Umeå University's logo

umu.sePublications
Change search
Link to record
Permanent link

Direct link
de Luna, Xavier, ProfessorORCID iD iconorcid.org/0000-0003-3187-1987
Alternative names
Publications (10 of 82) Show all publications
Ecker, K., de Luna, X. & Schelin, L. (2024). Causal inference with a functional outcome. The Journal of the Royal Statistical Society, Series C: Applied Statistics, 73(1), 221-240
Open this publication in new window or tab >>Causal inference with a functional outcome
2024 (English)In: The Journal of the Royal Statistical Society, Series C: Applied Statistics, ISSN 0035-9254, E-ISSN 1467-9876, Vol. 73, no 1, p. 221-240Article in journal (Refereed) Published
Abstract [en]

This article presents methods to study the causal effect of a binary treatment on a functional outcome with observational data. We define a Functional Average Treatment Effect (FATE) and develop an outcome regression estimator. We show how to obtain valid inference on the FATE using simultaneous confidence bands, which cover the FATE with a given probability over the entire domain. Simulation experiments illustrate how the simultaneous confidence bands take the multiple comparison problem into account. Finally, we use the methods to infer the effect of early adult location on subsequent income development for one Swedish birth cohort.

Place, publisher, year, edition, pages
Oxford University Press, 2024
Keywords
early adult location, functional average treatment effect, lifetime income trajectory, simultaneous confidence bands
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-216053 (URN)10.1093/jrsssc/qlad092 (DOI)001090448500001 ()2-s2.0-85182646724 (Scopus ID)
Funder
Swedish Research Council, 2016-02851Swedish Research Council, 2008-7491Riksbankens JubileumsfondUmeå University
Available from: 2023-11-01 Created: 2023-11-01 Last updated: 2024-02-13Bibliographically approved
Lestari, S. K., Eriksson, M., de Luna, X., Malmberg, G. & Ng, N. (2024). Volunteering and instrumental support during the first phase of the pandemic in Europe: the significance of COVID-19 exposure and stringent country’s COVID-19 policy. BMC Public Health, 24(1), Article ID 99.
Open this publication in new window or tab >>Volunteering and instrumental support during the first phase of the pandemic in Europe: the significance of COVID-19 exposure and stringent country’s COVID-19 policy
Show others...
2024 (English)In: BMC Public Health, E-ISSN 1471-2458, Vol. 24, no 1, article id 99Article in journal (Refereed) Published
Abstract [en]

Background: The COVID-19 control policies might negatively impact older adults’ participation in volunteer work, instrumental support provision, and the likelihood of receiving instrumental support. Studies that quantify changes in these activities and the related factors are limited. The current study aimed to examine the level of volunteering, instrumental support provision and receipt before and during the first phase of the COVID-19 pandemic in Europe and to determine whether older adults’ volunteering, instrumental support provision and receipt were associated with individual exposure to COVID-19 and the stringency of country’s COVID-19 control policy during the first phase of the COVID-19 pandemic.

Methods: A cross-sectional survey using data from the Survey of Health, Ageing and Retirement in Europe (SHARE) Corona Survey 1 was designed to focus on community-dwelling Europeans aged ≥50 years. History of participation in volunteering work and instrumental support provision or receipt was assessed from the previous SHARE Wave data. The country’s COVID-19 control policy stringency index (S-Index) was from the Oxford COVID-19 Government Response Tracker database. A total of 45,669 respondents from 26 European countries were included in the volunteering analysis. Seventeen European countries were included in the analyses of instrumental support provision (N = 36,518) and receipt (N = 36,526). The multilevel logistic regression model was fitted separately to analyse each activity.

Results: The level of volunteering and instrumental support provision was lower during the pandemic, but instrumental support receipt was higher. The country S-Index was positively associated with support provision (OR:1.13;95%CI:1.02–1.26) and negatively associated with support receipt (OR:0.69;95%CI:0.54–0.88). Exposure to COVID-19 was positively associated with support receipt (OR:1.64;95%CI:1.38–1.95). COVID-19 exposure on close ones positively associated with volunteering (OR:1.47;95%CI:1.32–1.65), support provision (OR:1.28;95%CI:1.19–1.39), and support receipt (OR:1.25;95%CI:1.15–1.35).

