An introductory guide to survey research in anaesthesia.
Surveys allow convenient and inexpensive research. Surveys include mail-out questionnaires, email questionnaires, telephone interviews, andpersonal interviews. Despite a widespread perception that surveys are easy to conduct, good surveys need rigorous design, implementation and analysis. This requires substantial planning, time and effort. The most important step in designing a survey is to clearly define the question(s) the survey aims to answer. The target population, measured variables and types of associations being investigated should be specific and unambiguous. Investigators should concentrate on what they need to know rather than what would be nice to know During development surveys should be piloted to identify problems. The main goal when implementing a survey is to maximize the response rate to avoid misleading results. Evidence-based strategies, including brief personalized surveys with stamped return envelopes, can be used to maximize the response rate. A poorly conducted survey can lead to misleading or invalid conclusions and may undermine participation in subsequent surveys by the target population.
Key Words: survey, questionnaire, methodology, design, data collection, epidemiological
A survey is a research method or tool used to gather epidemiological, attitudinal or knowledge-based information from a sample of individuals (1). In anaesthesia, surveys have been used to gather information about current practice, determine the incidence or prevalence of a specific event or disease state and to investigate the beliefs and attitudes of practitioners or patients. For the purposes of this review, surveys include mail-out questionnaires, email questionnaires, telephone interviews, and personal interviews.
Surveys allow convenient and inexpensive research with several potential advantages. First, investigators can study a large population quickly (2). Second, a survey may be the only method for researching a particular topic (3). such as attitudes to a health service, medical product or intervention. Third, surveys can be used to generate hypotheses or be preliminary studies for subsequent observational or interventional research. Finally, if the data collection instrument is standardized, it is possible to compare the characteristics of populations that differ in demographics, place or time (2).
Despite these advantages, survey research has several disadvantages. Survey data may be biased or misleading because surveys use 'self-reported' data, (3). relying on the memory and honesty of respondents. (4). Surveys can also be inflexible, unless open-ended questions or personal interviews are conducted. Further, although a survey can establish a relationship between two variables it cannot be used to determine causality (3). These limitations therefore require careful consideration when embarking on this type of research.
There is a widespread perception that surveys are easier to conduct than other research methodologies. However, in reality, rigorously designed, implemented and analysed survey research requires substantial planning, time and effort. Conclusions drawn from survey research may be misleading or invalid if the design of the questionnaire is poor, the sample
inadequate, or unrepresentative of the target population. In addition, poorly designed questionnaires may discourage individuals from returning subsequent surveys which may be considerably better designed and of greater clinical importance (5).
Surprisingly, there are few studies on questionnaire quality in epidemiological journals6 and the attention given to questionnaire development is often inadequate (7). Although a number of guidelines have been published on conducting and reporting surveys (1,4,8) (table 1) there is no established methodology for conducting and reporting medical surveys (9). this review aims to provide a practical guide for conducting survey research in the field of anaesthesia.
TYPE OF SURVEYS
The three major types of medical survey are epidemiological surveys, surveys on attitudes to a health service or intervention and questionnaires assessing knowledge on a particular issue or topic (1).
A. Epidemiological surveys
Epidemiological surveys are used to assess the characteristics of a population such as the frequency and distribution of a disease state, symptom or therapy. examples include changes in anaesthetic approaches to caesarean section (10). the practice of aspiration prophylaxis in obstetric anaesthesia (11). and the assessment of risk factors for difficult airways (12). epidemiological surveys can be used to assess changes that occur demographically, geographically and temporally.
B. Attitudes of patients or staff members
A survey is often the most practical way of acquiring information regarding the attitudes of staff or patients to a medical intervention, product, service or system (3). Attitudes to some extent will dictate behaviour and will provide some insight into the predicted pattern of use by a doctor or patient (2). Attitudes tend to be stable over months, but may change as new information becomes available and developments arise. Surveys on attitudes can also be used to assess the knowledge base of an intervention or health service and the value that people place on it. Finally surveys on attitudes can provide information on how an individual will act based on their attitude; i.e., their intended future behaviour (2). examples of such surveys include attitudes toward adverse event reporting (13). and patients' attitudes toward the use of local anaesthesia for peripheral cannula insertion (14).
