Dataset Information

Identifying patterns of item missing survey data using latent groups: an observational study.

ABSTRACT: OBJECTIVES:To examine whether respondents to a survey of health and physical activity and potential determinants could be grouped according to the questions they missed, known as 'item missing'. DESIGN:Observational study of longitudinal data. SETTING:Residents of Brisbane, Australia. PARTICIPANTS:6901 people aged 40-65 years in 2007. MATERIALS AND METHODS:We used a latent class model with a mixture of multinomial distributions and chose the number of classes using the Bayesian information criterion. We used logistic regression to examine if participants' characteristics were associated with their modal latent class. We used logistic regression to examine whether the amount of item missing in a survey predicted wave missing in the following survey. RESULTS:Four per cent of participants missed almost one-fifth of the questions, and this group missed more questions in the middle of the survey. Eighty-three per cent of participants completed almost every question, but had a relatively high missing probability for a question on sleep time, a question which had an inconsistent presentation compared with the rest of the survey. Participants who completed almost every question were generally younger and more educated. Participants who completed more questions were less likely to miss the next longitudinal wave. CONCLUSIONS:Examining patterns in item missing data has improved our understanding of how missing data were generated and has informed future survey design to help reduce missing data.

SUBMITTER: Barnett AG

PROVIDER: S-EPMC5665304 | biostudies-literature | 2017 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Identifying patterns of item missing survey data using latent groups: an observational study.

Barnett Adrian G AG McElwee Paul P Nathan Andrea A Burton Nicola W NW Turrell Gavin G

BMJ open 20171030 10

<h4>Objectives</h4>To examine whether respondents to a survey of health and physical activity and potential determinants could be grouped according to the questions they missed, known as 'item missing'.<h4>Design</h4>Observational study of longitudinal data.<h4>Setting</h4>Residents of Brisbane, Australia.<h4>Participants</h4>6901 people aged 40-65 years in 2007.<h4>Materials and methods</h4>We used a latent class model with a mixture of multinomial distributions and chose the number of classes ...[more]

PMID: 29084795

Similar Datasets

Project description:Recently, there has been interest in using the short message service (SMS or text messaging), to gather frequent information on the clinical course of individual patients. One possible role for identifying clinical course patterns is to assist in exploring clinically important subgroups in the outcomes of research studies. Two previous studies have investigated detailed clinical course patterns in SMS data obtained from people seeking care for low back pain. One used a visual analysis approach and the other performed a cluster analysis of SMS data that had first been transformed by spline analysis. However, cluster analysis of SMS data in its original untransformed form may be simpler and offer other advantages. Therefore, the aim of this study was to determine whether cluster analysis could be used for identifying clinical course patterns distinct from the pattern of the whole group, by including all SMS time points in their original form. It was a 'proof of concept' study to explore the potential, clinical relevance, strengths and weakness of such an approach.This was a secondary analysis of longitudinal SMS data collected in two randomised controlled trials conducted simultaneously from a single clinical population (n?=?322). Fortnightly SMS data collected over a year on 'days of problematic low back pain' and on 'days of sick leave' were analysed using Two-Step (probabilistic) Cluster Analysis.Clinical course patterns were identified that were clinically interpretable and different from those of the whole group. Similar patterns were obtained when the number of SMS time points was reduced to monthly. The advantages and disadvantages of this method were contrasted to that of first transforming SMS data by spline analysis.This study showed that clinical course patterns can be identified by cluster analysis using all SMS time points as cluster variables. This method is simple, intuitive and does not require a high level of statistical skill. However, there are alternative ways of managing SMS data and many different methods of cluster analysis. More research is needed, especially head-to-head studies, to identify which technique is best to use under what circumstances.

Project description:Background: Overuse of antibiotics significantly fuels the development of Antimicrobial resistance, which threating the global population health. Great variations existed in antibiotic prescribing practices among physicians, indicating improvement potential for rational use of antibiotics. This study aims to identify antibiotic prescribing patterns of primary care physicians and potential determinants. Methods: A cross-sectional survey was conducted on 551 physicians from 67 primary care facilities in Hubei selected through random cluster sampling, tapping into their knowledge, attitudes and prescribing practices toward antibiotics. Prescriptions (n = 501,072) made by the participants from 1 January to March 31, 2018 were extracted from the medical records system. Seven indicators were calculated for each prescriber: average number of medicines per prescription, average number of antibiotics per prescription, percentage of prescriptions containing antibiotics, percentage of antibiotic prescriptions containing broad-spectrum antibiotics, percentage of antibiotic prescriptions containing parenteral administered antibiotics, percentage of antibiotic prescriptions containing restricted antibiotics, and percentage of antibiotic prescriptions containing antibiotics included in the WHO "Watch and Reserve" list. Two-level latent profile analyses were performed to identify the antibiotic prescribing patterns of physicians based on those indicators. Multi-nominal logistic regression models were established to identify determinants with the antibiotic prescribing patterns. Results: On average, each primary care physician issued 909 (ranging from 100 to 11,941 with a median of 474) prescriptions over the study period. The mean percentage of prescriptions containing antibiotics issued by the physicians reached 52.19% (SD = 17.20%). Of those antibiotic prescriptions, an average of 82.29% (SD = 15.83%) contained broad-spectrum antibiotics; 71.92% (SD = 21.42%) contained parenteral administered antibiotics; 23.52% (SD = 19.12%) contained antibiotics restricted by the regional government; and 67.74% (SD = 20.98%) contained antibiotics listed in the WHO "Watch and Reserve" list. About 28.49% of the prescribers were identified as low antibiotic users, compared with 51.18% medium users and 20.33% high users. Higher use of antibiotics was associated with insufficient knowledge, indifference to changes, complacency with satisfied patients, low household income and rural location of the prescribers. Conclusion: Great variation in antibiotic prescribing patterns exists among primary care physicians in Hubei of China. High use of antibiotics is not only associated with knowledge shortfalls but also low socioeconomic status of prescribers.

Dataset Information

Identifying patterns of item missing survey data using latent groups: an observational study.

Publications

Identifying patterns of item missing survey data using latent groups: an observational study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets