1
|
Zhu ZY, Shan HH, Wang J, Zhu HJ, Liu SG, Lin F. Graph modeling of relational structures among functioning variables with low back pain: an exploratory analysis based on International Classification of Functioning, Disability and Health. Eur J Phys Rehabil Med 2024; 60:487-495. [PMID: 38551517 PMCID: PMC11258909 DOI: 10.23736/s1973-9087.24.08089-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/14/2023] [Revised: 10/02/2023] [Accepted: 03/08/2024] [Indexed: 07/12/2024]
Abstract
BACKGROUND Given the complex etiology, multidimensional impact, and widespread prevalence of low back pain (LBP), it is crucial to prioritize intervention targets based on understanding the relationships between functional impairments in patients. This prioritization maximizes the physical and psychological benefits for patients, and graph modeling holds promise in achieving these objectives. AIM The aim of this study was establishing a graphical model of functioning variables for LBP based on the International Classification of Functioning, Disability, and Health (ICF) to identify the most influential items (i.e., functioning variables) on the physical and mental well-being of patients. Exploring feasible intervention measures by understanding the dysfunction correlations among these variables. DESIGN Cross-sectional survey. SETTING Nine hospitals in Jiangsu Province, China. POPULATION Three hundred and six persons with LBP aged ≥18 years. METHODS All patients were assessed using the Comprehensive ICF Core Sets for LBP. The scoring system was converted to dichotomous data, with 1 indicating dysfunction and 0 indicating no dysfunction. In the graphical model, network parameters and the results of Item Response Theory modeling (as detailed in our other article) were used to determine the importance of items, while partial correlations were utilized to estimate the dysfunction correlations between functioning variables. RESULTS 1) A total of 56 ICF items were located in the backbone structure of LBP, among which d430 (Lifting and carrying objects) occupied the most central position, followed by b126 (Temperament and personality functions). 2) In the main component of backbone structure, d430 has moderate dysfunction correlation with looking after one's health (0.6027), social norms, practices and ideologies (0.597), stability of joint functions (0.5759), and emotional functions (0.4078). b126 has moderate dysfunction correlation with basic interpersonal interactions (0.6595). CONCLUSIONS d430 and b126 significantly impact the physical and mental well-being of LBP patients. To improve d430, maintaining exercise habits, reducing working hours, enhancing lumbar stability, and overcoming fear-related emotions are recommended. Similarly, improving b126 can be achieved through enhancing interpersonal relationships. CLINICAL REHABILITATION IMPACT Through the identification of crucial functioning variables and the associated dysfunctional correlation relationships, graphical model of Comprehensive ICF Core Set for LBP can offer healthcare decision-makers valuable insights into potential treatment targets and pathways aimed at improving the condition of LBP patients.
Collapse
Affiliation(s)
- Zi-Yan Zhu
- School of Rehabilitation Medicine, Nanjing Medical University, Nanjing, China
- Department of Rehabilitation Medicine, Sir Run Run Hospital, Nanjing Medical University, Nanjing, China
| | - Hui-Hui Shan
- School of Rehabilitation Medicine, Nanjing Medical University, Nanjing, China
- Department of Rehabilitation Medicine, Sir Run Run Hospital, Nanjing Medical University, Nanjing, China
| | - Jin Wang
- School of Rehabilitation Medicine, Nanjing Medical University, Nanjing, China
- Department of Rehabilitation Medicine, The Affiliated Jiangsu Shengze Hospital of Nanjing Medical University, Suzhou, China
| | - Hong-Jun Zhu
- Department of Rehabilitation Medicine, The First Affiliated Hospital of Soochow University, Suzhou, China
| | - Shou-Guo Liu
- Department of Rehabilitation Medicine, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China
| | - Feng Lin
- School of Rehabilitation Medicine, Nanjing Medical University, Nanjing, China -
- Department of Rehabilitation Medicine, Sir Run Run Hospital, Nanjing Medical University, Nanjing, China
- Department of Rehabilitation Medicine, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China
| |
Collapse
|
2
|
Faulks D, Scambler S, Daly B, Jamieson L, Hennequin M, Tsakos G. Measuring oral health-How can the International Classification of Functioning help? Community Dent Oral Epidemiol 2023; 51:153-164. [PMID: 35112389 DOI: 10.1111/cdoe.12732] [Citation(s) in RCA: 5] [Impact Index Per Article: 5.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/09/2021] [Revised: 01/07/2022] [Accepted: 01/24/2022] [Indexed: 11/28/2022]
Abstract
