ORCID Profile
0000-0003-3890-6206
Current Organisations
London School of Hygiene and Tropical Medicine
,
University of Oxford
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: SAGE Publications
Date: 31-07-2023
DOI: 10.1177/09622802231188518
Abstract: Clinical trials that investigate physical activity interventions often use accelerometers to measure step count at a very granular level, for ex le in 5-second epochs. Participants typically wear the accelerometer for a week-long period at baseline, and for one or more week-long follow-up periods after the intervention. The data is aggregated to provide daily or weekly step counts for the primary analysis. Missing data are common as participants may not wear the device as per protocol. Approaches to handling missing data in the literature have defined missingness on the day level using a threshold on daily weartime, which leads to loss of information on the time of day when data are missing. We propose an approach to identifying and classifying missingness at the finer epoch-level and present two approaches to handling missingness using multiple imputation. Firstly, we present a parametric approach which accounts for the number of missing epochs per day. Secondly, we describe a non-parametric approach where missing periods during the day are replaced by donor data from the same person where possible, or data from a different person who is matched on demographic and physical activity-related variables. Our simulation studies show that the non-parametric approach leads to estimates of the effect of treatment that are least biased while maintaining small standard errors. We illustrate the application of these different multiple imputation strategies to the analysis of the 2017 PACE-UP trial. The proposed framework is likely to be applicable to other digital health outcomes and to other wearable devices.
Publisher: Paediatrician Publishers LLC
Date: 06-05-2023
Abstract: Background . Well-written and transparent case reports (1) reveal early signals of potential benefits, harms, and information on the use of resources (2) provide information for clinical research and clinical practice guidelines, and (3) inform medical education. High-quality case reports are more likely when authors follow reporting guidelines. During 2011–2012, a group of clinicians, researchers, and journal editors developed recommendations for the accurate reporting of information in case reports that resulted in the CARE (CAse REport) Statement and Checklist. They were presented at the 2013 International Congress on Peer Review and Biomedical Publication, have been endorsed by multiple medical journals, and translated into nine languages. Objectives . This explanation and elaboration document has the objective to increase the use and dissemination of the CARE Checklist in writing and publishing case reports. Article design and setting . Each item from the CARE Checklist is explained and accompanied by published ex les. The explanations and ex les in this document are designed to support the writing of high-quality case reports by authors and their critical appraisal by editors, peer reviewers, and readers. Results and conclusion. This article and the 2013 CARE Statement and Checklist, available from the CARE website [www.care-statement.org] and the EQUATOR Network [www.equator-network.org], are resources for improving the completeness and transparency of case reports. Source . This article is a translation of the original paper «CARE guidelines for case reports: explanation and elaboration document» in the Journal of Clinical Epidemiology (doi: 10.1016/j.jclinepi.2017.04.026), prepared under the permission of the copyright holder (Elsevier Inc.), with supervision from the Scientific Editor by Professor E.G. Starostina, MD, PhD (translator) (Moscow, Russia). Present translation was first published in Digital Diagnostics. doi: 10.17816/DD105291. It is published with minor changes related to the literary editing of the translation itself. Keywords: case report case study EQUATOR network health research reporting guidelines CARE guideline timelines N-of-1 For citation: Riley David S., Barber Melissa S., Kienle Gunver S., Aronson Jeffrey K., von Schoen-Angerer Tido, Tugwell Peter, Kiene Helmut, Helfand Mark, Altman Douglas G., Sox Harold, Werthmann Paul G., Moher David, Rison Richard A., Shamseer Larissa, Koch Christian A., Sun Gordon H., Hanaway Patrick, Sudak Nancy L., Kaszkin-Bettag Marietta, Carpenter James E., Gagnier Joel J. CARE Guidelines for Case Reports: Explanation and Elaboration Document. Translation into Russian. Voprosy sovremennoi pediatrii — Current Pediatrics. 2023 (2):88–108. (In Russ). doi: 0.15690/vsp.v22i2.2540
