Health Interview Survey 2004 Belgium Research Protocol Version 3: March 2005
Table of contents 1. GENERAL INFORMATION... 4 COORDINATES OF SPONSORS... 4 INSTITUTE IN CHARGE OF STUDY... 4 MANAGER... 4 STUDY DIRECTOR... 4 TEAM MEMBERS... 4 PROTOCOL REVIEW... 5 2. INTRODUCTION TO THE HEALTH INTERVIEW SURVEY... 6 BACKGROUND... 19 OBJECTIVES OF THE SURVEY... 19 THEORETICAL AND CONCEPTUAL BACKGROUND... 19 Health status... 19 Health determinants... 19 Medical prevention... 19 Health consumption... 19 Health and society... 19 ETHICAL CONSIDERATIONS... 19 Population surveys... 19 Informed consent... 19 Collection of Health Information... 19 Disclosure of Health Information... 19 3. DESCRIPTION OF THE SAMPLE... 19 TARGET POPULATION... 19 SAMPLING FRAME... 19 STUDY POPULATION... 19 SAMPLING DESIGN... 19 Introduction... 19 Sample Size... 19 Stratified Clustered Multi-stage design... 19 4. NETWORK OF INTERVIEWERS... 19 COMPOSING THE NETWORK... 19 Call for candidates... 19 Selection of candidates... 19 TRAINING OF INTERVIEWERS... 19 MANUAL FOR INTERVIEWERS... 19 5. FIELDWORK PROCEDURES... 20 SECRETARIAT... 20 THE ANNOUNCEMENT OF THE SURVEY... 20 CONTACT PROCEDURE... 21 SELECTION OF RESPONDENTS WITHIN THE HOUSEHOLD... 24 REPLACEMENT OF HOUSEHOLDS... 24 COMMUNICATION FORM... 25 RECEPTION FORM... 25 RECEPTION OF QUESTIONNAIRES... 25 FOLLOW-UP OF INTERVIEWERS... 26 Help-desk... 26 Newsletter... 26 Individual contacts... 26 MONITORING OF INTERVIEWERS... 26 2
6. QUESTIONNAIRES... 28 CONTENT OF QUESTIONNAIRES... 28 TARGET GROUPS AGE LIMITS... 29 USE OF PROXIES... 29 INTERPRETERS... 29 BUILDING OF QUESTIONNAIRES... 29 PRINTING AND DISTRIBUTION QUESTIONNAIRES... 29 7. DATA ENTRY... 31 VERTICAL CONTROL... 31 HORIZONTAL CONTROL... 31 8. DATA MANAGEMENT... 32 INTERVIEWER S DATABASE... 32 SAMPLE DATABASE... 32 CLUSTERED DATABASE... 32 FIELDWORK DATABASE... 33 RECEPTION DATABASE... 34 HEALTH SURVEY DATABASE... 34 9. PROCEDURE IMPLEMENTATION... 35 STRATIFICATION... 35 MULTI STAGE SAMPLING... 36 SYSTEMATIC SAMPLE... 37 CLUSTERING: SELECTION OF A SETS OF HOUSEHOLDS WITHIN EACH MUNICIPALITY... 37 SELECTION OF THE HOUSEHOLD MEMBERS... 41 10. DATA ANALYSIS... 43 11. REPORTING... 44 12. DISSEMINATION OF RESULTS... 45 GLOBAL REPORT... 45 PROVINCIAL REPORTS... 45 SUMMARY REPORT... 45 PRESS CONFERENCE... 46 COMMUNICATION WITH PARTICIPATING HOUSEHOLDS... 46 SCIENTIFIC CONTRIBUTIONS... 46 INTERNET SITE... 46 EXTERNAL USERS... 46 13. REFERENCES... 48 3
1. General information Coordinates of sponsors 1 Minister of Consumer Affairs, Public Health and Environment Federal Government Minister of Social Affairs and Pensions Federal Government Minister of Wellbeing, Health and Equal chances Flemish Government Minister of Social Affairs and Health Walloon Government Minister of Youth Support and Health French Community Minister of Youth and Family, Monuments and Sites, Health and Social Affairs German-speaking Community Ministers of Common Community Commission in charge of Health - Brussels Capital Region Represented in a Commission of Commissioners, chaired by Dr Pol Gerits Federal Public Service - Health, Security of the Food Chain and Environment Directorate-General Organisation of Health Care Provision Rijksadministratief Centrum Vesaliusgebouw Oratoriënberg 20 1010 Brussel Tel: 32 2 210 47 75 Fax : 32 2 210 64 00 pol.gerits@health.fgov.be Institute in charge of study Manager Scientific Institute of Public Health (IPH) Unit of Epidemiology J. Wytsmanstraat 14 1050 Brussels Prof. Dr. Herman Van Oyen, Head Unit of Epidemiology, IPH Tel: 32 2 642 50 29 Fax: 32 2 642 54 10 herman.vanoyen@iph.fgov.be Study Director Dr. J. Tafforeau, Unit of Epidemiology, IPH Tel: 32 2 642 57 33 Fax: 32 2 642 54 10 jean.tafforeau@iph.fgov.be Team members Mrs J. Buziarsist, Unit of Epidemiology, IPH Mr S. Demarest, Unit of Epidemiology, IPH Mrs L. Gisle, Unit of Epidemiology, IPH Dr E. Hesse, Unit of Epidemiology, IPH 1 See additional information in annex 1.1 and 1.2 4
Dr J. Van der Heyden, Unit of Epidemiology, IPH Protocol review MSc L. Bruckers, Center of Statistics, Limburgs Universitair Centrum Prof Dr G Molenberghs, Center of Statistics, Limburgs Universitair Centrum 5
2. Introduction to the Health Interview Survey Background Health information and research has been defined by the 43rd World Health Assembly as a process for obtaining systematic knowledge and technology that can be used for improvement of the health of the individuals or groups of population. Health information can thus be considered as one of the tools to be used for health promotion and disease prevention. Due to the lack of high quality and timely health data in Belgium it has been decided in the nineties to develop of a new tool aiming at gathering useful information for the decision makers when designing the public health programs. Several countries facing the same problem have successfully answered this by developing health interview surveys. The pioneering countries in this domain were Canada, Denmark, The Netherlands and the United Kingdom, where the health surveys progressively became the necessary supplement to routine information systems in order to develop consistent public health policies. Health surveys provide information on health as perceived by the population. The added value of the health survey is the horizontal approach of the data collection: several types of information (health status, health determinants, personal characteristics, health consumption,...) are collected simultaneously for the same person. The outcome is a global picture of the health of the population allowing identifying priority domains. In addition, because the data are gathered periodically overtime, changes in health and effects of health policies and interventions can be monitored. The health interview survey brings together the arguments for an increased investment in health promotion and prevention, and rationalisation of health care and expenditures. This information is thus a powerful framework for a rational policy decision-making process (1). Objectives of the survey The main objective of the Health Interview Survey (HIS) is to give a description of the health status of the population in Belgium in general and of the three regional subpopulations (Flemish, Walloon and Brussels region) in particular. The idea is to obtain information on how people experience their health, to what extent they make use of health care facilities, and how they look after their health by adopting a certain life style or relying on preventive and other health services. More specifically, the goals of the survey can be summarized as follows: identification of health problems; description of the health status and health needs of the population; estimation of prevalence and distribution of health indicators; analysis of social (in)equality in health and access to the health services; study of health consumption and its determinants; study of possible trends in the health status of the population. A health interview survey provides one possible channel through which such information can be obtained. On the basis of this survey, assessing a large variety of personal, social and material characteristics, life habits and conditions, determinants for public health can be traced and identified. The ultimate goal of the health interview survey is to be an integrated instrument in 6