C. Knowledge questionnaires
Surveys can be used to assess knowledge or understanding in a particular field or branch of medicine. the information obtained from such surveys can form the basis for subsequent educational activities or risk management programs. examples of surveys that assess knowledge of a particular topic include junior doctors' knowledge of difficult airway management(15). and the knowledge and practice of prophylactic [beta]-blockers for non-cardiac surgery (16).
PLANNING THE SURVEY
The key to a successful survey is to carefully plan the process in a logical and sequential fashion(2-4) (Figure
1). the first step is deciding on the study's objectives. this will help to clarify the research question(s) and the type of information required, as well as aiding identification of the target population. It is also important to clearly identify where the research question(s) lie within the framework of current knowledge and how clinically relevant the findings of the questionnaire will be (5). this can be done via a comprehensive literature review and by consultation with experts in the field. the literature review can also identify validated research tools (instruments) that can be used as part of the survey. Investigators need to identify what analyses are required for their data, because this will determine the response format of the questions and the design of the survey. the type of analysis to apply may not always be straightforward so it is strongly advised that statistical advice is sought in the planning stages (4). Several guidelines have been published on survey research (1,4,8,9,17-19) (Table 1).
[FIGURE 1 OMITTED]
DEFINING THE SURVEY OBJECTIVES
The most important step in designing a survey is to clearly define the question(s) the survey aims to answer". The target population, outcomes measured, and the types of associations being investigated should be specific and unambiguous. The questions in the survey should be designed and written to accurately meet the objectives. Examples of study objectives include determining "the knowledge of junior doctors of the principles of advanced cardiac life support", "the incidence of symptoms of gastroesophageal reflux in patients undergoing anaesthesia" and "the attitudes and pattern of use of temperature measuring devices by nursing staff".
CHOOSING THE METHOD OF DATA COLLECTION
Data collection is the most time-consuming and costly part of conducting a survey. The method of data collection may influence the response rate of the survey, and has the potential to bias the response to the items of the questionnaire (2). The choice of method of data collection will be dictated by the amount of
A. Personal interview
Personal interviews allow complex questions to be asked and permit the presentation of visual material (3). It is the most expensive technique of data collection per individual surveyed, especiallyif a large population is surveyed. In addition, interviewers must be trained and supervised to avoid biasing the response to the questionnaire. Personal interview is the method most appropriate for small and geographically concentrated samples (2).
B. Telephone interview
Telephone interviews retain some of the advantages of the personal interview but with less expense and intensity of interviewer training. The results of the survey can be compiled relatively quickly and a larger target population can usually be reached for a given budget. Furthermore, a geographically dispersed population can be sampled via telephone much more conveniently than via personal interview. Like personal interviews, telephone interviews allow clarification or explanation of terms or questions (4). The major limitation of telephone interviews is that they do not permit presentation of visual material. In addition, it is more difficult for the interviewer to establish rapport with the respondent.
C. Mail-out data collection
The main advantage of a mail-out survey is the capacity to reach a much larger target population for a given budget (2). In addition, it is possible to present limited amounts of visual material. The major disadvantages of this technique are the potential for non-response and the time required for data compilation. Mail surveys require an accurate list of people to be constructed to ensure adequate and representative coverage of the target population (1). Mail surveys should be accompanied by a'cover letter' that explains why the survey is being conducted'. This cover letter should also include information regarding who is conducting the research, confidentiality and the name and contact details of an available person who can be contacted to answer questions regarding the survey.