There is a need for a theoretically informed, contextualized approach to measuring oral health from a multidisciplinary perspective that goes beyond the commonly used clinical indices and sociodental measures. This commentary aims to discuss the potential for the WHO's International Classification of Functioning, Disability and Health (ICF) to provide a model for the development of indicators for oral health. It is suggested that the ICF might provide both a theoretical model and an operational classification for indicators of oral health. The ICF model states that human experience of physical, cognitive and social functioning is universal and, thus, can be described and qualified. Human function is given social and environmental context within the model at both an individual and population level. The ICF can not only capture data regarding oral health and function at the physiological level (e.g. chewing) but also at the social level (e.g. sharing meals). It is able not only to capture aspects of preventive behaviour (e.g. caring for teeth) but also aspects of social facilitation (e.g. economic self-sufficiency) or ability to fulfil a social role (e.g. remunerative employment). It also includes aspects of social environment, such as healthcare services or political, economic and legal systems. Case studies are given as examples of the potential use of the ICF in the oral health domain. Examples are also given of the first steps that have been made towards operationalization of the ICF in data collection and oral health research. The challenges of encompassing such a comprehensive model into a practical oral health measure are discussed.
Collapse
Affiliation(s)
- Denise Faulks
- Service d'Odontologie, CHU Clermont-Ferrand, Clermont-Ferrand, France
- Centre de Recherche en Odontologie Clinique (CROC) EA4847, Université Clermont Auvergne, UFR d'Odontologie, Clermont-Ferrand, France
| | - Sasha Scambler
- Faculty of Dentistry, Oral and Craniofacial Sciences, King's College London, Tower Wing, Guy's Campus, London, UK
| | - Blánaid Daly
- Trinity College Dublin, School of Dental Sciences and Dublin Dental University Hospital, Dublin 2, Ireland
| | - Lisa Jamieson
- Australian Research Centre for Population Oral Health, The University of Adelaide, Adelaide, South Australia, Australia
| | - Martine Hennequin
- Service d'Odontologie, CHU Clermont-Ferrand, Clermont-Ferrand, France
- Centre de Recherche en Odontologie Clinique (CROC) EA4847, Université Clermont Auvergne, UFR d'Odontologie, Clermont-Ferrand, France
| | - Georgios Tsakos
- Department of Epidemiology and Public Health, UCL, London, UK
| |
Collapse
|
3
|
You M, Fang W, Wang X, Yang T. Modelling of the ICF core sets for chronic ischemic heart disease using the LASSO model in Chinese patients. Health Qual Life Outcomes 2018; 16:139. [PMID: 29996874 PMCID: PMC6042460 DOI: 10.1186/s12955-018-0957-0] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/11/2017] [Accepted: 06/13/2018] [Indexed: 11/30/2022] Open
Abstract
Background This study aimed to examine the associations among the International Classification of Functioning, Disability, and Health (ICF) core sets relevant to chronic ischemic heart disease (CIHD) using the least absolute shrinkage and selection operator (LASSO) model based on the ICF core sets scale in Chinese patients. Methods This was a prospective study of 120 patients with CIHD selected from January 2013 to June 2014 at the Fada Institute of Forensic Medicine & Science (Beijing, China). Functioning was qualified using the ICF core sets checklist for CIHD (Chinese version). The variables of core set categories of the ICF assessment scale for CIHD were entered into the LASSO model for mining dependencies among those variables. Graphical modeling was applied using LASSO generalized linear models. Results “Muscle endurance functions”, “sensations associated with cardiovascular and respiratory functions”, “blood vessel functions”, and “heart functions” were the most injured in CIHD status. “Recreation and leisure” and “intimate relationships” were the most affected in CIHD status. “General social support services, systems, and policies” and “acquaintances, peers, colleagues, neighbors, and community members” were important for the outcome of functional status of the CIHD patient. “Economic self-sufficiency” and “family relationships” of the CIHD patient were not undermined in most cases. Conclusions Graphical modeling can be used to describe associations between different areas of functioning in CIHD patients. The results suggest that these associations could be used as basis to improve rehabilitation and provide a deeper understanding of functioning in Chinese CIHD patients. Electronic supplementary material The online version of this article (10.1186/s12955-018-0957-0) contains supplementary material, which is available to authorized users.
