepidemiological studies to detect a clinically meaningful treatment effect is an important consideration. Research questions in this field have distinct structures that must be accounted for when calculating power. The formula presented differs from existing instrumental variable power formulae due to its parametrization, which is designed specifically for ease of use by pharmacoepidemiologists., (© The Author 2017. Published by Oxford University Press on behalf of the International Epidemiological Association)

The Two-Sample Two-Stage Least Squares (TS2SLS) data combination estimator is a popular estimator for the parameters in linear models when not all variables are observed jointly in one single data set. Although the limiting normal distribution has been established, the asymptotic variance formula has only been stated explicitly in the literature for the case of conditional homoskedasticity. By using the fact that the TS2SLS estimator is a function of reduced form and first-stage OLS estimators, we derive the variance of the limiting normal distribution under conditional heteroskedasticity. A robust variance estimator is obtained, which generalises to cases with more general patterns of variable (non-)availability. Stata code and some Monte Carlo results are provided in an Appendix. Stata code for a nonlinear GMM estimator that is identical to the TS2SLS estimator in just identified models and asymptotically equivalent to the TS2SLS estimator in overidentified models is also provided there.

Despite condoms being cheap and effective in preventing HIV, there remains an 8billion shortfall in condom use in risky sex-acts. Social marketing organisations apply private sector marketing approaches to sell public health products. This paper investigates the impact of marketing tools, including promotion and pricing, on demand for male and female condoms in 52 countries between 1997 and 2009. A static model differentiates drivers of demand between products, while a dynamic panel data estimator estimates their short- and long-run impacts. Products are not equally affected: female condoms are not affected by advertising, but highly affected by interpersonal communication and HIV prevalence. Price and promotion have significant short- and long-run effects, with female condoms far more sensitive to price than male condoms. The design of optimal distribution strategies for new and existing HIV prevention technologies must consider both product and target population characteristics., (Copyright © 2016 The Authors. Published by Elsevier B.V. All rights reserved.)

We consider testing for weak instruments in a model with multiple endogenous variables. Unlike Stock and Yogo (2005), who considered a weak instruments problem where the rank of the matrix of reduced form parameters is near zero, here we consider a weak instruments problem of a near rank reduction of one in the matrix of reduced form parameters. For example, in a two-variable model, we consider weak instrument asymptotics of the form [Formula: see text] where [Formula: see text] and [Formula: see text] are the parameters in the two reduced-form equations, [Formula: see text] is a vector of constants and [Formula: see text] is the sample size. We investigate the use of a conditional first-stage [Formula: see text]-statistic along the lines of the proposal by Angrist and Pischke (2009) and show that, unless [Formula: see text], the variance in the denominator of their [Formula: see text]-statistic needs to be adjusted in order to get a correct asymptotic distribution when testing the hypothesis [Formula: see text]. We show that a corrected conditional [Formula: see text]-statistic is equivalent to the Cragg and Donald (1993) minimum eigenvalue rank test statistic, and is informative about the maximum total relative bias of the 2SLS estimator and the Wald tests size distortions. When [Formula: see text] in the two-variable model, or when there are more than two endogenous variables, further information over and above the Cragg-Donald statistic can be obtained about the nature of the weak instrument problem by computing the conditional first-stage [Formula: see text]-statistics.

The use of genetic markers as instrumental variables (IV) is receiving increasing attention from economists, statisticians, epidemiologists and social scientists. Although IV is commonly used in economics, the appropriate conditions for the use of genetic variants as instruments have not been well defined. The increasing availability of biomedical data, however, makes understanding of these conditions crucial to the successful use of genotypes as instruments. We combine the econometric IV literature with that from genetic epidemiology, and discuss the biological conditions and IV assumptions within the statistical potential outcomes framework. We review this in the context of two illustrative applications., (Copyright © 2015 The Authors. Published by Elsevier B.V. All rights reserved.)