Conclusions: The COVID-19 pandemic impacted older Europeans’ volunteering, instrumental support provision, and instrumental support receipt from outside their household. When someone close to them was exposed to COVID-19, older Europeans were likely to receive instrumental support and to volunteer and provide instrumental support. A stricter country’s COVID-19 control policy might motivate older adults to provide instrumental support, but it prevents them from receiving instrumental support from outside their households. 

Place, publisher, year, edition, pages
BioMed Central (BMC), 2024
Keywords
s COVID-19, Social support, Social participation, Volunteering, Older population, SHARE, Europe, Ageing population
National Category
Public Health, Global Health, Social Medicine and Epidemiology
Identifiers
urn:nbn:se:umu:diva-200954 (URN)10.1186/s12889-023-17507-5 (DOI)2-s2.0-85181485748 (Scopus ID)
Funder
EU, Horizon 2020, 101015924
Note

Originally included in thesis in manuscript form.

Available from: 2022-11-11 Created: 2022-11-11 Last updated: 2024-01-24Bibliographically approved
Gorbach, T., de Luna, X., Waernbaum, I. & Karvanen, J. (2023). Contrasting identifying assumptions of average causal effects: robustness and semiparametric efficiency. Journal of machine learning research, 24(197), 1-65
Open this publication in new window or tab >>Contrasting identifying assumptions of average causal effects: robustness and semiparametric efficiency
2023 (English)In: Journal of machine learning research, ISSN 1532-4435, E-ISSN 1533-7928, Vol. 24, no 197, p. 1-65Article in journal (Refereed) Published
Abstract [en]

Semiparametric inference on average causal effects from observational data is based on assumptions yielding identification of the effects. In practice, several distinct identifying assumptions may be plausible; an analyst has to make a delicate choice between these models. In this paper, we study three identifying assumptions based on the potential outcome framework:  the back-door assumption, which uses pre-treatment covariates, the front-door assumption, which uses mediators, and the two-door assumption using pre-treatment covariates and mediators simultaneously. We provide the efficient influence functions and the corresponding semiparametric efficiency bounds that hold under these assumptions, and their combinations. We demonstrate that neither of the identification models provides uniformly the most efficient estimation and give conditions under which some bounds are lower than others. We show when semiparametric estimating equation estimators based on influence functions  attain the bounds, and study the robustness of the estimators to misspecification of the nuisance models. The theory is complemented with simulation experiments on the finite sample behavior of the estimators. The results obtained are relevant for an analyst facing a choice between several plausible identifying assumptions and corresponding estimators. Our results show that this choice implies a trade-off between efficiency and robustness to misspecification of the nuisance models. 

Keywords
causal inference, efficiency bound, robustness, back-door, front-door
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-190082 (URN)
Funder
Forte, Swedish Research Council for Health, Working Life and Welfare, 2018-00852Swedish Research Council, 2018-02670Swedish Research Council, 2016-00703Marianne and Marcus Wallenberg Foundation, 2015.0060Academy of Finland, 311877
Available from: 2021-12-03 Created: 2021-12-03 Last updated: 2023-09-04Bibliographically approved
Mohammad, G., Moosavi, N. & de Luna, X. (2023). Convolutional neural networks for valid and efficient causal inference. Journal of Computational And Graphical Statistics
Open this publication in new window or tab >>Convolutional neural networks for valid and efficient causal inference
2023 (English)In: Journal of Computational And Graphical Statistics, ISSN 1061-8600, E-ISSN 1537-2715Article in journal (Other academic) Epub ahead of print
Abstract [en]

Convolutional neural networks (CNN) have been successful in machine learning applications including image classification. When it comes to images, their success relies on their ability to consider the space invariant local features in the data. Here, we consider the use of CNN to fit nuisance models in semiparametric estimation of a one dimensional causal parameter: the average causal effect of a binary treatment. In this setting, nuisance models are functions of pre-treatment covariates that need to be controlled for. In an application where we want to estimate the effect of early retirement on a health outcome, we propose to use CNN to control for time-structured covariates. Thus, CNN is used when fitting nuisance models explaining the treatment assignment and the outcome. These fits are then combined into an augmented inverse probability weighting estimator yielding efficient and uniformly valid inference. Theoretically, we contribute by providing rates of convergence for CNN equipped with the rectified linear unit activation function and compare it to an existing result for feedforward neural networks. We also show when those rates guarantee uniformly valid inference for the proposed estimator. A Monte Carlo study is provided where the performance of the proposed estimator is evaluated and compared with other strategies. Finally, we give results on a study of the effect of early retirement on later hospitalization using a database covering the whole Swedish population.