Collapse
Affiliation(s)
- Meng You
- Collaborative Innovation Center of Judicial Civilization, Key Laboratory of Evidence Science, China University of Political Science and Law, Ministry of Education, 25 Xitucheng Road, Beijing, 100040, China
| | - Wen Fang
- Beijing Jiaotong University, Beijing, China
| | - Xu Wang
- Collaborative Innovation Center of Judicial Civilization, Key Laboratory of Evidence Science, China University of Political Science and Law, Ministry of Education, 25 Xitucheng Road, Beijing, 100040, China
| | - Tiantong Yang
- Collaborative Innovation Center of Judicial Civilization, Key Laboratory of Evidence Science, China University of Political Science and Law, Ministry of Education, 25 Xitucheng Road, Beijing, 100040, China.
| |
Collapse
|
4
|
Saal S, Meyer G, Beutner K, Klingshirn H, Strobl R, Grill E, Mann E, Köpke S, Bleijlevens MHC, Bartoszek G, Stephan AJ, Hirt J, Müller M. Development of a complex intervention to improve participation of nursing home residents with joint contractures: a mixed-method study. BMC Geriatr 2018; 18:61. [PMID: 29490617 PMCID: PMC5831216 DOI: 10.1186/s12877-018-0745-z] [Citation(s) in RCA: 10] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2017] [Accepted: 02/13/2018] [Indexed: 11/10/2022] Open
Abstract
BACKGROUND Joint contractures in nursing home residents limit the capacity to perform daily activities and restrict social participation. The purpose of this study was to develop a complex intervention to improve participation in nursing home residents with joint contractures. METHODS The development followed the UK Medical Research Council framework using a mixed-methods design with re-analysis of existing interview data using a graphic modelling approach, group discussions with nursing home residents, systematic review of intervention studies, structured 2-day workshop with experts in geriatric, nursing, and rehabilitation, and group discussion with professionals in nursing homes. RESULTS Graphic modelling identified restrictions in the use of transportation, walking within buildings, memory functions, and using the hands and arms as the central target points for the intervention. Seven group discussions with 33 residents revealed various aspects related to functioning and disability according the International Classification of Functioning, Disability and Health domains body functions, body structures, activities and participation, environmental factors, and personal factors. The systematic review included 17 studies with 992 participants: 16 randomised controlled trials and one controlled trial. The findings could not demonstrate any evidence in favour of an intervention. The structured 2-day expert workshop resulted in a variety of potential intervention components and implementation strategies. The group discussion with the professionals in nursing homes verified the feasibility of the components and the overall concept. The resulting intervention, Participation Enabling CAre in Nursing (PECAN), will be implemented during a 1-day workshop for nurses, a mentoring approach, and supportive material. The intervention addresses nurses and other staff, residents, their informal caregivers, therapists, and general practitioners. CONCLUSIONS In view of the absence of any robust evidence, the decision to use mixed methods and to closely involve both health professionals and residents proved to be an appropriate means to develop a complex intervention to improve participation of and quality of life in nursing home residents. We will now evaluate the PECAN intervention for its impact and feasibility in a pilot study in preparation for an evaluation of its effectiveness in a definitive trial. TRIAL REGISTRATION German clinical trials register, reference number DRKS00010037 (12 February 2016).
