scholarly journals Clustering repeated ordinal data: Model based approaches using finite mixtures

2021 ◽  
Author(s):  
◽  
Roy Ken Costilla Monteagudo
Keyword(s):  
Dirichlet Process ◽  
Ordinal Data ◽  
Expectation Maximization Algorithm ◽  
Information Criteria ◽  
Repeated Measurements ◽  
Finite Mixtures ◽  
Cross Sectional ◽  
Advantages And Disadvantages ◽  
Model Based ◽  
Self Reported Health Status

<p>Model based approaches to cluster continuous and cross-sectional data are abundant and well established. In contrast to that, equivalent approaches for repeated ordinal data are less common and an active area of research. In this dissertation, we propose several models to cluster repeated ordinal data using finite mixtures. In doing so, we explore several ways of incorporating the correlation due to the repeated measurements while taking into account the ordinal nature of the data.   In particular, we extend the Proportional Odds model to incorporate latent random effects and latent transitional terms. These two ways of incorporating the correlation are also known as parameter and data dependent models in the time-series literature. In contrast to most of the existing literature, our aim is classification and not parameter estimation. This is, to provide flexible and parsimonious ways to estimate latent populations and classification probabilities for repeated ordinal data.   We estimate the models using Frequentist (Expectation-Maximization algorithm) and Bayesian (Markov Chain Monte Carlo) inference methods and compare advantages and disadvantages of both approaches with simulated and real datasets. In order to compare models, we use several information criteria: AIC, BIC, DIC and WAIC, as well as a Bayesian Non-Parametric approach (Dirichlet Process Mixtures). With regards to the applications, we illustrate the models using self-reported health status in Australia (poor to excellent), life satisfaction in New Zealand (completely agree to completely disagree) and agreement with a reference genome of infant gut bacteria (equal, segregating and variant) from baby stool samples.</p>

scholarly journals Clustering repeated ordinal data: Model based approaches using finite mixtures

2021 ◽  
Author(s):  
◽  
Roy Ken Costilla Monteagudo
Keyword(s):  
Dirichlet Process ◽  
Ordinal Data ◽  
Expectation Maximization Algorithm ◽  
Information Criteria ◽  
Repeated Measurements ◽  
Finite Mixtures ◽  
Cross Sectional ◽  
Advantages And Disadvantages ◽  
Model Based ◽  
Self Reported Health Status

<p>Model based approaches to cluster continuous and cross-sectional data are abundant and well established. In contrast to that, equivalent approaches for repeated ordinal data are less common and an active area of research. In this dissertation, we propose several models to cluster repeated ordinal data using finite mixtures. In doing so, we explore several ways of incorporating the correlation due to the repeated measurements while taking into account the ordinal nature of the data.   In particular, we extend the Proportional Odds model to incorporate latent random effects and latent transitional terms. These two ways of incorporating the correlation are also known as parameter and data dependent models in the time-series literature. In contrast to most of the existing literature, our aim is classification and not parameter estimation. This is, to provide flexible and parsimonious ways to estimate latent populations and classification probabilities for repeated ordinal data.   We estimate the models using Frequentist (Expectation-Maximization algorithm) and Bayesian (Markov Chain Monte Carlo) inference methods and compare advantages and disadvantages of both approaches with simulated and real datasets. In order to compare models, we use several information criteria: AIC, BIC, DIC and WAIC, as well as a Bayesian Non-Parametric approach (Dirichlet Process Mixtures). With regards to the applications, we illustrate the models using self-reported health status in Australia (poor to excellent), life satisfaction in New Zealand (completely agree to completely disagree) and agreement with a reference genome of infant gut bacteria (equal, segregating and variant) from baby stool samples.</p>