D. Data collection methods and bias
The method of data collection may bias the results of the survey in a number of ways. The major source of bias in surveys is non-response bias (2). The response rate to a survey is one of the most important factors dictating the reliability of the survey results (14). Nonresponse bias is likely to be highest in mail-out surveys and lowest in personal interviews. Population sampling bias can occur if the population surveyed is too small or not representative of the overall target population. Another potential source of bias is excluding those without a telephone or email, or those who are away from their usual contact number or email, or those with inadequate language skills. Finally, the interviewer may influence the response of the individual being surveyed (interviewer bias). Appropriate interviewer training can minimize this bias (2).
DEVELOPING THE SURVEY INSTRUMENT
The survey instrument is the hub of the survey process and is the interface between the study objectives and the response of those surveyed. For the research findings to be credible the survey instrument needs to be valid, reliable and generalizable. Validity refers to the degree to which a study accurately reflects or assesses the concept that the investigators are attempting to measure. Reliability, or repeatability, reflects the extent to which the survey instrument yields the same result when applied at different times or to different populations (3 5 19,20). It may be possible to use or modify a previously published validated instrument (3,9) such as the SF-36 (for measuring quality of life). The latter has been validated in ICU patientS (21) as well as primary care (22) and normal population (23). Investigators should seek specialist advice when choosing an instrument. Some valid and reliable survey instruments have not been published in a peer reviewed journal. Conversely, some survey instruments that have been published in the peerreviewed literature are neither valid nor reliable. A systematic review of studies evaluating patient satisfaction data found that 94% of published studies used inadequately tested instruments. Another term, generalizability, refers to the extent that the results from the survey can be generalized to the target population.
A questionnaire should be carefully designed and constructed with the following considerations.
A. Questionnaire structure and format.
The questions (or items)' in the survey should be designed to answer the objectives of the study. The survey should be as short as possible to limit responder fatigue and increase response-rates (2). Accordingly, investigators should concentrate on issues that they need to know rather than the ones that would be nice to know (8). Introductory questions should be easy to answer and address the most important issues of the survey (3). Questions about demographics or personal information should be placed toward the end of the survey. Similar questions should be grouped together into common themes, with each theme flowing smoothly onto the next. Considerable effort should be made to ensure the questionnaire appears professional and is easy to read
B. Principles for writing questions
Each question in the survey should be as short, simple and unambiguous as possible (1-3,8)stion (item) is composed of a stem (the question) and a response (4). stion stems should be unidimensional; they should ask only one question at a time. For example, in an epidemiological survey examining the prevalence of symptoms consistent with ischaemic heart disease, a poorly designed question would ask, Have you ever experienced shortness of breath or chest pain?" A yes response to this question would not indicate which of the two symptoms the respondent has experienced. This question would be better divided into two parts requiring two separate answers. Questions should not excessively test the respondents' memory by forcing them to recall information over an extended period, as this will increase the inaccuracy of the data (1).
C. Language of question writing
Vocabulary used to construct the questions of the survey should be as simple as possible. Researchers should avoid abbreviations, technical terms, acronyms and leading questions that include modifiers (4) such as 'always', 'sometimes','never', or 'occasionally'. Similarly, questions should not contain emotive or subjective terminology (2).
D. Response formats
Questions may elicit either an open or closed response (4,19,20) (Figure 2). An open response format allows the respondent to answer in free text and does not restrict the response of the subject. open response formats are more flexible than closed response formats but require more time and effort to compile and interpret (4). Closed response formats provide a limited number of responses from which the respondent must choose. Closed responses formats should exhaust all possible options (1) and these should be mutually exclusive (3). Closed response formats permit easier interpretation and presentation of results (4).
TARGEt POPULATION AND SAMPLE
The goal of a survey is usually to describe the characteristics, attitudes or knowledge of a population3 though carefully designed surveys can also focus upon organizations or systems. In the majority of cases it is impractical to survey an entire population and a representative sample of the target population is assessed. Before a survey can be conducted, a list of people (sample frame)3 must be created. From this sample frame, inferences are made about the overall population from which it is selected. the quality of the sample frame is probably the dominant feature for ensuring adequate representation of the target population (1). Construction of a sample frame can be difficult, as the membership of the population can change (25).