Collapse
Affiliation(s)
- Susanne Saal
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany.
| | - Gabriele Meyer
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany
| | - Katrin Beutner
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany
| | - Hanna Klingshirn
- Institute for Medical Information Processing, Biometry and Epidemiology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377, Munich, Germany
| | - Ralf Strobl
- Institute for Medical Information Processing, Biometry and Epidemiology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377, Munich, Germany
| | - Eva Grill
- Institute for Medical Information Processing, Biometry and Epidemiology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377, Munich, Germany.,German Center for Vertigo and Balance Disorders, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377, Munich, Germany
| | - Eva Mann
- Institute for General, Family and Preventive Medicine, Paracelsus Medical University, Strubergasse 21, 5020, Salzburg, Austria
| | - Sascha Köpke
- Institute of Social Medicine and Epidemiology, Nursing research group, University of Lübeck, Ratzeburger Allee 160, 23538, Lübeck, Germany
| | - Michel H C Bleijlevens
- CAPHRI Care and Public Health Research Institute, Department of Health Services Research, Living Lab on Ageing and Long-Term Care, Maastricht University, Universiteitssingel 40, 6229 ER, Maastricht, The Netherlands
| | - Gabriele Bartoszek
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany.,University of Applied Sciences for Social Work, Education and Nursing, Dürerstraße 25, 01307, Dresden, Germany
| | - Anna-Janina Stephan
- Institute for Medical Information Processing, Biometry and Epidemiology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377, Munich, Germany
| | - Julian Hirt
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany
| | - Martin Müller
- Institute of Health and Nursing Sciences, Medical Faculty, University of Halle-Wittenberg, Magdeburger Straße 8, 06112, Halle (Saale), Germany.,Faculty of Applied Health and Social Sciences, Rosenheim University of Applied Sciences, Hochschulstraße 1, 83024, Rosenheim, Germany
| |
Collapse
|
5
|
Ehrmann C, Bickenbach J, Stucki G. Graphical modelling: a tool for describing and understanding the functioning of people living with a health condition. Eur J Phys Rehabil Med 2017; 55:131-135. [PMID: 29144108 DOI: 10.23736/s1973-9087.17.04970-x] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/08/2022]
Abstract
Rehabilitation aims to optimize people's lived experience of health or functioning. A comprehensive understanding of people's functioning is thus fundamental for rehabilitation clinicians and scientists. Over the past ten years it has been shown that graphical modelling is a promising technique for modelling data on people's functioning. It can contribute to our understanding of the complex associations between domains of functioning and the identification of potential targets for rehabilitation interventions both at the level of the person and the environment. The objective of this methodological note is to demonstrate how graphical modelling can be used by rehabilitation clinicians and scientists in the description, understanding and influencing of people's functioning. The application of graphical modelling and the interpretation of results is illustrated using the Spinal Cord Injury Independence Measure - Self Report used in the Swiss Spinal Cord Injury Cohort Study. Finally, we discuss the potential of graphical modelling for the planning of studies that expand our understanding of functioning and for rehabilitation interventions.
