umu.sePublications
Change search
Link to record
Permanent link

Direct link
BETA
de Luna, Xavier, ProfessorORCID iD iconorcid.org/0000-0003-3187-1987
Alternative names
Publications (10 of 67) Show all publications
Genbäck, M. & de Luna, X. (2019). Causal inference accounting for unobserved confounding after outcome regression and doubly robust estimation. Biometrics, 75(2), 506-515
Open this publication in new window or tab >>Causal inference accounting for unobserved confounding after outcome regression and doubly robust estimation
2019 (English)In: Biometrics, ISSN 0006-341X, E-ISSN 1541-0420, Vol. 75, no 2, p. 506-515Article in journal (Refereed) Published
Abstract [en]

Causal inference with observational data can be performed under an assumption of no unobserved confounders (unconfoundedness assumption). There is, however, seldom clear subject-matter or empirical evidence for such an assumption. We therefore develop uncertainty intervals for average causal effects based on outcome regression estimators and doubly robust estimators, which provide inference taking into account both sampling variability and uncertainty due to unobserved confounders. In contrast with sampling variation, uncertainty due to unobserved confounding does not decrease with increasing sample size. The intervals introduced are obtained by modeling the treatment assignment mechanism and its correlation with the outcome given the observed confounders, allowing us to derive the bias of the estimators due to unobserved confounders. We are thus also able to contrast the size of the bias due to violation of the unconfoundedness assumption, with bias due to misspecification of the models used to explain potential outcomes. This is illustrated through numerical experiments where bias due to moderate unobserved confounding dominates misspecification bias for typical situations in terms of sample size and modeling assumptions. We also study the empirical coverage of the uncertainty intervals introduced and apply the results to a study of the effect of regular food intake on health. An R-package implementing the inference proposed is available.

Place, publisher, year, edition, pages
Wiley-Blackwell, 2019
Keywords
Average causal effects, double robust, ignorability assumption, regular food intake, sensitivity analysis, uncertainty intervals
National Category
Probability Theory and Statistics
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-153868 (URN)10.1111/biom.13001 (DOI)000483730600018 ()
Funder
Forte, Swedish Research Council for Health, Working Life and Welfare, 2013-2506Marianne and Marcus Wallenberg Foundation
Available from: 2018-12-06 Created: 2018-12-06 Last updated: 2019-10-09Bibliographically approved
Ghosh, T., Ma, Y. & de Luna, X. (2019). Sufficient Dimension Reduction for Feasible and Robust Estimation of Average Causal Effect. Statistica sinica
Open this publication in new window or tab >>Sufficient Dimension Reduction for Feasible and Robust Estimation of Average Causal Effect
2019 (English)In: Statistica sinica, ISSN 1017-0405, E-ISSN 1996-8507Article in journal (Refereed) Epub ahead of print
Abstract [en]

When estimating the treatment effect in an observational study, we use a semi- parametric locally efficient dimension reduction approach to assess both the treat- ment assignment mechanism and the average responses in both treated and non- treated groups. We then integrate all results through imputation, inverse prob- ability weighting and double robust augmentation estimators. Double robust estimators are locally efficient while imputation estimators are super-efficient when the response models are correct. To take advantage of both procedures, we introduce a shrinkage estimator to automatically combine the two, which re- tains the double robustness property while improving on the variance when the response model is correct. We demonstrate the performance of these estima- tors through simulated experiments and a real dataset concerning the effect of maternal smoking on baby birth weight.

Place, publisher, year, edition, pages
Taipei: Academia Sinica, Institute of Statistical Science, 2019
National Category
Probability Theory and Statistics
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-163592 (URN)10.5705/ss.202018.0416 (DOI)
Available from: 2019-09-27 Created: 2019-09-27 Last updated: 2019-10-03
Häggström, C., Garmo, H., de Luna, X., Van Hemelrijck, M., Söderkvist, K., Aljabery, F., . . . Holmberg, L. (2019). Survival after radiotherapy versus radical cystectomy for primary muscle-invasive bladder cancer: A Swedish nationwide population-based cohort study. Cancer Medicine, 8(5), 2196-2204
Open this publication in new window or tab >>Survival after radiotherapy versus radical cystectomy for primary muscle-invasive bladder cancer: A Swedish nationwide population-based cohort study
Show others...
2019 (English)In: Cancer Medicine, ISSN 2045-7634, E-ISSN 2045-7634, Vol. 8, no 5, p. 2196-2204Article in journal (Refereed) Published
Abstract [en]

BACKGROUND: Studies of survival comparing radical cystectomy (RC) and radiotherapy for muscle-invasive bladder cancer have provided inconsistent results and have methodological limitations. The aim of the study was to investigate risk of death after radiotherapy as compared to RC.