decision making, while mapping out an adequate health policy. The aim is stating priorities in policy development and to monitor the progress of populations health. Theoretical and conceptual background A health survey is an investigation in which information on health related issues is systematically collected by means of interviews and/or examinations in a representative sample of the population. Most of the European countries are performing surveys exclusively oriented towards health (status and consumption) and its determinants. In some countries such as Germany and the UK there are multipurpose surveys with a specific module on health. The advantage of a specific health survey is that it allows more in depth and broader study of the health related issues, as there is no competition with other domains of investigation (2). In Belgium during the period 2000-2003, several multipurpose surveys have allowed investigating specific health topics: Health perception, chronic morbidity, functional status and informal care in the Socio- Economic Survey (Census) 2001 Health perception, depression in the Community Household Panel (2000) Chronic morbidity, handicap in the Labour Force Survey (2002). The number of items and questions in those surveys were however quite limited. The health survey in Belgium is performed via interviews only. The potential advantages of examinations (e.g. blood or urine samples) coupled with the interview have not yet been considered due to the elevated cost of such an investigation. Five main domains have been considered in the conceptual framework of the Health Interview Survey (HIS) in Belgium: Health status Health determinants Medical prevention Health consumption Health and society Health status Paying attention to the health status of the population is necessary in accordance to the WHO definition of health and the global approach of the health problems. Measuring health consumption only is no longer sufficient and an instrument such as the HIS is an essential instrument to complement the information usually collected by health care providers, registries and vital statistics. The HIS allows measuring the health status of the population in general and not only in relation with specific health problems. Such a difference is described in the literature as the distinction between health status and state of the health (3). Even if health is the main subject matter of the survey and despite the positive approach of health recommended by WHO, most of the domains investigated in the HIS have to do with ill-health and diseases. A positive conceptual framework was effectively considered when elaborating the HIS but it has unfortunately not been possible to fulfil the concepts due notably to the lack of available instruments (4;5). 7
One of the main characteristics of the survey comes from the fact that most of the information gathered is provided by the individuals themselves with all the potential subjectivity. Their experience and their sensitivity in relation to their own health is thus playing a major role. It is possible however to differentiate from each other relatively objective questions (height and weight for example) from purely subjective ones (self perceived health). Most of the topics investigated in the HIS lie between those two extremities. An other basic concept of the HIS is the differential approach of the health status of the population related to the medically diagnosed diseases on one side and their consequences on the functional status of the individual on the other side. Here also it is referred to the medical diagnosis declared by the person answering to the question without any objective verification of the medical records. The measurement of the health status of the population within the HIS is mainly focused on chronic conditions; due to their long duration these conditions have a bigger impact on health expenditures and represent a higher burden at the population level. Health determinants Life styles are intrinsic components of the daily life of individuals. They are closely linked to the values and the priorities of each person, to the opportunities and constraints inherent to the culture and the socio-economic status. Life styles are in fact shaped by the social acquirements and interpersonal interactions. It is thus misleading to believe that a specific behaviour is only determined by a simple personal decision as far as the refusal or the adoption of health related life styles (deterministic approach). Life styles are however health determinants: some aspects of daily life contribute to the preservation of a good state of health, the prevention of specific conditions and the improvement of the psychological well-being. At the opposite specific behaviours are harmful to health especially if they are excessive or chronic. Better life styles are the main source of improvement in the health of the population, probably before the progresses of the medicines by itself. This is why health promotion is one of the most important components of the public health programmes especially the Health for All initiative of the WHO aiming at improving individual health related behaviours. It is this essential for the public health decision makers as well as for the institutions in charge of the implementation of health promotion programmes to regularly measure the prevalence of specific health related behaviours at the population level, their distribution of specific population subgroups and their trend over time. Such measurement is imperative for the evaluation of programmes and policies. It is not possible to prove the relationship between a programme and a specific trend in the population via health surveys but these are however useful tools for the monitoring of health related behaviours. Medical prevention Clinicians have understood since a long time that preventive medicine is playing a major role. They are every day facing health problems difficult to solve and are thus looking for opportunities allowing to proceed earlier in the disease process or even before the diseases itself. The advantages of preventive medicine became more and more apparent during the last 30 to 40 years. This new approach modifies deeply the way to solve problems such as infectious diseases (with the immunisation programmes for example). Early diseases detection became also an essential component of preventive medicine with striking results as far as morbidity and mortality are concerned (6). 8