Advantages and limitations of the endoscopic tympanoscopy a prospective study

2020 ◽  
Vol 1 (1) ◽  
pp. 012-017
Author(s):  
Najib Zouhair ◽  
Anass Chaouki ◽  
Amine M’khatri ◽  
Youssef Oukessou ◽  
Sami Rouadi ◽  
...  
Keyword(s):  
Middle Ear ◽  
Operating Time ◽  
Cross Sectional Study ◽  
Cross Sectional ◽  
Invasive Approach ◽  
Intraoperative Bleeding ◽  
Advantages And Disadvantages ◽  
Early Return ◽  
The Mean ◽  
Endoscopic Tympanoplasty

Tympanoplasty is one of the most performed procedures in ENT. The aggressiveness of its microscopic approach has led otologists to adopt the endoscopic approach as a less invasive alternative. The purpose of this work is to appreciate the advantages and disadvantages of this surgical technique. We conducted a prospective descriptive cross-sectional study on 20 interventions within the ENT department of August the 20th 1953 Hospital of Casablanca from April 2019 to June 2019. The average age of operated patients was 36.3 years. Perforations were unilateral in (71%) of the cases with a predominance of the anterior (29%) and subtotal (36%) locations. The tympanoplasties were performed by 3 different senior otologic surgeons, and were left in (57%). The mean operating time was (59.5 min) and the mean anesthesia duration was 75.1 min. Intraoperative vision allowed us to fully visualize the margins of all perforations (100%) and anatomical structures of the middle ear in almost all interventions. The first procedures carried out were filled with difficulties whose management of intraoperative bleeding was the main one in (42.8%) of the cases. (57%) procedures were described as easy. No complication was detected intraoperatively or immediately postoperatively. Endoscopic tympanoplasty has several advantages, including: Minimally invasive approach to the middle ear; panoramic perioperative vision; Gain of operating time; decrease in the duration of anesthesia; Valuable educational tool; postoperative comfort; Decrease in hospital stay and early return to daily activities; Better aesthetic rendering; cost and transportability. However, we also note a number of disadvantages of endoscopic tympanoplasty, particularly: performing the procedure with one hand; difficulty passing through the EAC; 2D vision that alters the perception of depth; management of intraoperative bleeding; fogging; learning curve.

scholarly journals Self-reported health status, treatment decision and survival in asymptomatic and symptomatic patients with aortic stenosis in a Western Norway population undergoing conservative treatment: a cross-sectional study with 18 months follow-up

BMJ Open ◽  
2017 ◽  
Vol 7 (8) ◽  
pp. e016489 ◽  
Author(s):  
Kjersti Oterhals ◽  
Rune Haaverstad ◽  
Jan Erik Nordrehaug ◽  
Geir Egil Eide ◽  
Tone M Norekvål
Keyword(s):  
Mental Health ◽  
Aortic Valve ◽  
Treatment Decision ◽  
Physical And Mental Health ◽  
Cross Sectional ◽  
Western Norway ◽  
Asymptomatic Patients ◽  
Reported Health ◽  
Self Reported Health Status

ObjectivesTo investigate symptoms and self-reported health of patients conservatively treated for aortic stenosis (AS) and to identify factors associated with treatment decision and patient outcomes.DesignA cross-sectional survey with an 18-month follow-up.SettingOne tertiary university hospital in Western Norway.ParticipantsIn all, 1436 patients were diagnosed with AS between 2000 and 2012, and those 245 still under conservative treatment in 2013 were included in this study.Primary and secondary outcome measuresPrimary outcome measures were symptoms and self-reported health status. Secondary outcomes were treatment decision and patient survival after 18 months.ResultsA total of 136 patients with mean (SD) age 79 (12) years, 52% men responded. Among conservatively treated patients 77% were symptomatic. The symptom most frequently experienced was dyspnoea. Symptomatic patients reported worse physical and mental health compared with asymptomatic patients (effect size 1.24 and 0.74, respectively). In addition, symptomatic patients reported significantly higher levels of anxiety and depression compared with asymptomatic patients. However, symptom status did not correlate with haemodynamic severity of AS. After 18 months, 117 (86%) were still alive, 20% had undergone surgical aortic valve replacement (AVR) and 7% transcatheter aortic valve implantation (TAVI). When adjusting for age, gender, symptomatic status, severity of AS and European system for cardiac operative risk evaluation (EuroSCORE), patients with severe AS had more than sixfold chance of being scheduled for AVR or TAVI compared with those with moderate AS (HR 6.3, 95% CI 1.9 to 21.2, p=0.003). Patients with EuroSCORE ≥11 had less chance for undergoing AVR or TAVI compared with those with EuroSCORE ≤5 (HR 0.06, 95% CI 0.01 to 0.46, p=0.007).ConclusionsSymptoms affected both physical and mental health in conservatively treated patients with AS. Many patients with symptomatic severe AS are not scheduled for surgery, despite the recommendations in current guidelines. The referral practice for AVR is a path for further investigation.