We investigated the role of common genetic variation in educational attainment and household income. We used data from 5,458 participants of the National Child Development Study to estimate: 1) the associations of rs9320913, rs11584700 and rs4851266 and socioeconomic position and educational phenotypes; and 2) the univariate chip-heritability of each phenotype, and the genetic correlation between each phenotype and educational attainment at age 16. The three SNPs were associated with most measures of educational attainment. Common genetic variation contributed to 6 of 14 socioeconomic background phenotypes, and 17 of 29 educational phenotypes. We found evidence of genetic correlations between educational attainment at age 16 and 4 of 14 social background and 8 of 28 educational phenotypes. This suggests common genetic variation contributes both to differences in educational attainment and its relationship with other phenotypes. However, we remain cautious that cryptic population structure, assortative mating, and dynastic effects may influence these associations.

Introduction: Smoking is a major avoidable cause of ill-health and premature death. Treatments that help patients successfully quit smoking have an important effect on health and life expectancy. Varenicline is a medication that can help smokers successfully quit smoking. However, there are concerns that it may cause adverse effects, such as increase in the occurrence of depression, self-harm and suicide and cardiovascular disease. In this study we aim to examine the effects of varenicline versus other smoking cessation pharmacotherapies on smoking cessation, health service use, all-cause and cause-specific mortality and physical and mental health conditions., Methods: In this project we will investigate the effects of varenicline compared to nicotine replacement therapies on: (1) long-term smoking cessation and whether these effects differ by area level deprivation; and (2) the following clinically-important outcomes: rate of general practice and hospital attendance; all-cause mortality and death due to diseases of the respiratory system and cardiovascular disease; and a primary care diagnosis of respiratory illness, myocardial infarction or depression and anxiety. The study is based on a cohort of patients prescribed these smoking cessation medications from the Clinical Practice Research Datalink (CPRD). We will use three methods to overcome confounding: multivariable adjusted Cox regression, propensity score matched Cox regression, and instrumental variable regression. The total expected sample size for analysis will be at least 180,000. Follow-up will end with the earliest of either an 'event' or censoring due to the end of registration or death., Ethics and Dissemination: Ethics approval was not required for this study. This project has been approved by the CPRD's Independent Scientific Advisory Committee (ISAC). We will disseminate our findings via publications in international peer-reviewed journals and presentations at international conferences., (Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/)

Instrumental variable estimates of causal effects can be biased when using many instruments that are only weakly associated with the exposure. We describe several techniques to reduce this bias and estimate corrected standard errors. We present our findings using a simulation study and an empirical application. For the latter, we estimate the effect of height on lung function, using genetic variants as instruments for height. Our simulation study demonstrates that, using many weak individual variants, two-stage least squares (2SLS) is biased, whereas the limited information maximum likelihood (LIML) and the continuously updating estimator (CUE) are unbiased and have accurate rejection frequencies when standard errors are corrected for the presence of many weak instruments. Our illustrative empirical example uses data on 3631 children from England. We used 180 genetic variants as instruments and compared conventional ordinary least squares estimates with results for the 2SLS, LIML, and CUE instrumental variable estimators using the individual height variants. We further compare these with instrumental variable estimates using an unweighted or weighted allele score as single instruments. In conclusion, the allele scores and CUE gave consistent estimates of the causal effect. In our empirical example, estimates using the allele score were more efficient. CUE with corrected standard errors, however, provides a useful additional statistical tool in applications with many weak instruments. The CUE may be preferred over an allele score if the population weights for the allele score are unknown or when the causal effects of multiple risk factors are estimated jointly., (© 2014 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.)

The most widely used measure of segregation is the so-called dissimilarity index. It is now well understood that this measure also reflects randomness in the allocation of individuals to units (i.e. it measures deviations from evenness, not deviations from randomness). This leads to potentially large values of the segregation index when unit sizes and/or minority proportions are small, even if there is no underlying systematic segregation. Our response to this is to produce adjustments to the index, based on an underlying statistical model. We specify the assignment problem in a very general way, with differences in conditional assignment probabilities underlying the resulting segregation. From this, we derive a likelihood ratio test for the presence of any systematic segregation, and bias adjustments to the dissimilarity index. We further develop the asymptotic distribution theory for testing hypotheses concerning the magnitude of the segregation index and show that the use of bootstrap methods can improve the size and power properties of test procedures considerably. We illustrate these methods by comparing dissimilarity indices across school districts in England to measure social segregation.