METHODS: We selected patients with muscle-invasive urothelial carcinoma without distant metastases, treated with radiotherapy or RC from 1997 to 2014 in the Bladder Cancer Data Base Sweden (BladderBaSe) and estimated absolute and relative risk of bladder cancer death and all-cause death. In a group of patients, theoretically eligible for a trial comparing radiotherapy and RC, we calculated risk difference in an instrumental variable analysis. We have not investigated chemoradiotherapy as this treatment was not used in the study time period.

RESULTS: The study included 3 309 patients, of those 17% were treated with radiotherapy and 83% with RC. Patients treated with radiotherapy were older, had more advanced comorbidity, and had a higher risk of death as compared to patients treated with RC (relative risks of 1.5-1.6). In the "trial population," all-cause death risk difference was 6 per 100 patients lower after radiotherapy at 5 years of follow-up, 95% confidence interval -41 to 29.

CONCLUSION(S): Patient selection between the treatments make it difficult to evaluate results from conventionally adjusted and propensity-score matched survival analysis. When taking into account unmeasured confounding by instrumental variable analysis, no differences in survival was found between the treatments for a selected group of patients. Further clinical studies are needed to characterize this group of patients, which can serve as a basis for future comparison studies for treatment recommendations.

Place, publisher, year, edition, pages
John Wiley & Sons, 2019
Keywords
bladder cancer, muscle-invasive, radical cystectomy, radiotherapy, urothelial carcinoma
National Category
Cancer and Oncology
Identifiers
urn:nbn:se:umu:diva-158779 (URN)10.1002/cam4.2126 (DOI)000469272500024 ()30938068 (PubMedID)
Funder
Swedish Cancer Society, CAN 2013/472
Available from: 2019-05-08 Created: 2019-05-08 Last updated: 2019-06-20Bibliographically approved
Gorbach, T. & de Luna, X. (2018). Inference for partial correlation when data are missing not at random. Statistics and Probability Letters, 141, 82-89
Open this publication in new window or tab >>Inference for partial correlation when data are missing not at random
2018 (English)In: Statistics and Probability Letters, ISSN 0167-7152, E-ISSN 1879-2103, Vol. 141, p. 82-89Article in journal (Refereed) Published
Abstract [en]

We introduce uncertainty regions to perform inference on partial correlations when data are missing not at random. These uncertainty regions are shown to have a desired asymptotic coverage. Their finite sample performance is illustrated via simulations and real data example. (C) 2018 Elsevier B.V. All rights reserved.

Place, publisher, year, edition, pages
ELSEVIER SCIENCE BV, 2018
Keywords
Nonignorable dropout, Uncertainty region, Change-change analysis, Brain markers, Cognition
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-151035 (URN)10.1016/j.spl.2018.05.027 (DOI)000440961600011 ()2-s2.0-85048717692 (Scopus ID)
Available from: 2018-09-04 Created: 2018-09-04 Last updated: 2019-04-26Bibliographically approved
Genbäck, M., Ng, N., Stanghellini, E. & de Luna, X. (2018). Predictors of decline in self-reported health: addressing non-ignorable dropout in longitudinal studies of ageing. European Journal of Ageing, 15(2), 211-220
Open this publication in new window or tab >>Predictors of decline in self-reported health: addressing non-ignorable dropout in longitudinal studies of ageing
2018 (English)In: European Journal of Ageing, ISSN 1613-9372, E-ISSN 1613-9380, Vol. 15, no 2, p. 211-220Article in journal (Refereed) Published
Abstract [en]