Publisher: BMJ
Date: 29-06-2009
DOI: 10.1136/BMJ.B2393
Publisher: Oxford University Press (OUP)
Date: 08-07-2010
DOI: 10.1093/AJE/KWQ137
Abstract: Multiple imputation is increasingly recommended in epidemiology to adjust for the bias and loss of information that may occur in analyses restricted to study participants with complete data ("complete-case analyses"). However, little guidance is available on applying the method, including which variables to include in the imputation model and the number of imputations needed. Here, the authors used multiple imputation to analyze the prevalence of wheeze among 81-month-old children in the Avon Longitudinal Study of Parents and Children (Avon, United Kingdom 1991-1999) and the association of wheeze with gender, maternal asthma, and maternal smoking. The authors examined how inclusion of different types of variables in the imputation model affected point estimates and precision, and assessed the impact of number of imputations on Monte Carlo variability. Inclusion of variables associated with the outcome in the imputation model increased odds ratios and reduced standard errors. When only 5 or 10 imputations were used, variability due to the imputation procedure was substantial enough to affect conclusions. Careful preliminary analysis identified the scope for multiple imputation to reduce bias and improve efficiency and provided guidance for building the imputation model. When data are missing, such preliminary analyses should be routinely undertaken and reported, regardless of whether multiple imputation is used in the final analysis.
Publisher: BMJ
Date: 22-07-2011
DOI: 10.1136/BMJ.D4002
Publisher: Elsevier BV
Date: 12-2009
DOI: 10.1016/J.OPHTHA.2009.10.022
Abstract: To compare the visual outcomes after verteporfin photodynamic therapy (VPDT) administered in routine clinical practice with those observed in the Treatment of Age-related macular degeneration with Photodynamic therapy (TAP) trials and to quantify the effects of clinically important baseline covariates on outcome. A prospective longitudinal study of patients treated with VPDT in 45 ophthalmology departments in the United Kingdom with expertise in the management of neovascular age-related macular degeneration (nAMD). Patients with wholly or predominantly classic choroidal neovascularization (CNV) of any cause with a visual acuity >or=20/200 in the eye to be treated. Refracted best-corrected visual acuity (BCVA) and contrast sensitivity were measured in VPDT-treated eyes at baseline and subsequent visits. Eyes were retreated at 3 months if CNV was judged to be active. Baseline angiograms were graded to quantify the percentages of classic and occult CNV. Treated eyes were categorized as eligible or ineligible for TAP, or unclassifiable. Best-corrected visual acuity and contrast sensitivity during 1 year of follow-up after initial treatment. A total of 7748 treated patients were recruited. Data from 4043 patients with a diagnosis of nAMD were used in the present analysis. Reading center determination of lesion type showed that 87% were predominantly classic CNV. Eyes received 2.4 treatments in year 1 and 0.4 treatments in year 2. Deterioration of BCVA over 1 year was similar to that observed in the VPDT arms of the TAP trials and was not influenced by TAP eligibility classification. Best-corrected visual acuity deteriorated more quickly in current smokers with increasing proportion of classic CNV, increasing age, and better baseline BCVA and when the fellow eye was the better eye. Patients in the cohort who would have been eligible for the TAP trials demonstrated deterioration in BCVA similar to VPDT-treated TAP participants but with fewer treatments. Clinical covariates with a significant impact on BCVA outcomes were identified.