Models of the determinants of individuals' primary care costs can be used to set capitation payments to providers and to test for horizontal equity. We compare the ability of eight measures of patient morbidity and multimorbidity to predict future primary care costs and examine capitation payments based on them. The measures were derived from four morbidity descriptive systems: 17 chronic diseases in the Quality and Outcomes Framework (QOF); 17 chronic diseases in the Charlson scheme; 114 Expanded Diagnosis Clusters (EDCs); and 68 Adjusted Clinical Groups (ACGs). These were applied to patient records of 86,100 individuals in 174 English practices. For a given disease description system, counts of diseases and sets of disease dummy variables had similar explanatory power. The EDC measures performed best followed by the QOF and ACG measures. The Charlson measures had the worst performance but still improved markedly on models containing only age, gender, deprivation and practice effects. Comparisons of predictive power for different morbidity measures were similar for linear and exponential models, but the relative predictive power of the models varied with the morbidity measure. Capitation payments for an individual patient vary considerably with the different morbidity measures included in the cost model. Even for the best fitting model large differences between expected cost and capitation for some types of patient suggest incentives for patient selection. Models with any of the morbidity measures show higher cost for more deprived patients but the positive effect of deprivation on cost was smaller in better fitting models., (Copyright © 2014 The Authors. Published by Elsevier B.V. All rights reserved.)

Objectives: To investigate whether physicians' prescribing preferences were valid instrumental variables for the antidepressant prescriptions they issued to their patients., Study Design and Setting: We investigated whether physicians' previous prescriptions of (1) tricyclic antidepressants (TCAs) vs. selective serotonin reuptake inhibitors (SSRIs) and (2) paroxetine vs. other SSRIs were valid instruments. We investigated whether the instrumental variable assumptions are likely to hold and whether TCAs (vs. SSRIs) were associated with hospital admission for self-harm or death by suicide using both conventional and instrumental variable regressions. The setting for the study was general practices in the United Kingdom., Results: Prior prescriptions were strongly associated with actual prescriptions: physicians who previously prescribed TCAs were 14.9 percentage points (95% confidence interval [CI], 14.4, 15.4) more likely to prescribe TCAs, and those who previously prescribed paroxetine were 27.7 percentage points (95% CI, 26.7, 28.8) more likely to prescribe paroxetine, to their next patient. Physicians' previous prescriptions were less strongly associated with patients' baseline characteristics than actual prescriptions. We found no evidence that the estimated association of TCAs with self-harm/suicide using instrumental variable regression differed from conventional regression estimates (P-value = 0.45)., Conclusion: The main instrumental variable assumptions held, suggesting that physicians' prescribing preferences are valid instruments for evaluating the short-term effects of antidepressants., (Copyright © 2013 The Authors. Published by Elsevier Inc. All rights reserved.)