Investigators need to consider the methods they will use to obtain a representative sample. A list of potential participants should be constructed randomly to reduce selection bias. Random selection can be achieved in a number of ways. the simplest method is to select every nth person from a list or to draw names from a hat. More sophisticated techniques include the use of computer based random number generators (3). there are also free online randomizers that investigators can use (such as www.randomizer. org). the size of the sample frame is largely dictated by resource and budget restraints, but should take into account the degree of variation expected in the population to be studied (1). In addition, the sample size will be governed by the confidence level assigned to achieve the specified precision (25). For example, the size of a survey on postoperative vomiting will depend on the required precision for the estimate of the frequency of vomiting. Data for a departmental audit can be less precise and therefore have a smaller sample than a survey comparing vomiting after different drug treatments. If quantitative and epidemiological data are being collected by the survey, a statistician should be consulted to estimate the minimum sample size needed (19).
Figure 2: examples of open and closed response formats for survey questions. Open Format Responses 1. Text open end How do you feel about the waiting time in our clinic? 2. Numeric open end How long do you usually wait to see a doctor in our clinic? Closed Format Responses 1. Multiple choice How many minutes do you usually wait to see a doctor in our clinic?  less than 5  5-14  15-30  more than 30 2. Rating scales How would you rate the waiting time to see a doctor in our clinic? 1 poor 2 average 3 good 4 very good 3. Pictorial scales How happy are you about the waiting time to see a doctor in our clinic? [ILLUSTRATION OMITTED] 4. Agreement scales (Likert-type) scales the waiting times in our medical clinic are acceptable, (please circle one) strongly disagree uncertain agree strongly disagree agree 5. Visual analogue scale (10 cm line) How would you rate the waiting time to see a doctor in our clinic? very poor --x-- very good
Once the questionnaire has been drafted, the next step is to test the survey before distribution, i.e., pilot the survey (7). The pilot involves assembling 6-12 people that are similar to the sample frame (population) to be surveyed (3). This assembly forms a focus group (1) to predict the emotional responses and comprehension of the target population (9). Pilot tests help identify poorly worded or redundant questions (9), ascertain whether questions are too personal (1), and help identify potential logistical problems with conducting the survey (5). Managing this process will be easier if the survey is clearly labelled and relabelled, such as Version 1 and Version 2, during its evolution. One approach to a focus group is to use a'moderator' skilled in managing group dynamics (1). The moderator's task is to help determine how the members of the focus group comprehend the questionnaire and arrive at their answer (9). Participation of all members of the focus group is encouraged and questions should be open-ended (1). To ascertain whether the questions in the survey are leading or biased, the focus group should be asked to guess the hypothesis that the study is seeking to answer (3). The feedback obtained from the focus group can be recorded on video or audiotape, or by manual note-taking.
ADMINISTERING THE SURVEY
The main goal of any survey implementation plan is to maximize the response rate (3). Low response rates increase the risk of non-response bias. That is, the characteristics of non-respondents may differ to those of respondents, leading to misleading results (3). A response rate of greater than 60% for self completion questionnaires and 75% for interviews is generally considered to be acceptable because it is sufficient to permit generalization of the sample frame to the population being studied (4,18).
All surveys should be accompanied by an introduction (verbal or written) explaining the objectives of the survey, why the respondent has been chosen to complete the survey, and why the response of the individual is important to the study. An indication of the amount of time required to respond to the survey should be provided. Most surveys are confidential and, under these circumstances, it should be stressed that individual responses will remain confidential'. On occasion a survey may be designed that does not require confidentiality. Under these circumstances this must be emphasised to potential respondents and permission obtained from respondents to use the information for the stated purpose.