Collapse
Affiliation(s)
- Cristina Ehrmann
- Department of Health Sciences and Health Policy, Faculty of Humanities and Social Sciences, University of Lucerne, Lucerne, Switzerland - .,Swiss Paraplegic Research (SPF), Nottwil, Switzerland -
| | - Jerome Bickenbach
- Department of Health Sciences and Health Policy, Faculty of Humanities and Social Sciences, University of Lucerne, Lucerne, Switzerland.,Swiss Paraplegic Research (SPF), Nottwil, Switzerland.,ICF Research Branch, a cooperation partner within the WHO Collaborating Center for the Family of International Classifications in Germany (at DIMDI), Nottwil, Switzerland
| | - Gerold Stucki
- Department of Health Sciences and Health Policy, Faculty of Humanities and Social Sciences, University of Lucerne, Lucerne, Switzerland.,Swiss Paraplegic Research (SPF), Nottwil, Switzerland.,ICF Research Branch, a cooperation partner within the WHO Collaborating Center for the Family of International Classifications in Germany (at DIMDI), Nottwil, Switzerland
| |
Collapse
|
6
|
Dong Y, Zhang CJ, Shi J, Deng J, Lan CN. Clinical application of ICF key codes to evaluate patients with dysphagia following stroke. Medicine (Baltimore) 2016; 95:e4479. [PMID: 27661012 PMCID: PMC5044882 DOI: 10.1097/md.0000000000004479] [Citation(s) in RCA: 6] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/26/2022] Open
Abstract
This study was aimed to identify and evaluate the International Classification of Functioning (ICF) key codes for dysphagia in stroke patients. Thirty patients with dysphagia after stroke were enrolled in our study. To evaluate the ICF dysphagia scale, 6 scales were used as comparisons, namely the Barthel Index (BI), Repetitive Saliva Swallowing Test (RSST), Kubota Water Swallowing Test (KWST), Frenchay Dysarthria Assessment, Mini-Mental State Examination (MMSE), and the Montreal Cognitive Assessment (MoCA). Multiple regression analysis was performed to quantitate the relationship between the ICF scale and the other 7 scales. In addition, 60 ICF scales were analyzed by the least absolute shrinkage and selection operator (LASSO) method. A total of 21 ICF codes were identified, which were closely related with the other scales. These included 13 codes from Body Function, 1 from Body Structure, 3 from Activities and Participation, and 4 from Environmental Factors. A topographic network map with 30 ICF key codes was also generated to visualize their relationships. The number of ICF codes identified is in line with other well-established evaluation methods. The network topographic map generated here could be used as an instruction tool in future evaluations. We also found that attention functions and biting were critical codes of these scales, and could be used as treatment targets.
Collapse
Affiliation(s)
- Yi Dong
- Mawangdui Hospital of Hunan Province
| | - Chang-Jie Zhang
- Department of Rehabilitation, The Second Xiangya Hospital of Central South University, Changsha, China
| | - Jie Shi
- Department of Rehabilitation, The Second Xiangya Hospital of Central South University, Changsha, China
| | | | - Chun-Na Lan
- Department of Rehabilitation, The Second Xiangya Hospital of Central South University, Changsha, China
- Correspondence: Chun-Na Lan, Department of Rehabilitation, The Second Xiangya Hospital of Central South University, No. 139 Renmin Road, Changsha 410011, China (e-mail: )
| |
Collapse
|
7
|
Kirschneck M, Sabariego C, Singer S, Tschiesner U. Assessment of functional outcomes in patients with head and neck cancer according to the International Classification of Functioning, Disability and Health Core Sets from the perspective of the multi-professional team: results of 4 Delphi surveys. Head Neck 2013; 36:954-68. [PMID: 23733325 DOI: 10.1002/hed.23399] [Citation(s) in RCA: 12] [Impact Index Per Article: 1.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/27/2012] [Revised: 04/11/2013] [Accepted: 05/23/2013] [Indexed: 11/07/2022] Open
Abstract
BACKGROUND The International Classification of Functioning, Disability and Health Core Set for Head and Neck Cancer (ICF-HNC) covers the typical spectrum of problems in functioning in head and neck cancer. This study is part of a multistep process to develop practical guidelines in Germany. The purpose of this study was to identify instruments for the assessment of functioning using the ICF-HNC as reference. METHODS Four Delphi surveys with physicians, physiotherapists, psychologists, and social workers were performed to identify which aspects of the ICF-HNC are being treated and which assessment tools are recommended for the assessment of functioning. RESULTS Ninety-seven percent categories of the ICF-HNC were treated by healthcare professionals participating in the current study. Altogether, 33 assessment tools were recommended for therapy monitoring, food intake, pain, further organic problems/laboratory tests, and psychosocial areas. CONCLUSION Although the ICF-HNC is being currently implemented by the head and neck cancer experts, several areas are not covered regularly. Additionally, validated tools were rarely recommended.