Place, publisher, year, edition, pages
Taylor & Francis, 2023
Keywords
Average causal effect, augmented inverse probability weighting, early retirement, rate double robustness, post-machine learning inference
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-199235 (URN)10.1080/10618600.2023.2257247 (DOI)2-s2.0-85174611652 (Scopus ID)
Funder
Marianne and Marcus Wallenberg FoundationSwedish Research Council
Note

Orinally included in thesis in manuscript form. 

Available from: 2022-09-08 Created: 2022-09-08 Last updated: 2023-10-30
Edström, F., Hellström, T. & de Luna, X. (2023). Robot causal discovery aided by human interaction. In: 2023 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN): . Paper presented at IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), Busan, Korea, August 28-31, 2023 (pp. 1731-1736). IEEE
Open this publication in new window or tab >>Robot causal discovery aided by human interaction
2023 (English)In: 2023 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), IEEE, 2023, p. 1731-1736Conference paper, Published paper (Refereed)
Abstract [en]

Causality is relatively unexplored in robotics even if it is highly relevant, in several respects. In this paper, we study how a robot’s causal understanding can be improved by allowing the robot to ask humans causal questions. We propose a general algorithm for selecting direct causal effects to ask about, given a partial causal representation (using partially directed acyclic graphs, PDAGs) obtained from observational data. We propose three versions of the algorithm inspired by different causal discovery techniques, such as constraint-based, score-based, and interventions. We evaluate the versions in a simulation study and our results show that asking causal questions improves the causal representation over all simulated scenarios. Further, the results show that asking causal questions based on PDAGs discovered from data provides a significant improvement compared to asking questions at random, and the version inspired by score-based techniques performs particularly well over all simulated experiments.

Place, publisher, year, edition, pages
IEEE, 2023
Series
IEEE RO-MAN proceedings, ISSN 1944-9445, E-ISSN 1944-9437
Keywords
human-robot-interaction (hri), causal discovery, causal inference
National Category
Robotics Computer Sciences Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-219029 (URN)10.1109/RO-MAN57019.2023.10309376 (DOI)001108678600221 ()2-s2.0-85187012918 (Scopus ID)9798350336702 (ISBN)9798350336719 (ISBN)
Conference
IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), Busan, Korea, August 28-31, 2023
Funder
Swedish Research Council
Available from: 2024-01-05 Created: 2024-01-05 Last updated: 2024-03-18Bibliographically approved
Moosavi, N., Häggström, J. & de Luna, X. (2023). The costs and benefits of uniformly valid causal inference with high-dimensional nuisance parameters. Statistical Science, 38(1), 1-12
Open this publication in new window or tab >>The costs and benefits of uniformly valid causal inference with high-dimensional nuisance parameters
2023 (English)In: Statistical Science, ISSN 0883-4237, E-ISSN 2168-8745, Vol. 38, no 1, p. 1-12Article in journal (Refereed) Published
Abstract [en]

Important advances have recently been achieved in developing procedures yielding uniformly valid inference for a low dimensional causal parameter when high-dimensional nuisance models must be estimated. In this paper, we review the literature on uniformly valid causal inference and discuss the costs and benefits of using uniformly valid inference procedures. Naive estimation strategies based on regularisation, machine learning, or a preliminary model selection stage for the nuisance models have finite sample distributions which are badly approximated by their asymptotic distributions. To solve this serious problem, estimators which converge uniformly in distribution over a class of data generating mechanisms have been proposed in the literature. In order to obtain uniformly valid results in high-dimensional situations, sparsity conditions for the nuisance models need typically to be made, although a double robustness property holds, whereby if one of the nuisance model is more sparse, the other nuisance model is allowed to be less sparse. While uniformly valid inference is a highly desirable property, uniformly valid procedures pay a high price in terms of inflated variability. Our discussion of this dilemma is illustrated by the study of a double-selection outcome regression estimator, which we show is uniformly asymptotically unbiased, but is less variable than uniformly valid estimators in the numerical experiments conducted. 