OPTIMIZING THE RESPONSE RATE TO THE SURVEY
A. Response rates and the method of data collection
There is evidence to suggest that the method of data collection has a qualitative and quantitative impact on the response to the questionnaire. In a Canadian survey of Family Physicians (26) the response rate for post-mail was greater than for email (53% versus 34%). However, email was returned faster and contained longer comments. A similar randomized controlled trial of survey data collection methodology conducted with Canadian anaesthetists2' found that the response rate of postal participants was twice that of email participants (69% versus 35%). Smeeth and co-workeer (28) compared response rates and levels of self-reported morbidity using postal survey or personal interview either by a lay interviewer or a nurse. Although response rate was highest for postal survey, the proportion of missing or invalid responses was highest using this method of data collection. In addition, levels of self-reported morbidity tended to be lower in the group interviewed by a nurse than with the other two methods.
Other factors may also affect the response rate of participants. For example, in a survey investigating perioperative anaesthesia complications and postoperative pain, the researchers examined whether the timing and sequence of interviews would improve the response rate of patients (29). During the control period, interviewing started at 9.00 a.m and concluded when at least one effort had been made to contact each eligible patient (usually around 1.30 p.m). During the intervention period, patient interviews coincided with nursing timetables for patient care. The response rate during the control period was 60% and 88% during the intervention period.
B. Optimizing response rates to mail-out surveys
Non-response to a survey is most likely in mail-out questionnaires unless several principles are adhered to (5,9,30,31) (Table 3). Edwards and co-workers" have recently completed a systematic review to identify methods to increase responses to postal questionnaires. A cover letter should accompany the questionnaire concisely explaining who is conducting the survey and why. Participants are twice as likely to respond if the questionnaire interests them. Confidentiality should be highlighted. The questionnaire should be titled to accurately indicate what the survey is about (19). A letter or email telling the recipient that they will soon receive the survey may improve the response rate (32).
The overall layout of the questionnaire is important. The font size should be legible to the visually impaired, and the layout should be clear and simple9. The use of coloured ink has been shown to increase response rates (31). The questionnaire should be as short as possible. This strategy can double the response rate (31). The questionnaire should contain items worded along the principles outlined above, and should not contain overly sensitive or personal questions unless they explicitly meet the studies objectives (31). Incentives such as money (31), lottery tickets (33), or continuing medical education (CME) points (34) may increase the response rate. A recent mail-out survey of Australian anaesthetists achieved a 85% response rate using several of these strategies (35).
COLLECTING AND COMPILING THE DATA
Data can be collected on survey sheets completed by the respondent or the interviewer, or may be directly entered onto a computer, particularly with telephone interviews (1). Internet technology can provide links between questionnaires distributed via email and websites. These websites can allow respondents to enter their responses directly into a computerized database. Data received from mail, telephone or interview questionnaires should be entered into a computer database to facilitate processing. Manual entry of data is laborious and error prone. Questionnaire sheets that can be directly scanned by computer are increasingly available but require expertise as well as specialized software and hardware (4). Investigators must decide how to handle missing data fields and the format of presentation of the study results (1). A statistician should be consulted for analysing the survey results (9).
Reporting the study findings should follow the format of other medical research: introduction, methods, results and discussion. The specific hypothesis or research question should be clearly stated, such as: "We tested the hypothesis that patients find waiting times in our clinic acceptable" or "We tested the hypothesis that anaesthetists prefer to use laryngeal mask airways for elective day case procedures".
A. Reporting the survey methodology
The methodology section should allow the reader to judge the quality of the survey (3). Researchers should describe the method of data collection, the format of the items and responses. Ideally, the questionnaire should be included in the publication, possibly in an appendix. A brief account of the process of pretesting and instrument revision should be provided (19). The characteristics of the sample frame should be described, as should the method of its construction. The method of administration and efforts to maximize response rate should be included.
B. Results and discussions
The response rate of the survey must be reported. Where possible researchers should comment on how the respondents and non-respondents may differ. This can be summarized in a recruitment diagram (19). The data obtained from the study and its implications for the original hypothesis should be discussed (3) as should the limitations of the survey. In particular, it is important to comment on the various forms of bias that may have occurred during the study. Finally, the importance of the findings should be outlined, including directions for future research and implications for clinical care, education and research.