Predictors of decline in health in older populations have been investigated in multiple studies before. Most longitudinal studies of aging, however, assume that dropout at follow-up is ignorable (missing at random) given a set of observed characteristics at baseline. The objective of this study was to address non-ignorable dropout in investigating predictors of declining self-reported health (SRH) in older populations (50 years or older) in Sweden, the Netherlands, and Italy. We used the SHARE panel survey, and since only 2895 out of the original 5657 participants in the survey 2004 were followed up in 2013, we studied whether the results were sensitive to the expectation that those dropping out have a higher proportion of decliners in SRH. We found that older age and a greater number of chronic diseases were positively associated with a decline in self-reported health in the three countries studies here. Maximum grip strength was associated with decline in self-reported health in Sweden and Italy, and self-reported limitations in normal activities due to health problems were associated with decline in self-reported health in Sweden. These results were not sensitive to non-ignorable dropout. On the other hand, although obesity was associated with decline in a complete case analysis, this result was not confirmed when performing a sensitivity analysis to non-ignorable dropout. The findings, thereby, contribute to the literature in understanding the robustness of longitudinal study results to non-ignorable dropout while considering three different population samples in Europe.

Place, publisher, year, edition, pages
Springer, 2018
Keywords
Longitudinal studies, Dropout, Sensitivity analysis, Chronic disease, Body mass index, SHARE
National Category
Probability Theory and Statistics Public Health, Global Health, Social Medicine and Epidemiology Gerontology, specialising in Medical and Health Sciences
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-127118 (URN)10.1007/s10433-017-0448-x (DOI)000433224700010 ()29867305 (PubMedID)
Projects
Paths to Healthy and Active Ageing
Funder
Forte, Swedish Research Council for Health, Working Life and Welfare, 2013-2506
Note

Originally included in thesis in manuscript form.

Available from: 2016-11-01 Created: 2016-10-31 Last updated: 2018-06-25Bibliographically approved
Lindmark, A., de Luna, X. & Eriksson, M. (2018). Sensitivity analysis for unobserved confounding of direct and indirect effects using uncertainty intervals. Statistics in Medicine, 37(10), 1744-1762
Open this publication in new window or tab >>Sensitivity analysis for unobserved confounding of direct and indirect effects using uncertainty intervals
2018 (English)In: Statistics in Medicine, ISSN 0277-6715, E-ISSN 1097-0258, Vol. 37, no 10, p. 1744-1762Article in journal (Refereed) Published
Abstract [en]

To estimate direct and indirect effects of an exposure on an outcome from observed data, strong assumptions about unconfoundedness are required. Since these assumptions cannot be tested using the observed data, a mediation analysis should always be accompanied by a sensitivity analysis of the resulting estimates. In this article, we propose a sensitivity analysis method for parametric estimation of direct and indirect effects when the exposure, mediator, and outcome are all binary. The sensitivity parameters consist of the correlations between the error terms of the exposure, mediator, and outcome models. These correlations are incorporated into the estimation of the model parameters and identification sets are then obtained for the direct and indirect effects for a range of plausible correlation values. We take the sampling variability into account through the construction of uncertainty intervals. The proposed method is able to assess sensitivity to both mediator‐outcome confounding and confounding involving the exposure. To illustrate the method, we apply it to a mediation study based on the data from the Swedish Stroke Register (Riksstroke). An R package that implements the proposed method is available.

Place, publisher, year, edition, pages
John Wiley & Sons, 2018
Keywords
direct effects, indirect effects, mediation, sensitivity analysis, sequential ignorability, unmeasured confounding
National Category
Probability Theory and Statistics
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-125929 (URN)10.1002/sim.7620 (DOI)000429730500011 ()29462839 (PubMedID)
Note

First published in thesis 2016 in manuscript form.

Available from: 2016-09-23 Created: 2016-09-22 Last updated: 2018-06-07Bibliographically approved
Barban, N., de Luna, X., Lundholm, E., Svensson, I. & Billari, F. C. (2017). Causal Effects of the Timing of Life-course Events: Age at Retirement and Subsequent Health. Sociological Methods & Research
Open this publication in new window or tab >>Causal Effects of the Timing of Life-course Events: Age at Retirement and Subsequent Health
Show others...
2017 (English)In: Sociological Methods & Research, ISSN 0049-1241, E-ISSN 1552-8294Article in journal (Refereed) Epub ahead of print
Abstract [en]