Publisher: Wiley
Date: 20-10-2009
DOI: 10.1002/PST.391
Abstract: The Points to Consider Document on Missing Data was adopted by the Committee of Health and Medicinal Products (CHMP) in December 2001. In September 2007 the CHMP issued a recommendation to review the document, with particular emphasis on summarizing and critically appraising the pattern of drop-outs, explaining the role and limitations of the 'last observation carried forward' method and describing the CHMP's cautionary stance on the use of mixed models. In preparation for the release of the updated guidance document, statisticians in the Pharmaceutical Industry held a one-day expert group meeting in September 2008. Topics that were debated included minimizing the extent of missing data and understanding the missing data mechanism, defining the principles for handling missing data and understanding the assumptions underlying different analysis methods. A clear message from the meeting was that at present, biostatisticians tend only to react to missing data. Limited pro-active planning is undertaken when designing clinical trials. Missing data mechanisms for a trial need to be considered during the planning phase and the impact on the objectives assessed. Another area for improvement is in the understanding of the pattern of missing data observed during a trial and thus the missing data mechanism via the plotting of data for ex le, use of Kaplan-Meier curves looking at time to withdrawal.
Publisher: Springer Science and Business Media LLC
Date: 23-04-2019
Publisher: Springer Science and Business Media LLC
Date: 29-04-2019
Publisher: National Institute for Health and Care Research
Date: 02-2012
DOI: 10.3310/HTA16060
Abstract: The verteporfin photodynamic therapy (VPDT) cohort study aimed to answer five questions: (a) is VPDT in the NHS provided as in randomised trials? (b) is 'outcome' the same in the nhs as in randomised trials? (c) is 'outcome' the same for patients ineligible for randomised trials? (d) is VPDT safe when provided in the NHS? and (e) how effective and cost-effective is VPDT? Treatment register. All hospitals providing VPDT in the NHS. All patients attending VPDT clinics. Infusion of verteporfin followed by infrared laser exposure is called VPDT, and is used to treat neovascular age-related macular degeneration (nAMD). The VPDT cohort study advised clinicians to follow patients every 3 months during treatment or active observation, retreating based on criteria used in the previous commercial 'TAP' (Treatment of Age-related macular degeneration with Photodynamic therapy) trials of VPDT. The primary outcome was logarithm of the minimum angle of resolution monocular best-corrected distance visual acuity (BCVA). Secondary outcomes were adverse reactions and events morphological changes in treated nAMD (wet) lesions and for a subset of patients, 6-monthly contrast sensitivity, generic and visual health-related quality of life (HRQoL) and resource use. Treated eyes were classified as eligible for the TAP trials (EFT), ineligible (IFT) or unclassifiable (UNC). Forty-seven hospitals submitted data for 8323 treated eyes in 7748 patients 4919 eyes in 4566 patients were treated more than 1 year before the last data submission or had completed treatment. Of 4043 eyes with nAMD in 4043 patients, 1227 were classified as EFT, 1187 as IFT and 1629 as UNC. HRQoL and resource use data were available for about 2000 patients. The mean number of treatments in years 1 and 2 was 2.3 and 0.4 respectively. About 50% of eyes completed treatment within 1 year. BCVA deterioration in year 1 did not differ between eligibility groups. EFT eyes lost 11.6 letters (95% confidence interval 10.1 to 13.0 letters) compared with 9.9 letters in VPDT-treated eyes in the TAP trials. EFT eyes had poorer BCVA at baseline than IFT and UNC eyes. Adverse reactions and events were reported for 1.4% of first visits - less frequently than those reported in the TAP trials. Associations between BCVA in the best-seeing eye with HRQoL and community health and social care resource use showed that the 11-letter difference in BCVA between VPDT and sham treatment in the TAP trials corresponded to differences in utility of 0.012 and health and social service costs of £60 and £92 in years 1 and 2, respectively. VPDT provided an incremental cost per quality-adjusted life-year (QALY) of £170,000 over 2 years. VPDT was administered less frequently than in the TAP trials, with less than half of those treated followed up for > 1 year in routine clinical practice. Deterioration in BCVA over time in EFT eyes was similar to that in the TAP trials. The similar falls in BCVA after VPDT across the pre-defined TAP eligibility groups do not mean that the treatment is equally effective in these groups because deterioration in BCVA can be influenced by the parameters that determined group membership. Safety was no worse than in the TAP trials. The estimated cost per QALY was similar to the highest previous estimate. Although VPDT is no longer in use as monotherapy for neovascular AMD, its role as adjunctive treatment has not been fully explored. VPDT also has potential as monotherapy in the management of vascular malformations of the retina and choroid and with trials underway in neovascularisation due to myopia and polypoidal choroidopathy. The National Institute for Health Research Health Technology Assessment programme.