Investigators should have some understanding of the ethical implications of their research and should consult specialist resources such as the National Statement on Ethical Conduct in Research Involving Humans, published by the National Health and Medical Research Council (36). Approval to conduct the survey should be obtained from an appropriate Human Research Ethics Committees (HREC). Most medical journals require HREC approval as a prerequisite for publication. Surveys of professionals, such as anaesthetists, should have HREC approval, usually from a hospital or university HREC. Surveys of patients in multiple hospitals will require approval from the HREC representing each hospital. It should be unnecessary to send a separate consent form with postal surveys because returning the questionnaire is assumed to imply consent to participating in the study (37) . However, the cover letter and survey materials will need to provide sufficient information for the respondent to make an informed decision (37).
The questionnaire should not cause offence and should not breach confidentiality (19). Confidentiality can be maintained by using codes, by destroying questionnaires after their data have been entered into a computer, or by omitting identifying information on the questionnaire (1). Postal surveys can include an "I do not wish to participate" slip to help avoid unwanted reminders. Further, the survey should not be excessively time-consuming so as to waste excessive amounts of the participants' time (19). Some have argued that surveys are unethical if they are unlikely to affect health care (38).
A survey is a relatively easy and inexpensive form of research that has the capacity to reach a large population relatively quickly. However, a poorly conducted survey may lead to misleading or invalid conclusions and may undermine participation in subsequent surveys by the target population, including Australasian anaesthetists. Accordingly, each stage of the survey process must be carefully planned to ensure that the objectives of the study are answered and that the responses of the cohort sampled are representative of the overall target population.
Accepted for publication on November 18, 2005.
(1.) American Statistical Association. What is a survey? Available at: www.amstat.org/sections/srms/whatsurvey.html. Accessed 26 May 2005.
(2.) Alreck PL, Settle RB. The Survey Research Handbook. 2nd Ed. New York: McGraw-Hill, 1995.
(3.) Doyle JK, Woods DW Introduction to survey methodology and design. Available at: www.sysurvey.com/tips/whitepapers.asp. Accessed 25 May 2005.
(4.) Passmore C, Dobbie AE, Parchman M, Tysinger J. Guidelines for constructing a survey. Fam Med 2002; 34:281-286.
(5.) Bruce J, Chambers WA. Questionnaire surveys. Anaesthesia 2002; 57:1049-1051.
(6.) Wilcox AJ. The quest for better questionnaires. Am J Epidemiol 1999; 150:1261-1262.
(7.) Olsen J. Epidemiology deserves better questionnaires. IEA European Questionnaire Group. International Epidemiological Association. Int J Epidemiol 1998; 27:935.
(8.) Day M. 20 top tips to writing effective surveys. Available at: www.ezinearticles.com/?expert=Martin Day. Accessed 25 May 2005.
(9.) Boynton PM. Administering, analysing, and reporting your questionnaire. BMJ 2004; 328:1372-1375.
(10.) Stamer UM, Wiese R, Stuber F, Wulf H, Meuser T Change in anaesthetic practice for Caesarean section in Germany. Acta Anaesthesiol Scand 2005; 49:170-176.
(11.) Helbo-Hansen HS, Bang U. Current Danish practice for aspiration prophylaxis in obstetric anaesthesia: a survey. Int J Obstet Anesth 1993; 2:233-235.
(12.) Cattano D, Panicucci E, Paolicchi A, Forfori F, Giunta F, Hagberg C. Risk factors assessment of the difficult airway: an Italian survey of 1956 patients. Anesth Analg 2004; 99:17741779.
(13.) Yong H, Kluger MT Incident reporting in anaesthesia: a survey of practice in New Zealand. Anaesth Intensive Care 2003; 31:555-559.
(14.) Brown J. Using lidocaine for peripheral i.v. insertions: patients' preferences and pain experiences. Medsurg Nurs 2003; 12:95100.