Public health policy has progressively been enlarged from the management of health care expenditures to the development of strategies aiming at the improvement of the health of the population. Such an approach is involving specific actions at the level of the biological factors, the physical and social environment, the individual behaviour but also at the level of the health services in their curative and preventive components (7). The WHO Health for All targets published in 1985 explicitly mention health promotion and diseases prevention programmes as priority programmes. For a conceptual point of view three areas can be mentioned in the field of preventive medicine (7). 1. Primary prevention: actions aiming at abrogating the cause of a disease in order to avoid emergence of new cases. 2. Secondary prevention: early detection and treatment of a specific disease before the apparition of the clinical symptoms and the complications. 3. Tertiary prevention: it is not strictly speaking prevention of diseases, but rather trying to limit their consequences. Some modules in the HIS are investigating specific action in the primary and secondary prevention. Several methods have been used in order to select the priority actions in the domain of preventive medicine. The frequency of the disease but also the importance of the problem at the individual and societal level, and the efficacy of the preventive methods are taken into account. The table below summarise the medical prevention priorities of the two communities in Belgium and the corresponding modules in the HIS: French Community Flemish Community Topics in HIS Immunisation Immunisation Immunisation Tuberculosis Cardiovascular diseases Screening for hypertension, cholesterol and diabetes Cancer Breast cancer Breast and cervix cancer screening Health consumption Information of health consumption is an essential part of the health information system in order to adapt available resources to the needs of the population. This covers three main domains: ambulatory care, institutional care and medical drugs consumption. Two different methods are usually available to measure health consumption: services statistics and health surveys. Several sources are accessible in Belgium for health services statistics: reimbursement expenditures (INAMI, RIZIV) for medical activities, reimbursement expenditures (Pharmanet) for prescribed medical drugs and hospital discharge records (RCM, RPM, RIM). It is generally admitted that health services statistics are more reliable than information coming from health surveys. This is due to the memory bias as well as the lack of medical knowledge of the individuals interviewed during the health surveys. However health surveys are the only potential source of information where data are collected on different health related aspects allowing to analyse the level of health consumption in correlation with several determinants such as health status, life styles or socio demographic characteristics. 9
Health and society The concept of health enlarged over time including progressively non medical components became a social phenomenon. Health and ill-health are now in relation with expectancies as far as the environment as well familial and professional insertion. The health status and the social level of the individual are closely linked. This can be studied through the accessibility to health care but also through the detailed analysis of health inequalities, social support, health and environment, traumatisms and violence, Ethical considerations Ethics is committed to ensuring that all research involving human participants is based on good science, meets ethical standards, and complies with international best practice. Ethical considerations that have to be considered are the following: Respect of confidentiality of the collected data and their management Information provided to the individual, including informed consent Accessibility of the data, including informatics' protection measures. Ethical committees are usually solicited to independently appreciate the social dimension of a research project, its impact on the well-being of the individual, the respect of individual rights. It also should verify that the project takes into account the technical and scientific progresses as well as of the possibilities of medicine. Ethics is considered as a part of philosophy that aims at distinguishing the good from the bad, at making choices of moral value. Four basic principles can be described (see Helsinki Declaration): 1. respect of autonomy: human rights, dignity and freedom 2. primum non nocere: not to cause any trouble to the individuals 3. respect of well-being: no hidden information, technical knowledge must aim at increasing well-being. 4. respect of justice: equity and honesty as far as results and utility are concerned. Research projects using animal or human participants, animal or human materials, personal information, or involving clinical trials, or combinations of such studies, require ethical approval (HRC Guidelines on Ethics for Health Research, New Zealand, 2002). Personal information falls into any of the different categories: Information from medical or other private or confidential files Information which may personally identify a research participant Information for which the participant has not given consent for the purposes of the research which is proposed Information which is considered by the participant to be sensitive or valuable in a personal, social, cultural or commercial sense. Protection of the welfare of human participants is a basic principle of ethical review of research. There is a need to balance potential risk of harm to individuals with the possible benefits to society at large. On occasions when there are major issues, there should be broader discussion with the community. 10