Aims: The UK Clinical Practice Research Datalink (CPRD) is increasingly being used to investigate suicide-related adverse drug reactions. No studies have comprehensively validated the recording of suicide and nonfatal self-harm in the CPRD. We validated general practitioners' recording of these outcomes using linked Office for National Statistics (ONS) mortality and Hospital Episode Statistics (HES) admission data., Methods: We identified cases of suicide and self-harm recorded using appropriate Read codes in the CPRD between 1998 and 2010 in patients aged ≥15 years. Suicides were defined as patients with Read codes for suicide recorded within 95 days of their death. International Classification of Diseases codes were used to identify suicides/hospital admissions for self-harm in the linked ONS and HES data sets. We compared CPRD-derived cases/incidence of suicide and self-harm with those identified from linked ONS mortality and HES data, national suicide incidence rates and published self-harm incidence data., Results: Only 26.1% (n = 590) of the 'true' (ONS-confirmed) suicides were identified using Read codes. Furthermore, only 55.5% of Read code-identified suicides were confirmed as suicide by the ONS data. Of the HES-identified cases of self-harm, 68.4% were identified in the CPRD using Read codes. The CPRD self-harm rates based on Read codes had similar age and sex distributions to rates observed in self-harm hospital registers, although rates were underestimated in all age groups., Conclusions: The CPRD recording of suicide using Read codes is unreliable, with significant inaccuracy (over- and under-reporting). Future CPRD suicide studies should use linked ONS mortality data. The under-reporting of self-harm appears to be less marked., (© 2012 The Authors. British Journal of Clinical Pharmacology © 2012 The British Pharmacological Society.)

Instrumental variables can be used to estimate the causal effects of exposures on outcomes in the presence of residual or uncontrolled confounding. To assess the validity of analyses using instrumental variables, specific information about whether underlying assumptions are met must be presented, in particular to demonstrate that the instrument is associated with the exposure but not with measured confounding factors. We systematically reviewed the epidemiological literature in Embase and Medline for articles containing the term "instrumental variable$" to investigate whether reporting of test statistics in studies using instrumental variables was sufficient to assess the validity of the results. We extracted the information each study reported about their instrumental variables, including specification tests used to check assumptions. The search found 756 studies of which 90 were relevant and were included. Only 25 (28%) studies reported appropriate tests of the strength of the associations between instruments and exposure. Forty-four (49%) studies reported associations between the instrumental variables and observed covariates. Studies using instrumental variables had wide confidence intervals and so effect estimates were imprecise. We propose a checklist of information and specification tests that studies using instrumental variables should report.

Background: Comorbidity is increasingly common in primary care. The cost implications for patient care and budgetary management are unclear., Aim: To investigate whether caring for patients with specific disease combinations increases or decreases primary care costs compared with treating separate patients with one condition each., Design: Retrospective observational study using data on 86 100 patients in the General Practice Research Database., Method: Annual primary care cost was estimated for each patient including consultations, medication, and investigations. Patients with comorbidity were defined as those with a current diagnosis of more than one chronic condition in the Quality and Outcomes Framework. Multiple regression modelling was used to identify, for three age groups, disease combinations that increase (cost-increasing) or decrease (cost-limiting) cost compared with treating each condition separately., Results: Twenty per cent of patients had at least two chronic conditions. All conditions were found to be both cost-increasing and cost-limiting when co-occurring with other conditions except dementia, which is only cost-limiting. Depression is the most important cost-increasing condition when co-occurring with a range of conditions. Hypertension is cost-limiting, particularly when co-occurring with other cardiovascular conditions., Conclusion: Three categories of comorbidity emerge, those that are: cost-increasing, mainly due to a combination of depression with physical comorbidity; cost-limiting because treatment for the conditions overlap; and cost-limiting for no apparent reason but possibly because of inadequate care. These results can contribute to efficient and effective management of chronic conditions in primary care.

Height has long been recognized as being associated with better outcomes: the question is whether this association is causal. We use children's genetic variants as instrumental variables to deal with possible unobserved confounders and examine the effect of child/adolescent height on a wide range of outcomes: academic performance, IQ, self-esteem, depression symptoms and behavioral problems. OLS findings show that taller children have higher IQ, perform better in school, and are less likely to have behavioral problems. The IV results differ: taller girls (but not boys) have better cognitive performance and, in contrast to the OLS, greater height appears to increase behavioral problems.