scholarly journals ‘At-risk’ individuals’ responses to direct to consumer advertising of prescription drugs: a nationally representative cross-sectional study

BMJ Open ◽  
2017 ◽  
Vol 7 (12) ◽  
pp. e017865 ◽  
Author(s):  
Neda Khalil Zadeh ◽  
Kirsten Robertson ◽  
James A Green
Keyword(s):  
At Risk ◽  
Health Status ◽  
Prescription Drugs ◽  
Cross Sectional ◽  
Drug Advertising ◽  
Behavioural Responses ◽  
Reported Health ◽  
Nationally Representative ◽  
Self Reported Health Status ◽  
Direct To Consumer Advertising

ObjectivesThe factors determining individuals’ self-reported behavioural responses to direct to consumer advertising of prescription drugs were explored with an emphasis on ‘at-risk’ individuals’ responses.DesignNationally representative cross-sectional survey.SettingCommunity living adults in New Zealand.Participants2057 adults (51% women).Primary outcome measuresSelf-reported behavioural responses to drug advertising (asking a physician for a prescription, asking a physician for more information about an illness, searching the internet for more information regarding an illness and asking a pharmacist for more information about a drug).MethodsMultivariate logistic regressions determined whether participants’ self-reported behavioural responses to drug advertising were predicted by attitudes towards advertising and drug advertising, judgements about safety and effectiveness of advertised drugs, self-reported health status, materialism, online search behaviour as well as demographic variables.ResultsIdentifying as Indian and to a less extent Chinese, Māori and ‘other’ ethnicities were the strongest predictors of one or more self-reported responses (ORs 1.76–5.00, Ps<0.05). Poorer self-reported health status (ORs 0.90–0.94, all Ps<0.05), favourable attitude towards drug advertising (ORs 1.34–1.61, all Ps<0.001) and searching for medical information online (ORs 1.32–2.35, all Ps<0.01) predicted all self-reported behavioural outcomes. Older age (ORs 1.01–1.02, Ps<0.01), less education (OR 0.89, P<0.01), lower income (ORs 0.89–0.91, Ps<0.05) and higher materialism (ORs 1.02–1.03, Ps<0.01) also predicted one or more self-reported responses.ConclusionsTaken together, the findings suggest individuals, especially those who are ‘at-risk’ (ie, with poorer self-reported health status, older, less educated, lower income and ethnic minorities), may be more vulnerable to drug advertising and may make uninformed decisions accordingly. The outcomes raise significant concerns relating to the ethicality of drug advertising and suggest a need for stricter guidelines to ensure that drug advertisements provided by pharmaceutical companies are ethical.

scholarly journals Protocol for the avatar acceptability study: a multiperspective cross-sectional study evaluating the acceptability of using patient-derived xenografts to guide personalised cancer care in Australia and New Zealand

BMJ Open ◽  
2018 ◽  
Vol 8 (8) ◽  
pp. e024064 ◽  
Author(s):  
Claire E Wakefield ◽  
Emma L Doolan ◽  
Joanna E Fardell ◽  
Christina Signorelli ◽  
Veronica F Quinn ◽  
...  
Keyword(s):  
Cancer Survivors ◽  
Cancer Care ◽  
New Technologies ◽  
Cross Sectional Study ◽  
Decisional Balance ◽  
Local Health ◽  
Sectional Study ◽  
Structured Interviews ◽  
Cross Sectional ◽  
Advantages And Disadvantages