Population surveys Some types of research require surveys to be undertaken on "total" populations or on samples of the population selected from public records. This is the case for the HIS. It is considered that direct approaches (for example, by telephone, postal questionnaire or visit interview) to people in the general population selected in this way do not require approval by any local health or medical body or individual practitioner. However, when possible, it may be appropriate to inform local health practitioners about the study. Investigators should consult and, where appropriate, obtain ethical approval from an accredited ethics committee for the research to proceed. The right of any person to decline taking part in such a survey or to withdraw from the survey at any time must always be respected. Where approaches involve visiting or telephoning research participants at their home, it is generally desirable that some advance notice be given and field staff must be provided with means of personal identification including a reference telephone number which the participant may call to establish the field worker's legitimacy. In some circumstances it may be appropriate to inform local police and other appropriate authorities. Informed consent In most cases research constitutes a health care procedure and, as such, written informed consent will be required unless there are good reasons to the contrary. Elements of informed consent include but are not limited to the following basic criteria: 1. The participants' legal competence and ability to understand 2. Information about the proposed research being comprehensively, properly and appropriately given, including any likely outcomes of participation in the research. 3. The participants' consent must be voluntary and not influenced by financial reward, or by duress in any manner, nor must dependent or vulnerable groups be used. 4. Participants must be able to withdraw from the investigation at any time without waiver of any rights and without giving reasons. 5. In the case of those who are unable to give their own consent, for example the mentally incapacitated or the unconscious patient, proxy consent should be sought from a person with appropriate legal authority. 6. In the case of research participants who are children the signature of the parent or guardian should be obtained in addition to the child s assent. Surveys may on occasion involve the physical examination or laboratory investigation of participants. In these circumstances informed consent from each participant must always be obtained before any examination is undertaken, and each participant must be informed of their right to withdraw without explanation from the research at any time without effect to their current or future health care. The research participant must be informed of any consequences to them due to their withdrawal from the research. Where clinical examination is involved, advance information about the survey for local practitioners and appropriate authorities is of special importance. Health information has at the core of its definition the notion that information relates to an identifiable individual. If information cannot be linked to an identifiable individual it will not come within the scope of privacy protection. Anonymised information which cannot be linked to any identifiable individual is thus outside the reach of the privacy protection. It is questionable if the linkage of this anonymised information with other sources requires informed consent from the individual because this linkage process goes to a partial desanonymisation of the information. 11
Collection of Health Information The researcher should only collect information necessary to the research project. Health information may be collected for research purposes from sources other than the individual concerned, if approval by an ethics committee (if required) has been given, and so long as it will not be published in a form that could reasonably be expected to identify the individual concerned. Where a researcher is due to collect information directly from the individual concerned, the researcher must take reasonable steps to ensure the individual knows that the information is being collected, why it is being collected, who will receive it, what consequences might follow if the information is not provided, and what are the individual s rights of access to and correction of the information. Disclosure of Health Information Individuals have rights of access to their own health information. These rights entitle the individual to obtain from the health agency confirmation of whether the agency holds the information and to have access to it. Also, the individual must be told of his or her rights to request correction of the information. Fair information principles The 1974 US Privacy Act s provisions were built on a set of "fair information practices," first published in a Department of Health, Education and Welfare report (Advisory Committee on Automated Personal Data Systems, 1973). They are: (1) there must be no personal-data record-keeping systems whose existence is a secret; (2) there must be a way for individuals to find out what information about them is collected, and how it is used; (3) there must be a way for individuals to prevent information obtained for one purpose from being used for other purposes without consent; (4) there must be a way for individuals to correct or amend identifiable information about themselves; and (5) organizations creating, maintaining, using or disseminating identifiable personal information must assure the reliability of the data for its intended use, and must take reasonable precautions to prevent misuse. Similar enumerations can be found in 1970s declarations in Britain, Canada, France, Germany and Sweden, and in reports by the Council of Europe and the Organization for Economic Cooperation and Development (Flaherty, 1989; Bennett, 1992). Following Bennett (1992), the various international renderings can be condensed into four generic items: (1) openness (that is, anti-secrecy), (2) access and correction (to/of information about oneself), (3) security (anti-access protections appropriate to the data in the system) and (4) minimalism (no more collection, use or disclosure of data than necessary to achieve the system s goals). To these one can add an over-arching fifth principle that flows from democratic values: consent. Either individuals must consent to practices that apply to them, or society as a whole must consent, via a regime of data protection legislation and regulation that sets the rules that will apply to all. 12