(15.) Rosenstock C, Ostergaard D, Kristensen MS, Lippert A, Ruhnau B, Rasmussen LS. Residents lack knowledge and practical skills in handling the difficult airway. Acta Anaesthesiol Scand 2004; 48:1014-1018.
(16.) Rapchuk I, Rabuka S, Tonelli M. Perioperative use of betablockers remains low: experience of a single Canadian tertiary institution. Can J Anaesth 2004; 51:761-767.
(17.) Bowling A. Research methods in health: investigating health and health services. Buckingham: Open University Press, 2000.
(18.) Kelley K, Clark B, Brown V, Sitzia J. Good practice in the conduct and reporting of survey research. Int J Qual Health Care 2003; 15:261-266.
(19.) Boynton PM, Greenhalgh T Selecting, designing, and developing your questionnaire. BMJ 2004; 328:1312-1315.
(20.) Writing Center at Colorado State University. Conducting surveys. Available at: http://writing.colostate.edu/references/ trad_research/intro/com2h4.cfm. Accessed 25 May 2005.
(21.) Chrispin PS, Scotton H, Rogers J, Lloyd D, Ridley SA. Short Form 36 in the intensive care unit: assessment of acceptability, reliability and validity of the questionnaire. Anaesthesia 1997; 52:15-23.
(22.) Brazier JE, Harper R, Jones NM et al. Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ 1992; 305:160-164.
(23.) Jenkinson C Coulter A, Wright L. Short form 36 (SF36) health survey questionnaire: normative data for adults of working age. BMJ 1993; 306:1437-1440.
(24.) Sitzia J. How valid and reliable are patient satisfaction data? An analysis of 195 studies. Int J Qual Health Care 1999;11:319328.
(25.) Burmeister LE Principles of successful sample surveys. Anesthesiology 2003; 99:1251-1252.
(26.) Seguin R, Godwin M, MacDonald S, McCall M. E-mail or snail mail? Randomized controlled trial on which works better for surveys. Can Fam Physician 2004; 50:414-419.
(27.) VanDenKerkhof EG, Parlow JL, Goldstein DH, Milne B. In Canada, anesthesiologists are less likely to respond to an electronic, compared to a paper questionnaire. Can J Anaesth 2004; 51:449-454.
(28.) Smeeth L, Fletcher AE, Stirling S et al. Randomised comparison of three methods of administering a screening questionnaire to elderly people: findings from the MRC trial of the assessment and management of older people in the community. BMJ 2001; 323:1403-1407.
(29.) Klock PA, Jr., Lichtor JL, Mingay DJ et al. Improving the efficiency of survey research in postoperative patients. J Clin Anesth 1998; 10:28-31.
(30.) Smeeth L, Fletcher AE. Improving the response rates to questionnaires. BMJ 2002; 324:1168-1169.
(31.) Edwards P, Roberts I, Clarke M et al. Increasing response rates to postal questionnaires: systematic review. BMJ 2002; 324:1183-1185.
(32.) Harrison RA, Cock D. Increasing response to a postal survey of sedentary patients-a randomised controlled trial. BMC Health Serv Res 2004; 4:31-33.
(33.) Kalantar JS, Talley NJ. The effects of lottery incentive and length of questionnaire on health survey response rates: a randomized study. J Clin Epidemiol 1999; 52:1117-1122.
(34.) McDermott MM, Greenland P, Hahn EA et al. The effects of continuing medical education credits on physician response rates to a mailed questionnaire. Health Mark Q 2003; 20:2742.
(35.) Myles PS, Symons JA, Leslie K. Anaesthetists' attitudes towards awareness and depth-of-anaesthesia monitoring. Anaesthesia 2003; 58:11-16.
(36.) National Health and Medical Research Council. National Statement on Ethical Conduct in Research Involving Humans. Canberra: Commonwealth of Australia, 1999.
(37.) Human Research Ethics Committee DoHS, Victoria. Preparing participant information and consent forms. http://www.health. vic.gov.au/ethics/downloads/picf-guidelines.pdf. Accessed 1 June 2005.