Publisher: European Respiratory Society (ERS)
Date: 22-10-2021
DOI: 10.1183/13993003.01586-2020
Abstract: Real-world data provide the potential for generating evidence on drug treatment effects in groups excluded from trials, but rigorous, validated methodology for doing so is lacking. We investigated whether non-interventional methods applied to real-world data could reproduce results from the landmark TORCH COPD trial. We performed a historical cohort study (2000–2017) of COPD drug treatment effects in the UK Clinical Practice Research Datalink (CPRD). Two control groups were selected from CPRD by applying TORCH inclusion/exclusion criteria and 1:1 matching to TORCH participants, as follows. Control group 1: people with COPD not prescribed fluticasone propionate (FP)-salmeterol (SAL) control group 2: people with COPD prescribed SAL only. FP-SAL exposed groups were then selected from CPRD by propensity score matching to each control group. Outcomes studied were COPD exacerbations, death from any cause and pneumonia. 2652 FP-SAL exposed people were propensity score matched to 2652 FP-SAL unexposed people while 991 FP-SAL exposed people were propensity score matched to 991 SAL exposed people. Exacerbation rate ratio was comparable to TORCH for FP-SAL versus SAL (0.85, 95% CI 0.74–0.97 versus 0.88, 0.81–0.95) but not for FP-SAL versus no FP-SAL (1.30, 1.19–1.42 versus 0.75, 0.69–0.81). In addition, active comparator results were consistent with TORCH for mortality (hazard ratio 0.93, 0.65–1.32 versus 0.93, 0.77–1.13) and pneumonia (risk ratio 1.39, 1.04–1.87 versus 1.47, 1.25–1.73). We obtained very similar results to the TORCH trial for active comparator analyses, but were unable to reproduce placebo-controlled results. Application of these validated methods for active comparator analyses to groups excluded from randomised controlled trials provides a practical way for contributing to the evidence base and supporting COPD treatment decisions.
Publisher: Wiley
Date: 27-02-2020
DOI: 10.1002/SIM.8503
Publisher: Wiley
Date: 25-01-2023
DOI: 10.1002/SIM.9658
Abstract: One of the main challenges when using observational data for causal inference is the presence of confounding. A classic approach to account for confounding is the use of propensity score techniques that provide consistent estimators of the causal treatment effect under four common identifiability assumptions for causal effects, including that of no unmeasured confounding. Propensity score matching is a very popular approach which, in its simplest form, involves matching each treated patient to an untreated patient with a similar estimated propensity score, that is, probability of receiving the treatment. The treatment effect can then be estimated by comparing treated and untreated patients within the matched dataset. When missing data arises, a popular approach is to apply multiple imputation to handle the missingness. The combination of propensity score matching and multiple imputation is increasingly applied in practice. However, in this article we demonstrate that combining multiple imputation and propensity score matching can lead to over‐coverage of the confidence interval for the treatment effect estimate. We explore the cause of this over‐coverage and we evaluate, in this context, the performance of a correction to Rubin's rules for multiple imputation proposed by finding that this correction removes the over‐coverage.