Coded personal data In Belgium, the protection of private life during scientific research studies is regulated by the following: MINISTERIE VAN JUSTITIE, [C - 2001/09176] N. 2001 701], 13 FEBRUARI 2001. Koninklijk besluit ter uitvoering van de wet van 8 december 1992 tot bescherming van de persoonlijke levenssfeer ten opzichte van de verwerking van persoonsgegevens. MINISTERE DE LA JUSTICE, [C - 2001/09176] F. 2001 701], 13 FEVRIER 2001. Arrêté royal portant exécution de la loi du 8 décembre 1992 relative à la protection de la vie privée à l égard des traitements de données à caractère personnel. The full text of this regulation is available on the website of the national commission for the protection of private life (http://www.privacy.fgov.be). Health Interview Survey (HIS) in Belgium The protocol of the HIS 2004 in Belgium has been submitted to and approved by the ethical committee of the Scientific Institute of Public Health (IPH). Authorization of the national commission for the protection of private life was also requested but this commission replied that they were not entitled to answer to this request. The National Statistical Institute (NSI) is indeed the institution officially in charge of the sampling and the data collection for the HIS. The identification of the respondents is thus kept only at the level of NSI. This institute is covered by a special regulation and the statistical secret. In the framework of the Health Interview Survey in Belgium, identification data used to identify and contact the individuals selected for the interviews are split from the health and behaviour data collected during the survey. Identification data are kept with the NSI whereas the survey data are deposited in the IPH. It is thus considered that the data kept within the IPH are coded personal data: De term «gecodeerde persoonsgegevens» heeft betrekking op persoonsgegevens die geen element bevatten op grond waarvan de persoon kan worden geïdentificeerd en waaraan een code is toegekend als enig middel om het gegeven met de betrokken persoon in verband te brengen. Le terme «données à caractère personnel codées» vise les données à caractère personnel démunies de tout élément permettant d identifier la personne et munies d un code, qui seul permet de relier la donnée à la personne concernée. As far as IPH is concerned there is thus no obligation to allow the individuals selected for the survey to verify and correct their own data. Such an obligation is only true as far as personal coded data are concerned for the institution in charge of the management of the identification data and the coding system. An informed consent was not requested from the respondents. Despite this, it might be possible to proceed to data linkage with other databases as long as the anonymity of the respondent is respected (a third party is using the coding system to match the data). If identification of the respondent is needed, by example for a follow-up survey, the consent of the respondent can be requested before the follow-up data collection. 13
3. Description of the sample Target population The objective of the Health Interview survey i.e. to give a description of the health status of the population in Belgium-, leads to the broad definition of the target population as all people residing in Belgium, regardless their place of birth, nationality or any other characteristic. Sampling frame The sampling frame consists of all households listed in the national register. Practically a household is defined as the people living at the address of a reference person. In the National Register some people are registered as living in collective households. For these collective households (e.g. homes for the elderly, prisons, religious convents,..) there is no reference person, but in the sampling frame each person belonging to a collective household is considered as a one person household. It is noted that people living in: an institution (including psychiatric institutions), with the exception of elderly people living in old people s homes, nursing homes and psychiatric nursing homes, a religious community or cloister with more than 8 persons, a prison are for practical reasons excluded from the survey, but this is done a posteriori, i.e. when the interviewer identifies that the selected household belongs to one of the indicated categories. People listed in the National Register without any physical address (mostly people living abroad, hence even not included in the target population) are a priori excluded from the sampling frame. In the National Register these can be identified through the field for the statistical sectors (starts with Z or 9 ). Within the selected households all members of the household are part of the sampling frame. Even though the most actual version of the National Register is used and the National Register is updated on a weekly basis, the real composition of a household can be different compared to the administrative composition. The sampling is organised per quarter. Hence, the procedure is followed 4 times, each time in the 2 weeks preceding the start of the quarter. Households selected in a previous quarter are excluded from the sampling frame for the next quarters. The general rule is that the list of household members is based on the information from the household itself. It should also include household members that are temporary (< 1 year) absent and elderly people (65 years and older) that have their official residence at the household address, but reside in fact in a home for the elderly, a nursing home for the elderly or a psychiatric nursing home. If according to the administrative information there are other household members than the ones mentioned by the household itself, the interviewer should ask for this without mentioning that there is a discrepancy between the situation as reported by the household and the National Register. Study population The study population the population that is reached by the study does not cover de target population completely. The following categories of persons are included in the target population, as defined above, but are not included in the study population. 1. All persons, living in Belgium but not listed in the National Register: homeless, illegal refugees, etc. 14