The literature that examines the relationship between child or adolescent Body Mass Index (BMI) and academic attainment generally finds mixed results. This may be due to the use of different data sets, conditioning variables, or methodologies: studies either use an individual fixed effects (FE) approach and/or an instrumental variable (IV) specification. Using one common dataset, the Avon Longitudinal Study of Parents and Children, and a common set of controls, this paper compares the different approaches (including using different types of IV's), discusses their appropriateness, and contrasts their findings. We show that, although the results differ depending on the approach, most estimates cannot be statistically distinguished from OLS, nor from each other. Examining the potential violations of key assumptions of the different approaches and comparing their point estimates, we conclude that fat mass is unlikely to be causally related to academic achievement in adolescence., (Copyright © 2012 Elsevier B.V. All rights reserved.)

Background: Weight gain during infancy may programme later health outcomes, but examination of this hypothesis requires appropriate lifecourse methods and detailed weight gain measures during childhood. We examined associations between weight gain in infancy and early childhood and blood pressure at the age of 6.5 years in healthy children born at term., Methods: We carried out an observational analysis of data from a cluster-randomized breastfeeding promotion trial in Belarus. Of 17 046 infants enrolled between June 1996 and December 1997, 13 889 (81.5%) had systolic and diastolic blood pressure measured at 6.5 years; 10 495 children with complete data were analysed. A random-effects linear spline model with three knot points was used to estimate each individual's birthweight and weight gain from birth to 3 months, 3 months to 1 year and 1-5 years. Path analysis was used to separate direct effects from those mediated through subsequent weight gain., Results: In boys, after controlling for confounders and prior weight gain, the change in systolic blood pressure per z-score increase in weight gain was 0.09 mmHg [95% confidence interval (95% CI) -0.14 to 0.31] for birthweight; 0.41 mmHg (95% CI 0.19-0.64) for birth to 3 months; 0.69 mmHg (95% CI 0.47-0.92) for 3 months to 1 year and 0.82 mmHg (95% CI 0.58-1.06) for 1-5 years. Most of the associations between weight gain and blood pressure were mediated through weight at the age of 6.5 years. Findings for girls and diastolic blood pressure were similar., Conclusions: Children who gained weight faster than their peers, particularly at later ages, had higher blood pressure at the age of 6.5 years, with no association between birthweight and blood pressure.

Background: If treatments are used to modify a trait, then patients with high-risk genotypes for the trait should be found at higher frequency in treatment groups than in the general population. The frequency ratio of high- to low-risk genotypes treated should reflect the mean threshold above which the treatment is given in the population. As an example, we hypothesized that because APOE (apolipoprotein E) alleles affect the LDL cholesterol (LDLc) concentration, APOE genotype frequencies in statin takers should act as a proxy for the prevailing treatment threshold of LDLc., Methods: We used LDLc, statin usage, and APOE genotype data from the British Women's Heart and Health Study (n=2289; age, 60-79 years) and calculated the genotype ratio treatment index (GRTI) by dividing the proportion of ε3/ε2 or ε3/ε4 participants prescribed a statin by the proportion of ε3/ε3 participants prescribed a statin, both overall and according to socioeconomic class, geographic region, and coronary heart disease (CHD) status. Genotype-specific LDLc distributions were used to calculate the mean LDLc treatment threshold., Results: For genotype ε3/ε2, the GRTI was 0.52 (95% CI, 0.30-0.87) for statin takers overall, 0.22 (95% CI, 0.00-0.56) for those without CHD, and 0.69 (95% CI, 0.31-1.18) for those with CHD. The GRTIs for those without and with CHD backcalculate to LDLc thresholds of 5.65 mmol/L (95% CI, 5.50-5.82 mmol/L) and 4.39 mmol/L (95% CI, 4.21-4.59 mmol/L), respectively. Scotland and North England showed dissimilar GRTIs, which backcalculated to LDLc thresholds of 5.06 mmol/L (95% CI, 4.83-5.28 mmol/L) and 5.44 mmol/L (95% CI, 5.19-5.69 mmol/L), respectively, for all women., Conclusions: The findings illustrate how genotype frequencies can be a proxy for treatment thresholds used in clinical practice. Genome-wide studies have identified>500 disease-relevant polymorphisms. GRTIs from cost-efficient genotyping, in combination with phenotypic data, may have wide potential in health services research.