Publisher: Wiley
Date: 30-07-2014
DOI: 10.1002/SIM.6265
Publisher: National Institute for Health and Care Research
Date: 08-2021
DOI: 10.3310/HTA25510
Abstract: Chronic obstructive pulmonary disease treatment is informed by randomised controlled trial results, but it is unclear if these findings apply to people excluded from these trials. We used data from the TORCH (TOwards a Revolution in COPD Health) randomised controlled trial to validate non-interventional methods for assessing the clinical effectiveness of chronic obstructive pulmonary disease treatment in the UK Clinical Practice Research Datalink, before applying these methods to the analysis of people who would have been excluded from TORCH. To validate the use of non-interventional Clinical Practice Research Datalink data and methods for estimating chronic obstructive pulmonary disease treatment effects against trial results, and, using validated methods, to determine treatment effects in people who would have been excluded from the TORCH trial. A historical non-interventional cohort design, including validation against randomised controlled trial results. The UK Clinical Practice Research Datalink. People aged ≥ 18 years with chronic obstructive pulmonary disease registered in Clinical Practice Research Datalink GOLD between January 2000 and January 2017. For objective 1, we prepared a cohort that was analogous to the TORCH trial cohort by applying TORCH trial inclusion/exclusion criteria followed by in idual matching to TORCH trial participants. For objectives 2 and 3, we prepared cohorts that were analogous to the TORCH trial that, nevertheless, would not have been eligible for the TORCH trial because of age, asthma, comorbidity or mild disease. The long-acting beta-2 agonist and inhaled corticosteroid combination product Seretide (GlaxoSmithKline plc) [i.e. fluticasone propionate plus salmeterol (FP-SAL)] compared with (1) no FP-SAL exposure or (2) exposure to salmeterol (i.e. the long-acting beta-2 agonist) only. Exacerbations, mortality, pneumonia and time to treatment change. For objective 1, the exacerbation rate ratio was comparable to that in the TORCH trial for FP-SAL compared with salmeterol (0.85, 95% confidence interval 0.74 to 0.97, vs. TORCH trial 0.88, 95% confidence interval 0.81 to 0.95), but not for FP-SAL compared with no FP-SAL (1.30, 95% confidence interval 1.19 to 1.42, vs. TORCH trial 0.75, 95% confidence interval 0.69 to 0.81). Active comparator results were also consistent with the TORCH trial for mortality (hazard ratio 0.93, 95% confidence interval 0.65 to 1.32, vs. TORCH trial hazard ratio 0.93, 95% confidence interval 0.77 to 1.13) and pneumonia (risk ratio 1.39, 95% confidence interval 1.04 to 1.87, vs. TORCH trial risk ratio 1.47, 95% confidence interval 1.25 to 1.73). For objectives 2 and 3, active comparator results were consistent with the TORCH trial for exacerbations, with the exception of people with milder chronic obstructive pulmonary disease, in whom we observed a stronger protective association (risk ratio 0.56, 95% confidence interval 0.46 to 0.70, vs. TORCH trial risk ratio 0.85, 95% confidence interval 0.74 to 0.97). For the analysis of mortality, we saw a lack of association with being prescribed FP-SAL (vs. being prescribed salmeterol), with the exception of those with prior asthma, for whom we observed an increase in mortality (hazard ratio 1.49, 95% confidence interval 1.21 to 1.85, vs. TORCH trial-analogous HR 0.93, 95% confidence interval 0.64 to 1.32). Routinely collected electronic health record data can be used to successfully measure chronic obstructive pulmonary disease treatment effects when comparing two treatments, but not for comparisons between active treatment and no treatment. Analyses involving patients who would have been excluded from trials mostly suggests that treatment effects for FP-SAL are similar to trial effects, although further work is needed to characterise a small increased risk of death in those with concomitant asthma. Some of our analyses had small numbers. The differences in treatment effects that we found should be investigated further in other data sets. Currently recommended chronic obstructive pulmonary disease inhaled combination therapy (other than FP-SAL) should also be investigated using these methods. This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment Vol. 25, No. 51. See the NIHR Journals Library website for further project information.