IntroductionPatient-derived xenografts (PDXs) have the potential to transform personalised cancer care, however, little is known about the acceptability of using PDXs to guide treatment decision-making. Given that patient and community preferences can influence satisfaction with care as well as the success of new technologies, we will evaluate the acceptability of PDXs in individuals affected by cancer and community comparisons.Methods and analysisThis comparative cross-sectional study will recruit 323 individuals affected by cancer (cancer survivors (of childhood or adult cancer) and parents of childhood cancer survivors) and 323 community comparisons (adults and parents). We will collect data via structured interviews and questionnaires. To determine the acceptability of PDXs, we will assess five domains: willingness to use PDXs when/if diagnosed with cancer, perceived advantages and disadvantages of PDXs, maximum acceptable out-of-pocket costs per patient, maximum acceptable turnaround time to receive results and maximum acceptable number of mice sacrificed per patient. The primary endpoint will be participants’ decisional balance ratio (calculated as participants’ advantages ratings divided by perceived disadvantages ratings).Ethics and disseminationThe study protocol has been approved by the South Eastern Sydney Local Health District Human Research Ethics Committee (HREC:12/173) and UNSW Sydney (HC15773). The results will be disseminated in peer-reviewed journals and at scientific conferences. A lay summary will be published on the Behavioural Sciences Unit website.

Cardiac and Non-Cardiac Determinants of Exercise Capacity in Chronic Kidney Disease

2021 ◽  
pp. ASN.2020091319
Author(s):  
Shanmugakumar Chinnappa ◽  
Nigel Lewis ◽  
Omer Baldo ◽  
Ming-Chieh Shih ◽  
Yu-Kang Tu ◽  
...  
Keyword(s):  
Heart Failure ◽  
Cardiac Output ◽  
Exercise Capacity ◽  
Significant Role ◽  
Exercise Physiology ◽  
Cardiopulmonary Exercise Test ◽  
Information Criteria ◽  
Healthy Controls ◽  
Cross Sectional ◽  
Patients With Heart Failure

Background Impaired exercise capacity is a significant symptom of CKD and is associated with poor survival. Furthermore, there is a growing interest in applying exercise as a diagnostic tool or as therapy in CKD. However, an in-depth understanding of exercise physiology in CKD is still lacking. Methods To evaluate the role of cardiac (central) and noncardiac (peripheral) determinants of exercise capacity in CKD, we conducted a cross-sectional study of 70 male CKD patients (stages 2-5) without diabetes or cardiac disease, 35 healthy controls, and 25 patients with heart failure. An integrated cardiopulmonary exercise test using a CO2 rebreathing technique was used to measure peak O2 consumption (VO2peak) and peak cardiac output simultaneously, and to calculate peak peripheral O2 extraction [C(a-v)O2], the peripheral determinant (the ability of exercising skeletal muscles to extract oxygen). We performed multiple regression analysis and used Bayesian information criteria (BIC) changes to quantitatively assess the individual contribution of central and peripheral factors. Results Compared with healthy controls, in patients with CKD, the VO2peak was impaired proportionate to its severity. Peak cardiac output was the predominant determinant of VO2peak in healthy controls and patients with heart failure, whereas C(a-v)O2 played a more significant role in determining VO2peak in CKD (β=0.68, P<0.001) compared with cardiac output (β=0.63, P<0.001). In addition, the magnitude of BIC reduction was greater for C(a v)O2 compared with cardiac output (BIC, 298.72 versus 287.68) in CKD. Conclusions In CKD, both peak cardiac output and peak C(a-v)O2 are independent predictors of VO2peak, and the more significant role played by peak C(a-v)O2 highlights the importance of noncardiac factors in determining exercise capacity in CKD.

Sign in / Sign up

Export Citation Format

Share Document