n this article, we combine the extensive literature on the analysis of life-course trajectories as sequences with the literature on causal inference and propose a new matching approach to investigate the causal effect of the timing of life-course events on subsequent outcomes. Our matching approach takes into account pre-event confounders that are both time-independent and time-dependent as well as life-course trajectories. After matching, treated and control individuals can be compared using standard statistical tests or regression models. We apply our approach to the study of the consequences of the age at retirement on subsequent health outcomes, using a unique data set from Swedish administrative registers. Once selectivity in the timing of retirement is taken into account, effects on hospitalization are small, while early retirement has negative effects on survival. Our approach also allows for heterogeneous treatment effects. We show that the effects of early retirement differ according to preretirement income, with higher income individuals tending to benefit from early retirement, while the opposite is true for individuals with lower income.

Place, publisher, year, edition, pages
Sage Publications, 2017
Keywords
life-course analysis, matching, propensity score, retirement, register data, sequence analysis
National Category
Probability Theory and Statistics
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-142215 (URN)10.1177/0049124117729697 (DOI)881251 (Local ID)881251 (Archive number)881251 (OAI)
Available from: 2017-11-26 Created: 2017-11-26 Last updated: 2019-04-04
Chaparro, M. P., de Luna, X., Häggström, J., Ivarsson, A., Lindgren, U., Nilsson, K. & Koupil, I. (2017). Childhood family structure and women's adult overweight risk: A longitudinal study. Scandinavian Journal of Public Health, 45(5), 511-519
Open this publication in new window or tab >>Childhood family structure and women's adult overweight risk: A longitudinal study
Show others...
2017 (English)In: Scandinavian Journal of Public Health, ISSN 1403-4948, E-ISSN 1651-1905, Vol. 45, no 5, p. 511-519Article in journal (Refereed) Published
Abstract [en]

AIM: The aim of this study was to investigate whether women's adult overweight and obesity risk was associated with their childhood family structure, measured as their mothers' marital status history, during the women's first 18 years of life.

METHODS: Using linked register data, we analyzed 30,584 primiparous women born in Sweden in 1975 who were between 19-35 years of age when their height and pre-pregnancy weight was recorded. The outcomes were women's overweight/obesity (body mass index (BMI) ≥ 25 kg/m(2)) and obesity (BMI ≥ 30 kg/m(2)) and the predictor was mothers' marital status history, which was summarized using sequence analysis. We carried out nested logistic regression models adjusting for women's age and maternal sociodemographic characteristics.

RESULTS: Mothers' marital status history was summarized into six clusters: stable marriage, stable cohabitation, married then divorcing, cohabiting then separating, varied transitions, and not with father. In fully adjusted models and compared with women whose mothers belonged to the stable marriage cluster: (1) women whose mothers belonged to the other marital status clusters had higher odds of overweight/obesity (odds ratio (OR) ranging 1.15-1.19; p < 0.05); and (2) women whose mothers belonged to the stable cohabitation (OR = 1.31; 95% confidence interval (CI) = 1.14-1.52), cohabiting then separating (OR = 1.23; 95% CI = 1.01-1.49), varied transitions (OR = 1.24; 95% CI = 1.11-1.39), and not with father (OR = 1.24; 95% CI = 1.00-1.54) clusters had higher odds of obesity.

CONCLUSIONS: Women whose mothers were not in stable marriage relationships had higher odds of being overweight or obese in adulthood. The finding that even women raised in the context of stable cohabitation had higher odds of being overweight or obese is intriguing as these relationships are socially accepted in Sweden.

Keywords
family structure, marital status, overweight, obesity, Sweden, sequence analysis
National Category
Public Health, Global Health, Social Medicine and Epidemiology
Identifiers
urn:nbn:se:umu:diva-135030 (URN)10.1177/1403494817705997 (DOI)000404652000007 ()28482752 (PubMedID)881251 (Local ID)881251 (Archive number)881251 (OAI)
Available from: 2017-05-16 Created: 2017-05-16 Last updated: 2019-02-15Bibliographically approved
Persson, E., Häggström, J., Waernbaum, I. & de Luna, X. (2017). Data-driven algorithms for dimension reduction in causal inference. Computational Statistics & Data Analysis, 105, 280-292
Open this publication in new window or tab >>Data-driven algorithms for dimension reduction in causal inference
2017 (English)In: Computational Statistics & Data Analysis, ISSN 0167-9473, E-ISSN 1872-7352, Vol. 105, p. 280-292Article in journal (Refereed) Published
Abstract [en]