Place, publisher, year, edition, pages
Institute of Mathematical Statistics, 2023
Keywords
Double robustness, Machine learning, Post-model selection inference, Regularization, Superefficiency
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-199231 (URN)10.1214/21-STS843 (DOI)000991879600001 ()2-s2.0-85152060424 (Scopus ID)
Funder
Marianne and Marcus Wallenberg Foundation
Note

Originally included in thesis in manuscript form.

Available from: 2022-09-08 Created: 2022-09-08 Last updated: 2023-09-05Bibliographically approved
Lee, S.-h., Ma, Y. & de Luna, X. (2022). Covariate balancing for causal inference on categorical and continuous treatments. Econometrics and Statistics
Open this publication in new window or tab >>Covariate balancing for causal inference on categorical and continuous treatments
2022 (English)In: Econometrics and Statistics, E-ISSN 2452-3062Article in journal (Refereed) In press
Abstract [en]

Novel estimators of causal effects for categorical and continuous treatments are proposed by using an optimal covariate balancing strategy for inverse probability weighting. The resulting estimators are shown to be consistent and asymptotically normal for causal contrasts of interest, either when the model explaining the treatment assignment is correctly specified, or when the correct set of bases for the outcome models has been chosen and the assignment model is sufficiently rich. For the categorical treatment case, the estimator attains the semiparametric efficiency bound when all models are correctly specified. For the continuous case, the causal parameter of interest is a function of the treatment dose. The latter is not parametrized and the estimators proposed are shown to have bias and variance of the classical nonparametric rate. Asymptotic results are complemented with simulations illustrating the finite sample properties. A data analysis suggests a nonlinear effect of BMI on self-reported health decline among the elderly.

Place, publisher, year, edition, pages
Elsevier, 2022
Keywords
Average causal effects, BMI, dose-response, double robust, self-reported health, semiparametric efficiency bound
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-193001 (URN)10.1016/j.ecosta.2022.01.007 (DOI)2-s2.0-85125446606 (Scopus ID)
Available from: 2022-03-15 Created: 2022-03-15 Last updated: 2024-04-26
Lestari, S. K., Eriksson, M., de Luna, X., Malmberg, G. & Ng, N. (2022). Frailty and types of social relationships among older adults in 17 European countries: A latent class analysis. Archives of gerontology and geriatrics (Print), 101, Article ID 104705.
Open this publication in new window or tab >>Frailty and types of social relationships among older adults in 17 European countries: A latent class analysis
Show others...
2022 (English)In: Archives of gerontology and geriatrics (Print), ISSN 0167-4943, E-ISSN 1872-6976, Vol. 101, article id 104705Article in journal (Refereed) Published
Abstract [en]

Background: Frailty is a syndrome commonly associated with old age. Social relationships are an essential determinant of frailty progression, and frailty can negatively affect social relationships.

Objectives: To identify social relationship types among older adults in Europe; to evaluate whether social relationship types differ across European regions; and to assess the association between frailty status and social relationship type.

Methods: We used data from 56,226 individuals from 17 European countries who participated in Wave 6 of the Survey of Health, Ageing and Retirement in Europe. We constructed social relationship types from social relationship variables (contacts frequency, perceived emotional support, participation in social activities, providing and receiving instrumental support) using latent class analysis (LCA). Associations between social relationship types and frailty were examined using multinomial regression analyses integrated with LCA.

Results: We identified four social relationship types: ‘poor’; ‘frequent and emotionally close’; ‘frequent, emotionally close, and supportive’; and ‘frequent, emotionally close, and active’. Type 3 is also characterised by participation in sport/social clubs (in the northern region) or receiving support (in the eastern region). Participation in volunteering/charity activities (in the central and northern regions) and instrumental support provision (in the northern region) are Type 4′s characteristics as well. In all regions, being frail was associated with less active social relationships (Types 1, 2, and 3) relative to the more ‘active’ type (Type 4).

Conclusion: Frailty status was associated with social relationship types. The identified types may help tailor intervention programmes for older adults to prevent worsening frailty.

