Survey development involved three stages including identification of important content areas, development of draft questions, and survey refinement.
Identifying important content areas for inclusion in the survey involved conducting a comprehensive English language literature search, consultation with relevant stakeholders (e.g. members of the Nova Scotia Department of Health), and input from subject matter experts at Dalhousie University. The literature review was conducted using the following bibliographic databases: PubMed, Cumulative Index to Nursing and Allied Health Literature (CINAHL), International Pharmaceutical Abstracts (IPA), and Web of Science Citation Databases. Hand and electronic searching of relevant journals was also conducted. Broad search terms were used without limits on publication date or place as nurse practitioner titles, roles and scopes of practice, and terminology regarding technology vary nationally and internationally. Some examples of terms used included nurse practitioner, nurse prescriber, nurse clinicians, district nurse, health visitor, drug information resources, drug information services, information needs, and information technology.
The draft survey was reviewed by the research team to reduce the number of items and improve clarity. The layout of the questionnaire was carefully examined to ensure that it was easy to follow and complete. Research results from a previous investigation of Nova Scotian physicians' behaviours regarding drug information were also used to further revise the survey . This draft questionnaire was pilot tested by two out of province NPs and one physician. The results of the pilot were used to make final revisions to the survey. Based on pilot-testing feedback and investigator consensus, the final survey was divided into 2 versions, one for personal digital assistant (PDA) users and one for computer users.
The 10 page surveys for PDA and computer users had 5 or 6 sections, respectively, and 37 questions, many with multiple parts. The survey content included demographics, computer or PDA use and experience, drug and therapeutic resource use and preferences, PDA future use, perceived barriers and facilitators to PDA use, and technology training preferences.
Section one contained demographic questions such as gender, age, job title, volume of patients, and EPR availability in the practice setting. Section two was designed to determine PDA or computer use and experience in the practice setting with questions regarding length of use, costs, and work versus home usage. This section also addressed usage and rating of different drug information resources. Resource ratings were based on the frequency of usage, usefulness, accessibility, credibility, and current/timeliness. Resources were grouped as print (i.e. books, journals, and clinical practice guidelines), online/electronic resources, and health professionals and other. Respondents used 5-point Likert scales (strongly agree to strongly disagree) for rating opinions related to resources. A rating of 6 (not applicable, I do not use this resource) was also included for respondents who did not use a particular resource. Frequency of searching for specific information was rated on a 3-point Likert scale (frequently to never). The final sections of the survey included categorical, open-ended, and Likert scale questions regarding preferred resources, technology barriers, PDA future use, and technology training preferences. Copies of the surveys are attached as an appendix in PDF format [see additional file 1 and 2] or can also be accessed from the Initiative for Medication Management, Policy Analysis, Research & Training (IMPART) website .
Ethics approval for the survey was granted through Dalhousie University Research Ethics Board on February 3, 2005.
The survey recruitment procedures were based on the methods of Dillman  and Salant and Dillman . Survey packages contained a cover letter, separate surveys for PDA and computer users, and a return self-addressed stamped envelope. The covering letter instructed respondents to self-select the appropriate survey (either PDA or computer) based on their drug information seeking behaviours. Participants who had used a PDA at any time were instructed to complete the PDA version of the survey. Those who had never used a PDA for drug information were instructed to complete the computer version of the survey. Several strategies were used to optimize response rate and included: personalized cover letters, coloured paper for surveys, stamped return envelopes, follow-up mailing, and a priority post mailing . The covering letter included coloured logos of Dalhousie University and the Nova Scotia Department of Health representing the investigator affiliations and endorsement of the project.
A master mailing list with names and addresses of NPs and their collaborating physician partners was created. To maintain confidentiality of respondents, a number placed on the bottom right corner of each survey corresponded to a name on the confidential master mailing sheet. The postage paid return envelopes were addressed to the research coordinator at the School of Nursing, Dalhousie University, who matched respondents to the mailing list from the first distribution. The cross-referenced mailing list was not accessible to those entering or analysing data. The research coordinator sent the second distribution to those who had not initially responded. A fluorescent coloured page was included in the second mailing to notify recipients of the second and final mailing status. The second mailing followed 2 weeks after the initial mailing (February 2005). The surveys were sent via Xpresspost™ through Canada Post.