Background: Bipolar disorder (BD) is characterized by episodes of mania and depression. The debilitating symptoms during an acute episode require intensive treatment, frequently leading to inpatient psychiatric care, which places significant demands on health and social care systems and incurs substantial costs. However, no study to date has estimated the economic impact of relapse., Objectives: To estimate the direct costs associated with relapse in the treatment of BD following an acute manic or mixed episode over a 21-month follow-up period in routine clinical practice in Europe, using data from a large, prospective, observational study., Methods: EMBLEM was a prospective, observational study on the outcomes of patients with a manic/mixed episode of BD conducted in 14 European countries. Patients eligible for analysis were those enrolled in the 21-month maintenance phase of the study, following the 3-month acute phase. Relapse was defined as achieving any one of the following criteria: (i) at least a one-point increase in Clinical Global Impression - Bipolar Disorder (CGI-BP) overall score from the previous visit, with a final rating of > or =4; (ii) inpatient admission for an acute episode of BD; or (iii) psychiatrists' confirmation of relapse. Data on healthcare resource use were recorded retrospectively for the four respective periods (3-6, 6-12, 12-18 and 18-24 month visits). Multivariate analyses were performed to compare the cost of resource use (inpatient stay, day care, psychiatrist visits and medication) for those who relapsed during the 21-month maintenance phase and those who never relapsed. A sensitivity analysis was also conducted to examine the 6-month costs during relapse. The analyses were adjusted for patient characteristics and took account of non-Normality of the cost data by using a log link function. UK unit costs were applied to resource use. The analysis was repeated after multiple imputation for missing data. All costs were presented as year 2007/08 values., Results: A total of 1379 patients completed all visits during the maintenance phase and were eligible for inclusion in the present analysis. Of these, over half (54.3%) experienced relapse during this period. A total of 792 patients without any missing data were eventually included in the final cost model. Costs incurred by patients who relapsed during the 21-month maintenance phase were approximately double those incurred by patients who never relapsed (pounds sterling 9140 vs pounds sterling 4457; p < 0.05). Of the cost difference, 80.3% was accounted for by inpatient stay. Estimates on the economic impact were higher (pounds sterling 11,781 vs pounds sterling 4789; p < 0.05) in the additional analysis with imputed missing data. The impact of relapse was even greater in the 6-month cost comparison. The average 6-month costs for patients who relapsed were found to be about three times higher than for those who did not relapse (pounds sterling 4083 vs pounds sterling 1298; p < 0.05)., Conclusions: Our findings confirm the significant economic impact of relapse in BD patients after an acute manic or mixed episode, even when considering direct costs only. Such costs were dominated by inpatient stay. Nevertheless, the use of UK unit costs requires caution when interpreting this costing in the context of a specific country, as resource use and the associated costs will differ by country.

Background: Relapse in schizophrenia is one of the greatest burdens of the illness., Aims: To estimate the costs associated with relapse in a pan-European naturalistic setting., Method: The SOHO study is a 3-year, prospective, observational study of 10,972 outpatients with schizophrenia across 10 European countries. The cost of resource use (inpatient stay, day care, psychiatrist visits and medication) for those who ever relapsed in three years was compared to those who never relapsed. One-year costs for both groups were also compared for a more stringent comparison. The analyses were adjusted for patient characteristics and took account of non-normality of the cost data by using a log-link function. UK unit costs were applied to resource use. The analysis was repeated after multiple imputation for missing data., Results: Costs incurred by patients who ever relapsed ( pound14,055) during three years were almost double to those incurred by patients who never relapsed ( pound7417). 61% of the cost difference was accounted for by hospital stay. The impact of relapse was even greater in the 1-year cost comparison. Results from the additional analysis with imputed missing data remained largely consistent., Conclusions: Our findings confirm the significant economic burden of relapse, and show such costs were mainly due to hospital stay. Nevertheless, the use of UK unit costs requires caution when interpreting this costing in the context of a specific country, as resource use and their associated costs will differ by country.