In observational studies, the causal effect of a treatment may be confounded with variables that are related to both the treatment and the outcome of interest. In order to identify a causal effect, such studies often rely on the unconfoundedness assumption, i.e., that all confounding variables are observed. The choice of covariates to control for, which is primarily based on subject matter knowledge, may result in a large covariate vector in the attempt to ensure that unconfoundedness holds. However, including redundant covariates can affect bias and efficiency of nonparametric causal effect estimators, e.g., due to the curse of dimensionality. In this paper, data-driven algo- rithms for the selection of sufficient covariate subsets are investigated. Under the assumption of unconfoundedness we search for minimal subsets of the covariate vector. Based on the framework of sufficient dimension reduction or kernel smoothing, the algorithms perform a backward elim- ination procedure testing the significance of each covariate. Their performance is evaluated in simulations and an application using data from the Swedish Childhood Diabetes Register is also presented.

Keywords
covariate selection, marginal co-ordinate hypothesis test, matching, kernel smoothing, type 1 diabetes mellitus
National Category
Probability Theory and Statistics
Research subject
Statistics
Identifiers
urn:nbn:se:umu:diva-80696 (URN)10.1016/j.csda.2016.08.012 (DOI)000385604500019 ()
Funder
Swedish National Infrastructure for Computing (SNIC), SNIC 2016/1-2Swedish Research Council, 2013-672Swedish Research Council, 07531Riksbankens Jubileumsfond, P11-0814:1
Available from: 2013-09-24 Created: 2013-09-24 Last updated: 2018-06-08Bibliographically approved
Gorbach, T., Pudas, S., Lundquist, A., Orädd, G., Josefsson, M., Salami, A., . . . Nyberg, L. (2017). Longitudinal association between hippocampus atrophy and episodic-memory decline. Neurobiology of Aging, 51, 167-176
Open this publication in new window or tab >>Longitudinal association between hippocampus atrophy and episodic-memory decline
Show others...
2017 (English)In: Neurobiology of Aging, ISSN 0197-4580, E-ISSN 1558-1497, Vol. 51, p. 167-176Article in journal (Refereed) Published
Abstract [en]

There is marked variability in both onset and rate of episodic-memory decline in aging. Structural magnetic resonance imaging studies have revealed that the extent of age-related brain changes varies markedly across individuals. Past studies of whether regional atrophy accounts for episodic-memory decline in aging have yielded inconclusive findings. Here we related 15-year changes in episodic memory to 4-year changes in cortical and subcortical gray matter volume and in white-matter connectivity and lesions. In addition, changes in word fluency, fluid IQ (Block Design), and processing speed were estimated and related to structural brain changes. Significant negative change over time was observed for all cognitive and brain measures. A robust brain-cognition change-change association was observed for episodic-memory decline and atrophy in the hippocampus. This association was significant for older (65-80 years) but not middle-aged (55-60 years) participants and not sensitive to the assumption of ignorable attrition. Thus, these longitudinal findings highlight medial-temporal lobe system integrity as particularly crucial for maintaining episodic-memory functioning in older age. 

Keywords
Aging, cognitive decline, episodic memory, hippocampus, longitudinal changes, non-ignorable attrition
National Category
Probability Theory and Statistics Neurosciences
Identifiers
urn:nbn:se:umu:diva-128725 (URN)10.1016/j.neurobiolaging.2016.12.002 (DOI)000397168600018 ()28089351 (PubMedID)
Funder
Swedish Research CouncilKnut and Alice Wallenberg FoundationRagnar Söderbergs stiftelse
Available from: 2016-12-15 Created: 2016-12-13 Last updated: 2019-01-25Bibliographically approved
Projects
Longitudinal studies of cognitive aging: Multivariate and fMRI outcomes with non-ignorable dropout [2012-05931_VR]; Umeå UniversityStatistical models and methods to study life trajectories in the labour market and health domains [2016-02851_VR]; Umeå UniversityThe Umeå SIMSAM Lab - Infrastructure for Microdata Research from Childhood into Lifelong Health and Welfare [IN16-0368:1_RJ]; Umeå University
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0003-3187-1987

Search in DiVA

Show all publications