Publisher: SAGE Publications
Date: 24-01-2012
Abstract: Estimation of the effect of a binary exposure on an outcome in the presence of confounding is often carried out via outcome regression modelling. An alternative approach is to use propensity score methodology. The propensity score is the conditional probability of receiving the exposure given the observed covariates and can be used, under the assumption of no unmeasured confounders, to estimate the causal effect of the exposure. In this article, we provide a non-technical and intuitive discussion of propensity score methodology, motivating the use of the propensity score approach by analogy with randomised studies, and describe the four main ways in which this methodology can be implemented. We carefully describe the population parameters being estimated — an issue that is frequently overlooked in the medical literature. We illustrate these four methods using data from a study investigating the association between maternal choice to provide breast milk and the infant's subsequent neurodevelopment. We outline useful extensions of propensity score methodology and discuss directions for future research. Propensity score methods remain controversial and there is no consensus as to when, if ever, they should be used in place of traditional outcome regression models. We therefore end with a discussion of the relative advantages and disadvantages of each.
Publisher: Johns Hopkins School Bloomberg School of Public Health, Center for Communication Programs
Date: 23-09-2019
Publisher: Cold Spring Harbor Laboratory
Date: 10-04-2020
DOI: 10.1101/2020.04.08.20033373
Abstract: Clinical trials generally each collect their own data despite routinely-collected health data (RCHD) increasing in quality and breadth. Our aim is to quantify UK-based randomised controlled trials (RCTs) accessing RCHD for participant data, characterise how these data are used and thereby recommend how more trials could use RCHD. We conducted a systematic review of RCTs accessing RCHD from at least one registry in the UK between 2013-2018, for the purposes of informing or supplementing participant data. A list of all registries holding RCHD in the UK was compiled. In cases where registries published release registers, these were searched for RCTs accessing RCHD. Where no release register was available, registries were contacted to request a list of RCTs. For each identified RCT, information was collected from all publicly available sources (release registers, websites, protocol etc.). The search and data extraction was undertaken between Jan-2019 and May-2019. We identified 160 RCTs accessing RCHD between 2013 and 2018 from a total of 22 registries this corresponds to only a very small proportion of all UK RCTs (approximately 3%). RCTs accessing RCHD were generally large (median s le size 1590), commonly evaluating treatments for cancer or cardiovascular disease. Most of the included RCTs accessed RCHD from NHS Digital (68%), and the most frequently accessed datasets were mortality (76%) and hospital visits (55%). RCHD was used to inform the primary trial (82%) and long-term follow-up (57%). There was substantial variation in how RCTs used RCHD to inform participant outcome measures. A limitation was the lack of information and transparency from registries and RCTs with respect to which datasets have been accessed and for what purposes. In the last five years, only a small minority of UK-based RCTs have accessed RCHD to inform participant data. We ask for improved accessibility, confirmed data quality and joined up thinking between the registries and the regulatory authorities. PROSPERO CRD42019123088
Publisher: Springer Science and Business Media LLC
Date: 12-05-2020
DOI: 10.1186/S13063-020-04329-8
Abstract: Clinical trials generally each collect their own data despite routinely collected health data (RCHD) increasing in quality and breadth. Our aim is to quantify UK-based randomised controlled trials (RCTs) accessing RCHD for participant data, characterise how these data are used and thereby recommend how more trials could use RCHD. We conducted a systematic review of RCTs accessing RCHD from at least one registry in the UK between 2013 and 2018 for the purposes of informing or supplementing participant data. A list of all registries holding RCHD in the UK was compiled. In cases where registries published release registers, these were searched for RCTs accessing RCHD. Where no release register was available, registries were contacted to request a list of RCTs. For each identified RCT, information was collected from all publicly available sources (release registers, websites, protocol etc.). The search and data extraction were undertaken between January and May 2019. We identified 160 RCTs accessing RCHD between 2013 and 2018 from a total of 22 registries this corresponds to only a very small proportion of all UK RCTs (about 3%). RCTs accessing RCHD were generally large (median s le size 1590), commonly evaluating treatments for cancer or cardiovascular disease. Most of the included RCTs accessed RCHD from NHS Digital (68%), and the most frequently accessed datasets were mortality (76%) and hospital visits (55%). RCHD was used to inform the primary trial (82%) and long-term follow-up (57%). There was substantial variation in how RCTs used RCHD to inform participant outcome measures. A limitation was the lack of information and transparency from registries and RCTs with respect to which datasets have been accessed and for what purposes. In the last five years, only a small minority of UK-based RCTs have accessed RCHD to inform participant data. We ask for improved accessibility, confirmed data quality and joined-up thinking between the registries and the regulatory authorities. PROSPERO CRD42019123088 .