Objective: To determine the cost utility of treating schizophrenic patients with olanzapine compared with other antipsychotics in a naturalistic outpatient setting., Methods: The pan-European SOHO study is a 3-year, prospective, outpatient, observational study of outcomes associated with antipsychotic treatment, focusing on olanzapine, in ten European countries. For the cost-utility analysis, healthcare resource use (inpatient care, day care, outpatient psychiatric consultations and antipsychotic and concomitant medication use) and EQ-5D data were collected at baseline and at 3, 6 and 12 months. The perspective was that of the health service payer. UK healthcare unit costs (year 2004 values) were applied to the resource use data for the ten countries. UK population tariffs were applied to the EQ-5D data to determine utility values.An Epoch analysis was used to analyze the longitudinal data. Multivariate regression analyses that adjusted for baseline covariates were used to estimate the incremental cost and utility gains for patients treated with olanzapine compared with each of the other antipsychotics (risperidone, quetiapine, amisulpride, clozapine and oral or depot typical antipsychotics)., Results: A total of 10 972 patients were enrolled at baseline, of which 9107 completed the 12-month study period. Treatment with olanzapine was more effective in terms of QALYs gained than all of the other antipsychotic treatments. Treatment with olanzapine dominated quetiapine and amisulpride. The incremental cost for olanzapine compared with risperidone was pound sterling 226 per patient over 12 months and the incremental cost per QALY gained was pound 5156, with bootstrap analyses showing 100% of the replications falling below a pound sterling 30 000 per QALY gained threshold. Compared with treatment with clozapine, olanzapine was found to be marginally more effective, at an additional cost of pound sterling 13 per patient over 12 months and to have an incremental cost per QALY gained of pound sterling 775. Bootstrap analyses showed that 81% of replications fell below a pound sterling 30 000 per QALY gained threshold. Comparing olanzapine with oral and depot typical antipsychotics, the incremental cost was pound sterling 849 and pound sterling 1106 per patient over 12 months and the incremental cost per QALY gained was pound sterling 15 696 and pound sterling 23 331, respectively. Bootstrap analyses showed that 98% of the replications fell below a pound sterling 30 000 per QALY gained threshold for the comparison with oral typical antipsychotics, and 79% of replications for the comparison with depot preparations., Conclusions: Among SOHO patients, if a funding threshold of pound sterling 30 000 per QALY gained is assumed, this analysis suggests that olanzapine has a high probability of being the most cost-effective treatment compared with other antipsychotic treatments. However, comparison of olanzapine with clozapine and typical depot antipsychotics should be viewed with caution because clozapine is a second-line treatment and depot treatment is used for patients who do not adhere to their oral medication.

Objectives: The objective of this study was to develop a method to allocate treatment effects when patients switch medication frequently in longitudinal observational studies and apply the approach to assess the cost-effectiveness of treatments in the Schizophrenia Outpatient Health Outcomes (SOHO) study., Methods: Data were collected on patients at entry to the SOHO study at 3, 6, and 12 months. The 12-month follow-up period was considered as three epochs: 0-3 months, 3-6 months, and 6-12 months. Patients who switched treatment at 3 months had their new treatment considered as a new baseline observation, as these two 3-month observations provide two sets of information on the cost-effectiveness of a drug in the first 3 months after initiation. Multivariate regression analysis was used to adjust for baseline covariates. The model allowed for flexible functional forms, and the cost data were modeled using an exponential mean function. Bootstrapping assessed the uncertainty of the estimated parameters and incremental cost-effectiveness analysis decision rule., Results and Conclusions: We show the feasibility of the epoch analysis approach using data from the SOHO study comparing two antipsychotics. Estimates for the incremental cost and effectiveness per epoch over the full 12-month period are presented. Using the estimates of 200 bootstrap samples, we demonstrate how one drug is cost-effective compared with another.