2. People belonging to households listed in the National Register, with no physical address (Codes Z and 9 ). 3. People, residing in: - an institution (including psychiatric institutions), with the exception of elderly people living in old people s homes, nursing homes and psychiatric nursing homes - a religious community or cloister with more than 8 persons - a prison 4. Persons in newly created households that are not yet included as such in the National Register Sampling Design Introduction Sample surveys can be distinguished from other statistical collections by their particular approach to two questions. The first concerns the units from which the population data are to be collected (the sample selection). The second relates to how to infer relevant conclusions, including estimates, concerning the population surveyed is from the data collected. Sampling theory is concerned with the answers to these two questions). In this chapter we will introduce and motivate the procedure used to select units from the target population that has been delineated above. The results of sample surveys are always subject to some uncertainty because only part of the population has been included and because of errors of measurement. Simply increasing the sample size costs both in terms of time and money. Hence, the specification of the degree of precision wanted in the results is an important consideration. Sample Size The total number of successful interviews for the sample is set to 10.000 (3.500 for Flanders, 3.500 for Wallonia, 3.000 for Brussels). This sample size is based on sample size calculations performed during pre-analyses for the health interview survey 1997, taking into account specific budget constraints and the available logistic means. In order to obtain more accurate estimations for the elderly and especially for the age group of 85 years and over the Ministry of Social Affairs asked for and financed an additional sample of 1250 elderly. Based on calculations of the 2001 survey and starting from a basic sample size of 10.000 this would lead to a sample size of more than 3000 persons of 65 years and older and almost 1000 persons of 85 years and over. The additional sample of 1250 is distributed among the 3 regions according to the distribution key of the basic sample (3500/3500/3000), leading to 450 extra interviews in Flanders and Wallonia and 350 extra interviews in Brussels. More specifications on the oversampling of the elderly are explained in annex 3.1. On the basis of the preliminary reports and the analysis of the HIS 1997, the efficiency obtained in estimation at the national and regional level appeared to be sufficient. It was however too small for estimation purposes at the provincial level. In order to answer specific requirements of the provinces in the HIS2001, provinces were encouraged to make extra funds available, enabling a province-specific analysis. In 2004 two provinces agreed to pay for the oversampling and to increase the number of interviews within their province: 450 for Limburg, and 900 for Luxemburg. Even though the provincial oversampling also results in extra elderly in the sample, this was not taken into account when estimations were done to calculate the size of the oversampling of the elderly. 15
As a result the final sample size for the HIS2004 is 12,600. This includes the basic sample (10,000), the oversampling of the elderly (1250) and the provincial oversampling (1350). Stratified Clustered Multi-stage design Motivation In the design of the sampling scheme both the coverage of the people in Belgium and the logistic feasibility of the fieldwork are important concerns. Even when a relatively exhaustive list is available (such as the National Register), a direct selection from this list would be too expensive, because the spread would be too wide. Cost savings may allow the investigators to use a larger sample size than they could use for a simple random sample of the same cost. Therefore, a multiple stage design is a preferable solution. In this design municipalities serve as primary selection units, while households within the municipalities and individuals within households are respectively second and third-stage units. Choosing a stratified sample instead of a simple random sample can be motivated as follows. Sample surveys displaying small variability among the measurements will produce small bounds on the errors of estimation. In other words, stratification may produce a smaller bound on the error of estimation than would be produced by a simple random sample of the same size. This result is enforced if strata are largely homogeneous. In the HIS2004 there are two stratification levels (at the regional and the provincial level). Within a region, a proportional representation per province in the base sample of 10,000 is sought. A simple random sample of municipalities within a region would ascertain this condition from the sampling framework point of view. Resulting differences are regarded as purely random. However, stratifying proportionally over provinces controlled this random variation further. Municipalities are established administrative units, they are stable (in general those units do not change during the time the survey is conducted), and they are easy to use in comparison with other specialized sources of data related to the survey. Municipalities are preferred to regions or provinces, because the latter are too large and too few. The great variation in the size of the municipalities is controlled for by systematically sampling within a province with a selection chance proportional to their size. Within each municipality, a sample of households is drawn so that groups of 50 individuals can be interviewed in total. Clustering also takes place at the household level since members of the same household are more alike than persons not belonging to the same household. Whereas the stratification effects and the systematic sampling according to municipalities have the effect of increasing the precision, the clustering effect (selecting groups of 50, selecting households in stead of individuals) might slightly reduce precision, since units will resemble each other more than in a simple random sample. However, since stratification is based on unequal probabilities (to guarantee meaningful sample size per stratum) a slight decrease in overall efficiency is to be expected. The effects due to clustering and stratification observed in the HIS1997 are very mild and do not outweigh the advantages. This design choice enables indeed persons to be sampled from abbreviated listings and, hence, reduces the survey field worker s travel distance significantly. In summary, in the light of the previous remarks multistage sampling is the appropriated way to get access to individuals. An overview of the steps in the selection procedures is given in the next section. 16