Collapse
Affiliation(s)
- Michaela Kirschneck
- Department of Medical Informatics, Biometry and Epidemiology - IBE, Chair for Public Health and Health Care Research, Ludwig-Maximilians-University, Munich, Germany
| | | | | | | |
Collapse
|
8
|
Baylor C, Yorkston K, Eadie T, Kim J, Chung H, Amtmann D. The Communicative Participation Item Bank (CPIB): item bank calibration and development of a disorder-generic short form. JOURNAL OF SPEECH, LANGUAGE, AND HEARING RESEARCH : JSLHR 2013; 56:1190-208. [PMID: 23816661 PMCID: PMC4377222 DOI: 10.1044/1092-4388(2012/12-0140)] [Citation(s) in RCA: 148] [Impact Index Per Article: 13.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 05/13/2023]
Abstract
PURPOSE The purpose of this study was to calibrate the items for the Communicative Participation Item Bank (CPIB; Baylor, Yorkston, Eadie, Miller, & Amtmann, 2009; Yorkston et al., 2008) using item response theory (IRT). One overriding objective was to examine whether the IRT item parameters would be consistent across different diagnostic groups, thereby allowing creation of a disorder-generic instrument. The intended outcomes were the final item bank and a short form ready for clinical and research applications. METHOD Self-report data were collected from 701 individuals representing 4 diagnoses: multiple sclerosis, Parkinson's disease, amyotrophic lateral sclerosis, and head and neck cancer. Participants completed the CPIB and additional self-report questionnaires. CPIB data were analyzed using the IRT graded response model. RESULTS The initial set of 94 candidate CPIB items were reduced to an item bank of 46 items demonstrating unidimensionality, local independence, good item fit, and good measurement precision. Differential item functioning analyses detected no meaningful differences across diagnostic groups. A 10-item, disorder-generic short form was generated. CONCLUSIONS The CPIB provides speech-language pathologists with a unidimensional, self-report outcomes measurement instrument dedicated to the construct of communicative participation. This instrument may be useful to clinicians and researchers wanting to implement measures of communicative participation in their work.
Collapse
|
9
|
Strobl R, Grill E, Mansmann U. Graphical modeling of binary data using the LASSO: a simulation study. BMC Med Res Methodol 2012; 12:16. [PMID: 22353192 PMCID: PMC3305667 DOI: 10.1186/1471-2288-12-16] [Citation(s) in RCA: 7] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/19/2011] [Accepted: 02/21/2012] [Indexed: 11/17/2022] Open
Abstract
Background Graphical models were identified as a promising new approach to modeling high-dimensional clinical data. They provided a probabilistic tool to display, analyze and visualize the net-like dependence structures by drawing a graph describing the conditional dependencies between the variables. Until now, the main focus of research was on building Gaussian graphical models for continuous multivariate data following a multivariate normal distribution. Satisfactory solutions for binary data were missing. We adapted the method of Meinshausen and Bühlmann to binary data and used the LASSO for logistic regression. Objective of this paper was to examine the performance of the Bolasso to the development of graphical models for high dimensional binary data. We hypothesized that the performance of Bolasso is superior to competing LASSO methods to identify graphical models. Methods We analyzed the Bolasso to derive graphical models in comparison with other LASSO based method. Model performance was assessed in a simulation study with random data generated via symmetric local logistic regression models and Gibbs sampling. Main outcome variables were the Structural Hamming Distance and the Youden Index. We applied the results of the simulation study to a real-life data with functioning data of patients having head and neck cancer. Results Bootstrap aggregating as incorporated in the Bolasso algorithm greatly improved the performance in higher sample sizes. The number of bootstraps did have minimal impact on performance. Bolasso performed reasonable well with a cutpoint of 0.90 and a small penalty term. Optimal prediction for Bolasso leads to very conservative models in comparison with AIC, BIC or cross-validated optimal penalty terms. Conclusions Bootstrap aggregating may improve variable selection if the underlying selection process is not too unstable due to small sample size and if one is mainly interested in reducing the false discovery rate. We propose using the Bolasso for graphical modeling in large sample sizes.
Collapse
Affiliation(s)
- Ralf Strobl
- Institute for Medical Informatics, Biometrics and Epidemiology, Ludwig-Maximilians-Universität München, Munich, Germany.
| | | | | |
Collapse
|