(38.) Peat J. Health Science Research: a handbook of quantitative methods. London: Sage Publications 2002.
D. JONES * STORY [dagger], O. CLAVISI [double dagger], R. JONES [section], E PEYTON ** Trials Group, Australian and New Zealand College of Anaesthetists, Melbourne, Victoria, Australia
* B.Sc.(Hons), M.B., B.S., Registrar, Department of Intensive Care, The Alfred Hospital, Melbourne.
[dagger] B.Med.Sci.(Hons), M.B., B.S.(Hons), M.D., EA.N.Z.C.A., Joint Director of Research, Department of Anaesthesia, and Principal Fellow, The University of Melbourne, Department of Surgery, Austin Health, Melbourne; Chair, Trials Group, Australian and New Zealand College of Anaesthetists.
[double dagger] B.Sc. (Hons), M.P.H., Research Coordinator, Trials Group, Australian and New Zealand College of Anaesthetists.
[section] B.Sc., Dip.Ed., B.Ed.(Hons) Ph.D., Director of Education, Australian and New Zealand College of Anaesthetists.
* * M.B., B.S., M.D., P.G.Dip.Echo., EA.N.Z.C.A., Joint Director of Research, Department of Anaesthesia, and Senior Lecturer, The University of Melbourne, Department of Surgery, Austin Health, Melbourne.
Authors on behalf of the Trials Group, Australian and New Zealand College of Anaesthetists.
Address for reprints: A/Prof David A. Story, Department of Anaesthesia, Austin Hospital, Studley Rd, Heidelberg, Victoria, Australia 3084.
Table 1 Twelve principles for conducting a survey 1. Clearly define the purpose and objectives of the survey. 2. Give the survey an appropriate, accurate title. 3. Make the survey as brief and simple as possible. 4. Keep each question short, simple, unambiguous and unidimensional. the questions should be designed to specifically answer the study objectives. 5. Avoid questions and data collection techniques that influence the answers. 6. Decide how the data will be compiled and analysed before conducting the survey. 7. Identify and target a representative and appropriately sized sample of the overall target population. Quantify the response rate to the survey and assess the characteristics of respondents and non-respondents. 8. Pilot (pretest) the questionnaire, ideally with a representative sample of the focus population that is not going to be included in the sample to be used for the final survey. 9. Revise the questionnaire following the results of the pilot. Re-pilot the revised questionnaire. 10. Distribute questionnaire to the broader sample from the target population. 11. Allow space for voluntary additional comments. 12. Always thank the respondents. TABLE 2 Characteristics of survey data collection methods Personal Phone Mail/email interview interview Cost +++ ++ + Need for interviewer recruitment and standardization +++ ++ + Time for data collection ++ + +++ Ability to obtain complex responses +++ ++ + Ability to ask personal questions + ++ +++ Potential for large sample + ++ +++ Ability to present visual material +++ - ++ Potential for non-response + + +++ TABLE 3 Strategies to improve response rates to postal questionnaires 1. Enclose a personalized cover letter. 2. Include stamped return envelopes. 3. Contact the participants before sending the questionnaires. 4. Send intermittent reminders to non-respondents and include an additional copy of the questionnaire with each reminder. 5. Ensure the cover letter and questionnaire appear professional and attractive: a. appropriate and accurate title for the questionnaire; b. coloured ink; c. clear and simple layout. 6. Make the questionnaire as short as possible. 7. Attempt to make the questionnaire interesting and relevant to the participant. 8. Avoid personal or sensitive questions. 9. Provide incentives for respondents.
|Printer friendly Cite/link Email Feedback|
|Author:||Jones, D.; Story, D.; Clavisi, O.; Jones, R.; Peyton, P.|
|Publication:||Anaesthesia and Intensive Care|
|Date:||Apr 1, 2006|
|Previous Article:||B-type natriuretic peptide concentrations and myocardial dysfunction in critical illness.|
|Next Article:||Quality of life at three years after cardiac surgery: relationship with preoperative status and quality of recovery.|