Please use this identifier to cite or link to this item: https://hdl.handle.net/2440/119075
Type: Thesis
Title: Surveillance of health status and health risk: The future of data collection using the telephone in Australia
Author: Dal Grande, Eleonora
Issue Date: 2017
School/Discipline: Adelaide Medical School
Abstract: Epidemiologically-designed, continuous and effective chronic disease and behavioural risk factor surveillance systems provide scientific evidence at the local level to assist government, health professionals and administrators, to respond effectively in reducing the burden associated with non-communicable diseases (NCDs). Many monitoring and surveillance systems utilise the telephone as the method of choice in obtaining population data. However, the dramatic change in telecommunication usage, diminishing coverage of telephone sampling frames and declining participation in household surveys, has led to methodological and statistical challenges. This has led to the present study that explores these challenges through an established telephone data collection system in Australia, the South Australian Monitoring and Surveillance System (SAMSS). The aim of this research is to determine how telephone surveys in Australia can continue to be used to reliably collect representative information on health indicators and other related health issues by exploring alternative efficient and cost effective methods. The first study, using face-to-face South Australian household survey data, found that using landline-based telephone number sampling frames excludes mobile-only households in Australia (27.8% of households in 2013). From 2006 to 2013, the proportion of mobile-only households has increased and this trend does not appear to be plateauing. This corresponds with the decrease in landline telephone coverage. Mobile-only households are demographically different in that respondents are more likely to be younger, never married and living in rented accommodation. By excluding this group, landline-based sampling frames may possibly produce biased health estimates for some health indicators, such as the proportion of people who are current smokers or who have a mental health condition. The second study found participation in SAMSS has decreased over a period of twelve years, with an 18.6% decrease in the response rate (from 68.9% in 2002 to 56.1% in 2014) and a 65.5% increase in the refusal rate. When demographic data are compared to Census data, SAMSS had a higher proportion of females, older people and people who rent, and these groups are increasingly being over-represented over time. The result from these studies imply that a mobile telephone sample needs to be incorporated. Unfortunately, there is no complete mobile telephone sampling frame in Australia with a geographical marker and only 7% of the currently used nationwide mobile telephone sampling frames are South Australian residents, making the sampling method uneconomic. This is compounded by lower participation in mobile telephone surveys compared to landline telephone surveys. Based on these methodological issues and corresponding with decline in participation, efficient methodological strategies need to be considered for smaller states like South Australia. The last two studies present two different cost effective and efficient methodological techniques, to minimise bias in health estimates due to nonresponse and sample coverage, and to increase participation in mobile telephone surveys. One study used raked weighting methodology to overcome, to some extent, the nonresponse biases and sampling coverage problems associated with telephone surveys. By incorporating more sociodemographic variables such as renting and marital status, besides the usual age, sex and area of residence, health estimates such as the proportion of current smokers corresponds well with other more expensive face-to-face surveys. The last study used a simple novel technique of sending a text message to prospective survey respondents to improve participation. This thesis has explored and shown, from a series of studies, that telephone surveys, with careful monitoring of procedures and use of innovative techniques and statistical methods, can still be used to collect and report information on chronic diseases and behavioural risk factors in Australia. The uniqueness of this body of works presents a detailed examination of the status of a current surveillance system by nonresponse rates, trends of nonresponse rates and coverage biases, and links this information to possible solutions to overcome nonresponse biases, with the aim of producing reliable and representative health estimates.
Advisor: Taylor, Anne
Dissertation Note: Thesis (Ph.D.) -- University of Adelaide, Adelaide Medical School, 2016
Keywords: Telephone sampling methodology
public health surveillance
raked weighting
nonresponse
representativeness
text messages
Provenance: This electronic version is made publicly available by the University of Adelaide in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. This thesis may incorporate third party material which has been used by the author pursuant to Fair Dealing exceptions. If you are the owner of any included third party copyright material you wish to be removed from this electronic version, please complete the take down form located at: http://www.adelaide.edu.au/legals
Appears in Collections:Research Theses

Files in This Item:
File Description SizeFormat 
Dal Grande2016_PhD.pdf3.59 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.