Study design and participants
This current study is nested within an international Internet-based cross-sectional study that aimed to better understand health information seeking behaviours and needs among individuals with CRC. We administered the survey internationally as a reflection of growing exchange of information and support between individuals with CRC from different countries, largely due to rising popularity online communities and use of social media (e.g. Facebook groups, Instagram).[14-16] Participants were eligible for the study if they: 1) were 18 years of age or older; 2) received a diagnosis of CRC; and 3) were able to complete the online health survey in English, French, Spanish, or Mandarin. Online, we recruited participants through Twitter, Facebook, and Instagram using our channels and those from partner CRC organizations (e.g., Colorectal Cancer Canada, COLONTOWN®, Fight Colorectal Cancer, Colorectal Cancer Alliance, and Young Adult Cancer Canada). Offline, we advertised the study using posters at waiting rooms, consultation rooms, and common spaces at two clinics providing CRC care in Vancouver, Canada. A researcher (KS) also recruited participants once a week at a local cancer centre in Vancouver, Canada, and those who consented completed the survey using a provided laptop. Finally, traditional media, which included newspaper, television, and radio interviews, were used to promote the study.
We administered an online health survey, which was hosted on Qualtrics, a survey platform supported by our institution and compliant with the British Columbia (BC) Freedom of Information and Protection of Privacy Act. Altogether, the survey included 12 pages online and consisted of four sections including demographic and CRC characteristics, health information needs, quality of life and health status. (Figure 1), as well as health information seeking behaviours, which we reported in our prior study. Incorporation of computer adaptive features in the survey platform facilitated administration in terms of tailoring specific items and/or sections based on participants’ responses to prior questions.
Demographic information and CRC characteristics
The section on demographic information and CRC characteristics comprised 12 questions on current age, sex, ethnicity, country of residence, marital status, education level, residence (e.g., urban, rural), age at CRC diagnosis, type of cancer (e.g., colon cancer, rectal cancer), cancer stage, treatment(s) received, and treatment status in terms of whether participants were currently receiving treatment for CRC or had completed treatment.
Health information needs
For the section on health information needs, we drew from van Mosel et al.’s 2012 scoping review of information needs across the CRC cancer care continuum, research team members’ clinical expertise (JL, SG), and input from patient research partners in designing questions and items (Supplementary Table 1). Questions were designed in matrix form, that is, a close-ended question that asked participants to evaluate rows of items using the same set of column choices. We explored three aspects of health information needs about CRC: 1) general information needs about CRC combined two questions that captured 14 items including survival information, risk of cancer for family members, sexual activity, fertility, work/employment, and mental health; 2) CRC treatment information needs included three possible questions with four (for surgery, radiation) or five (for chemotherapy) items that participants were prompted to answer (e.g., what to expect with treatment, what are treatment side effects) based on the treatment(s) they were currently receiving or had received for their CRC; and 3) specific information needs according to participant treatment status with seven items for participants undergoing treatment and five for those who had completed treatment, including similar items on exercise and physical activity, nutrition and diet, and bowel activity. We applied the response format of the Cancer Survivors Unmet Needs (CaSUN) questionnaire– specifically for each item, participants indicated one of the following five options: 1) information need has been met; 2) information need has not been met, need is weak; 3) information need has not been met, need is moderate; 4) information need has not been met, need is strong; and 5) information is not applicable (because it is not a need). Altogether, based on the adaptive design of our survey, the number of possible items on health information needs ranges from 25 to 34 for participants undergoing treatment and 22 to 31 for participants who have completed treatment. Finally, we also asked participants to indicate the amount of effort needed to obtain health information about CRC according to four-scales: no effort/easy to find (1), little effort/somewhat easy to find (2), moderate effort/not easy to find (3), and a lot of effort/information still not found (4).
Descriptive statistics, including counts and frequencies, were used to characterize participants’ sociodemographic and CRC characteristics, and group participants according to treatment status, namely those undergoing treatment and those who have competed treatment. For descriptive purposes, we applied a cut-off of having greater than 50% of participants indicating the response option of “information need has been met” to classify whether each item as met (and conversely, unmet). As there are no prior assessments on the extent of met health information needs among individuals of CRC to base a cut-off, we applied 50% as a conservative approach.
We then quantified met health information needs for each participant by first determining the number of specific items for each participant (e.g., number of health information need items). This was important since the number of health information needs items varied according to participants’ treatment status and the type of treatment modalities they received; as well some participants did not provide complete responses. Second, using the CaSUN response format, we determined whether a health information need item was met and then tallied the number of met items (e.g., number of met health information needs). Third, from these two values, we calculated average met health information needs for each participant as a relative value, ranging from 0 to 1 and expressed as a percentage, to ensure the same scale for all participants. We applied multivariable linear regression models, using the overall average number of met health information needs as the dependent variable, to determine factors associated with meeting health information needs in individuals with CRC. Factors considered included CRC characteristics (e.g., age at diagnosis, cancer type, and stage) as well as sociodemographic characteristics (e.g., location, education, marital status, and ethnicity). Separate models were computed according to treatment status. We used SAS 9.4 for the data analysis.
Ethical approval & data sharing
This study was approved by the University of British Columbia Research Ethics Board (#H18-02540). Confidentiality of all participants was ensured by the research team. Secure firewall servers were used to store all research files and only members of the research team had access to the data. The research data are not shared.