Overview of the design Summarized the final sampling scheme, i.e. the mechanism to get a random sample of households and respondents is a combination of several sampling techniques: stratification, multistage sampling and clustering. The selection process consists of the following steps: 1. Regional stratification. Belgium is divided into 3 regions, the Flemish Region, the Walloon Region and the Brussels Region, for which the number of interviews has been predetermined. The reason for this stratification is to ensure that for each region inference is possible with nearly the same precision. The number of interviews to be interviewed is fixed to 3500 for the Flemish and Walloon region and 3000 for the Brussels region. These figures do not include the oversampling. 2. Stratification at the level of the provinces. This second level of stratification is done to improve the quality of the sample over a simple random sample. In particular a balanced geographical spread is achieved. For the base sample, the sample size within the provincial stratification is proportional to the population size of the province. For the provinces agreeing with an oversampling, the number of interviews obtained via the proportional stratification is increased with the number of interviews the province is paying for. Furthermore, there is the special case of the province of Liege as the sample size of the German Community (which is geographically located within the province of Liège) has been predetermined. The province of Liège has been split into two strata: the German community and the rest of the province. 3. Then, within the strata, units are accessed in two (for the households (HH)) or three (for the individuals) stages: (1) Municipalities are selected with a selection chance proportional to their size, within each stratum. These municipalities are called the Primary Sampling Units (PSU). To keep the fieldwork feasible, each time a PSU is selected a group of 50 individuals have to be interviewed successfully during the year 2004. (2) Within each municipality a sample of households, the Secondary Sampling Units (SSU), is drawn such that 50 individuals per group can be interviewed in total. Households are stratified in 3 groups based on the age of the household members. The selection procedure within each stratum (see further) ensures that the required oversampling of elderly is obtained. (3) Finally, within each household, at most four individuals, the Tertiary Sampling Units (TSU), are chosen. Only questioning the reference person within a household would not enable us to give a good picture of a household's health status. For households with four members or less, all members are selected. For households with at least five members, the reference person and his/her partner (if any) are selected. Among the remaining household members a selection is made according to the birthday rule, so as to yield four household members selected. Interviewing more than four persons within a household is inefficient because of the familial correlation and the burden on the household would be too large. 4. To assure representativity over time, interviews are spread over the year so that each quarter is comparable in terms of number of successful interviews. The quarters are defined as follows: Q1: January-March; Q2: April-June; Q3: July-September and Q4: October-December. All of these aspects and the practical implications to comply with this procedure will be discussed in the next chapters. 17
4. Network of interviewers Composing the network Call for candidates After having selected the target municipalities for the sampling procedure, a pool of candidateinterviewers for those particular municipalities must be brought together. Candidates are recruited in different ways: (1) from the NIS database of interviewers that were active in previous surveys (HIS, Task Force, etc.), (2) from a database of previous candidates who never got the opportunity of actively taking part in a survey, (3) in state administrations and state-run institutions (police office, post office, schools, hospitals, etc.) and (4) if candidates are missing in particular districts where the survey must be carried out, a new recruitment mailing is locally sent out in public or private institutions or companies. The call for candidates mailing that is sent out includes a letter of information (about survey, content, time, wage ) and an application form that must be returned to ISP/NIS in a given delay. The application form provides with the following information: Name of candidate Spoken languages Professional settings (company, address) Private address, telephone and e-mail address Prior participation in surveys (yes/no, which?, when?) Municipalities in which they would like to work (among the list supplied of all participating municipalities), ordered by preference rank. This information is registered in an interviewers' database for HIS. Selection of candidates Candidates are then selected by the IPH in function of (1) predefined criteria (having experience with survey interviews, good survey performance (if available), employment, status etc. and (2) the match between their preference ranks for municipalities and the available sample groups in these municipalities. A detailed overview on the procedure used the select the interviewers can be found in annex 4.1. The selected candidates are further invited to join a training session, while others are kept as reserve-candidates for the replacement of dropouts. Basically 260 experienced and well-trained interviewers must be enrolled in order to collect the data from the households. Normally these interviewers are enrolled for the whole year of the fieldwork. In practice, interviewers may quit during the year, so replacements must be available. To be able to deal rapidly with interviewer turnover, a three- to fourfold replacement scheme should be foreseen. In this sense, municipalities with only 1 or 2 candidates per group (of 50 interviews) are targeted for a further recruitment process. 18