Publisher: Elsevier BV
Date: 06-2021
Publisher: ECO-Vector LLC
Date: 24-04-2022
DOI: 10.17816/DD105291
Abstract: BACKGROUND: Well-written and transparent case reports (1) reveal early signals of potential benefits, harms, and information on the use of resources (2) provide information for clinical research and clinical practice guidelines, and (3) inform medical education. High-quality case reports are more likely when authors follow reporting guidelines. During 20112012, a group of clinicians, researchers, and journal editors developed recommendations for the accurate reporting of information in case reports that resulted in the CARE (CAse REport) Statement and Checklist. They were presented at the 2013 International Congress on Peer Review and Biomedical Publication, have been endorsed by multiple medical journals, and translated into nine languages. OBJECTIVES: This explanation and elaboration document has the objective to increase the use and dissemination of the CARE Checklist in writing and publishing case reports. ARTICLE DESIGN AND SETTING: Each item from the CARE Checklist is explained and accompanied by published ex les. The explanations and ex les in this document are designed to support the writing of high-quality case reports by authors and their critical appraisal by editors, peer reviewers, and readers. RESULTS AND CONCLUSION: This article and the 2013 CARE Statement and Checklist, available from the CARE website [www.care-statement.org] and the EQUATOR Network [www.equator-network.org], are resources for improving the completeness and transparency of case reports. SOURCE: This article is a translation of the original paper CARE guidelines for case reports: explanation and elaboration document in the Journal of Clinical Epidemiology (doi: 10.1016/j.jclinepi.2017.04.026), prepared under the permission of the copyright holder (Elsevier Inc.), with supervision from the Scientific Editor by Professor E.G. Starostina, MD, PhD (translator) (Moscow, Russia).
Publisher: BMJ
Date: 12-10-2016
DOI: 10.1136/BMJ.I4919
Publisher: Wiley
Date: 24-02-2012
DOI: 10.1002/SIM.4504
Abstract: Propensity score methods are increasingly used to estimate the effect of a treatment or exposure on an outcome in non-randomised studies. We focus on one such method, stratification on the propensity score, comparing it with the method of inverse-probability weighting by the propensity score. The propensity score--the conditional probability of receiving the treatment given observed covariates--is usually an unknown probability estimated from the data. Estimators for the variance of treatment effect estimates typically used in practice, however, do not take into account that the propensity score itself has been estimated from the data. By deriving the asymptotic marginal variance of the stratified estimate of treatment effect, correctly taking into account the estimation of the propensity score, we show that routinely used variance estimators are likely to produce confidence intervals that are too conservative when the propensity score model includes variables that predict (cause) the outcome, but only weakly predict the treatment. In contrast, a comparison with the analogous marginal variance for the inverse probability weighted (IPW) estimator shows that routinely used variance estimators for the IPW estimator are likely to produce confidence intervals that are almost always too conservative. Because exact calculation of the asymptotic marginal variance is likely to be complex, particularly for the stratified estimator, we suggest that bootstrap estimates of variance should be used in practice.
Location: United Kingdom of Great Britain and Northern Ireland
Location: United Kingdom of Great Britain and Northern Ireland
No related grants have been discovered for James Carpenter.