Prospective observational studies, which provide information on the effectiveness of interventions in natural settings, may complement results from randomised clinical trials in the evaluation of health technologies. However, observational studies are subject to a number of potential methodological weaknesses, mainly selection and observer bias. This paper reviews and applies various methods to control for selection bias in the estimation of treatment effects and proposes novel ways to assess the presence of observer bias. We also address the issues of estimation and inference in a multilevel setting. We describe and compare the use of regression methods, propensity score matching, fixed-effects models incorporating investigator characteristics, and a multilevel, hierarchical model using Bayesian estimation techniques in the control of selection bias. We also propose to assess the existence of observer bias in observational studies by comparing patient- and investigator-reported outcomes. To illustrate these methods, we have used data from the SOHO (Schizophrenia Outpatient Health Outcomes) study, a large, prospective, observational study of health outcomes associated with the treatment of schizophrenia. The methods used to adjust for differences between treatment groups that could cause selection bias yielded comparable results, reinforcing the validity of the findings. Also, the assessment of observer bias did not show that it existed in the SOHO study. Observational studies, when properly conducted and when using adequate statistical methods, can provide valid information on the evaluation of health technologies.

The aim of this paper is to empirically analyse the responses by general practitioners to promotional activities for ethical drugs by pharmaceutical companies. Promotion can be beneficial as a means of providing information, but it can also be harmful in the sense that it lowers price sensitivity of doctors and it merely is a means of maintaining market share, even when cheaper, therapeutically equivalent drugs are available. A model is estimated that includes interactions of promotion expenditures and prices and that explicitly exploits the panel structure of the data, allowing for drug specific effects and dynamic adjustments, or habit persistence. The data used are aggregate monthly GP prescriptions per drug together with monthly outlays on drug promotion for the period 1994-1999 for 11 therapeutic markets, covering more than half of the total prescription drug market in the Netherlands. Identification of price effects is aided by the introduction of the Pharmaceutical Prices Act, which established that Dutch drugs prices became a weighted average of the prices in surrounding countries after June 1996. We conclude that GP drug price sensitivity is small, but adversely affected by promotion. Ltd., (Copyright (c) 2005 John Wiley & Sons, Ltd.)

We report an empirical analysis of the responses of the supply and demand for secondary care to waiting list size and waiting times. Whereas previous empirical analyses have used data aggregated to area level, our analysis focuses on the supply responses of a single hospital and the demand responses of the GP practices it serves, and distinguishes between first outpatient visits, inpatient admissions, day-case treatment and emergency admissions. The results are plausible and in line with the theoretical model. For example: the demand from practices for first outpatient visits is negatively affected by waiting times and distance to the hospital. Increases in waiting times and waiting lists lead to increases in supply; the supply of elective inpatient admissions is affected negatively by current emergency admissions and positively by lagged waiting list and waiting time. We use the empirical results to investigate the dynamic responses to one off policy measures to reduce waiting times and lists by increasing supply., ((c) 2005 John Wiley & Sons, Ltd.)

Many health-care systems allocate funding according to measures of need. The utilisation approach for measuring need rests on the assumptions that use of health care is determined by demand and supply and that need is an important element of demand. By estimating utilisation models which allow for supply it is possible to isolate the socio-economic and health characteristics which affect demand. A subset of these variables can then be identified by a combination of judgement and further analysis as needs variables to inform funding allocations. We estimate utilisation models using newly assembled data on admissions to acute hospitals, measures of supply, morbidity and socio-economic characteristics for 8414 small geographical areas in England. We make a number of methodological innovations including deriving additional measures of specific morbidities at small area level from individual level survey data. We compare models with different specifications for the effect of waiting times and provider characteristics, with total, planned and unplanned hospital admissions, and estimated at small area (ward) and primary care organisation (general practice) level. After allowing for waiting times, distance, capacity and the availability of private health care, measures of mortality, self-reported morbidity, low education and low income increase the use of health care. We find evidence of horizontal inequity with respect to ethnicity and employment and suggest a method for reducing its effects when deriving a needs-based allocation formula., (Copyright 2003 John Wiley & Sons, Ltd.)