Keywords
Frailty, Latent class analysis, Older age, SHARE, Social participation, Social support
National Category
Other Social Sciences not elsewhere specified Gerontology, specialising in Medical and Health Sciences
Identifiers
urn:nbn:se:umu:diva-194900 (URN)10.1016/j.archger.2022.104705 (DOI)000793742700007 ()35461166 (PubMedID)2-s2.0-85129513398 (Scopus ID)
Funder
Swedish Research Council, 2020-0254Forte, Swedish Research Council for Health, Working Life and Welfare, 2018-05196
Available from: 2022-06-01 Created: 2022-06-01 Last updated: 2023-03-24Bibliographically approved
Ecker, K., de Luna, X. & Westerlund, O. (2022). Regional differences in initial labour market conditions and dynamics in lifetime income trajectories. Longitudinal and Life Course Studies, 13(3), 352-379
Open this publication in new window or tab >>Regional differences in initial labour market conditions and dynamics in lifetime income trajectories
2022 (English)In: Longitudinal and Life Course Studies, E-ISSN 1757-9597, Vol. 13, no 3, p. 352-379Article in journal (Refereed) Published
Abstract [en]

We use longitudinal register data from Sweden to study patterns and dynamics in lifetime income trajectories. We examine divergences in these income trajectories by local economic conditions at labour market entry, in combination with other factors such as gender, education level and socio-economic background. We cannot assume that these relationships are constant over the course of individuals’ working lives. Therefore, we use methods from functional data analysis, allowing for a time-varying relationship between income and the explanatory variables. Our results show a large degree of heterogeneity in how lifetime income trajectories develop for different subgroups. We find that, for men, entering the labour market in an urban area is associated with higher cumulative lifetime income, especially later in life. The exception is men with only primary education, for whom those starting their working lives in a large city have lower incomes on average. This divergence increases in size over time. Women who enter into a large urban labour market receive higher lifetime income at all education levels. This relationship is strongest for women with primary education but decreases in strength over time for these women.

Place, publisher, year, edition, pages
Bristol University Press, 2022
Keywords
functional data analysis, lifetime income, local labour market, longitudinal data, urbanisation
National Category
Economics
Identifiers
urn:nbn:se:umu:diva-198263 (URN)10.1332/175795921X16427665823284 (DOI)000841657200002 ()35920642 (PubMedID)2-s2.0-85134214474 (Scopus ID)
Available from: 2022-08-02 Created: 2022-08-02 Last updated: 2024-01-23Bibliographically approved
Lestari, S. K., de Luna, X., Eriksson, M., Malmberg, G. & Ng, N. (2021). A longitudinal study on social support, social participation, and older Europeans' quality of life. SSM - Population Health, 13, Article ID 100747.
Open this publication in new window or tab >>A longitudinal study on social support, social participation, and older Europeans' quality of life
Show others...
2021 (English)In: SSM - Population Health, ISSN 2352-8273, Vol. 13, article id 100747Article in journal (Refereed) Published
Abstract [en]

The association between quality of life (QoL) and social relationships is well established. This paper further analyses whether and how participation in social activities as well as providing and receiving social support, independently, are associated with QoL among the older population in 16 European countries. QoL was measured using the CASP-12 scale. The baseline data came from Wave 6 and the outcome from Wave 7 of the Survey of Health, Ageing and Retirement in Europe (SHARE). The associations of interest were analysed using multivariable linear regression. The effect of possible non-ignorable dropout was tested. Then, doubly robust estimation and sensitivity analyses for unobserved confounding were performed to evaluate the possible causal interpretation of the associations found. Our findings show that participation in at least one of the socially productive activities was positively associated with QoL at two-year follow-up (Average Causal Effect, ACE: 0.474; 95%CI: 0.361, 0.587). The association was stronger among women, people aged 75+, and those in the Southern European region. Providing social support had a positive association with QoL, but only among people aged 75+ (ACE: 0.410; 95%CI: 0.031, 0.789). Conversely, receiving social support had a negative association (ACE: -0.321; 95%CI: -0.448, -0.195) with QoL, especially for men, people aged 75+, and those in Eastern European countries. Sensitivity analyses for unobserved confounders showed that the associations found cannot be attributed to causal effects.

Place, publisher, year, edition, pages
Elsevier, 2021
Keywords
Older population, Panel data, Quality of life, Social participation, Social support
National Category
Public Health, Global Health, Social Medicine and Epidemiology Gerontology, specialising in Medical and Health Sciences
Identifiers
urn:nbn:se:umu:diva-180792 (URN)10.1016/j.ssmph.2021.100747 (DOI)000636560000012 ()2-s2.0-85100764155 (Scopus ID)
Available from: 2021-02-25 Created: 2021-02-25 Last updated: 2023-09-05Bibliographically 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