Training of interviewers Many interviewers actually quit at the very beginning of the fieldwork, mostly because their expectations were different from the real daily work (8). This problem is partially solved by using the same interviewers as in the previous HIS, but it should also be tackled within the training course. The training should thus present a good picture of the overall workload and time investment of the job (average number of contacts with households, travel distances, response rates per districts, etc). The training of the interviewers is a very important issue for other reasons as well. As fieldworkers, the interviewers are at the very basis of retrieving the data from the households. From the candidate point of view, the training is regarded as indispensable, because of the complexity of the fieldwork. From the scientific point of view, all data should be retrieved in a standardised manner, including across interviewer's methods. In this sense, all candidate interviewers have to follow a collective training of at least 2 hours. Before attending the training, they receive the HIS questionnaires and the detailed interview guidelines (manual) which they are advised to read thoroughly. During the training sessions, different topics are explained, i.e., the aim of the survey, the rules for contacting the households, the content of the questionnaires, etc. Special attention is paid to the communication that must take place between the secretariat of the survey and the interviewers, and how to use the communication forms. Indeed, regular (weekly) contacts between the interviewers and the NIS secretariat are crucial for optimal fieldwork surveillance, but also for the follow-up procedures of non-contactable or refusing households. As the replacement of these households is based on information provided by the interviewers themselves, a fast and clear communication with NIS secretaries is essential. If an interviewer quits for any reason, the remaining households of his group are reassigned to another candidate-interviewer. This new comer also receives training, either together with others or individually. The training of new interviewers that takes place during the fieldwork period benefits from mistakes and misunderstanding from the active interviewers: mistakes pined out during the fieldwork surveillance or frequently asked questions from active interviewers are systematically explained in detail at the training sessions. Information about these mistakes is also transmitted to all interviewers by means of the newsletter. Moreover, the guideline manual should be appended with the encountered difficulties during the fieldwork or a listing of systematic recurrent errors. Manual for interviewers As part of the training, the candidate-interviewers are provided with a detailed guideline manual for the fieldwork, containing all aspects necessary to execute the survey correctly. The manual is divided in two parts: (1) the survey procedures and (2) the content of the questionnaires. The manual that is used during the training sessions can be found in annex 4.2 and 4.3. 19
5. Fieldwork procedures Secretariat In order to manage the fieldwork of the survey, a permanent office is installed at the NIS - Brussels. This office (3 full time collaborators) is responsible for the practical implementation of the fieldwork (announcing the survey, communication with interviewers, reception of communication forms, reception of questionnaires, data-control, ). The central office works under the responsibility of the director of the survey. Next to the central office, at both the NIS Gent and NIS Liège a collaborator (75%) is responsible for the reception and control of the questionnaires and for data-entry. For the data-entry, a varying number of NIS personnel is implicated. HISIS For the management of the fieldwork a SQL-based informatics system is developed, called HISIS. HISIS is an integrated application for the activation of households, the follow up of the fieldwork, the management of the network of interviewers and the data entry. More detailed information can be found in annex 5.1. The announcement of the survey An invitation letter (see annexes 5.2 to 5.4) and an information leaflet (see annexes 5.5 to 5.7) are sent in advance to the reference person of the selected households. In both documents the objectives of the Health Interview Survey are described and the importance to participate is stressed (although it is also stated that participation is not obligatory). Households are informed that an interviewer will contact them within a fortnight period. At the same time the list of addresses of reference persons is sent to the interviewer (together with names of all the members of the household, their day of birth and the address) together with a blank communication form (see annexes 5.8 to 5.10). This communication form lists already the HIS-identification number of the households. The information on the composition of the households is useful to estimate in advance the workload (how many respondents, how many respondents will have to complete the written questionnaire, ). More important: based on this list the interviewer can check whether the official composition of the household fits the real composition. This is in particular important in the context of the elderly who have their official address in a private household, but who are in fact institutionalised. From the moment the interviewer receives this list, he/she can start contacting the households. From a technical point of view, all households invited to participate (that is: to whom an invitation letter and an information leaflet is sent) receive the status "ACTIVATED". Based on the information transmitted by the interviewers via the communication forms, this status will be altered throughout the fieldwork (see next section). The objective pursued is that - at maximum 6 weeks after the activation of the household - a final status should be obtained (i.e. "refusal", "interviewed", "not contactable"). This period of 6 weeks is split into 3 two-weeks intervals. For each interval the interviewers have to communicate the actual status of all households he/she is responsible for to the central secretariat of the HIS. To do this the interviewer must use the communication form. For every household listed on this form (HIS-identification number), the interviewer must indicate the number of contact attempts with the household, the mode of contact (telephone, doorstep), the day of contact and the current status of the household. 20