Health information sources for different types of information used by Chinese patients with cancer and their family caregivers

Abstract Context Little is known about the information sources of Chinese patients with cancer and their family caregivers, yet this knowledge is critical for providing patient‐centred care. Objective To assess and compare the information sources used by Chinese patients with cancer and their family caregivers. Design The validated Health Information Wants Questionnaire (HIWQ) was translated and administered in March 2014. Setting The oncology department of a general hospital in south‐west China. Participants A convenience sample of 198 individuals, including 79 patients with cancer (mean age=55.24, SD=13.80) and 119 family caregivers (mean age=46.83, SD=14.61). Main outcome measures Ratings on the HIWQ items assessing information sources for different types of information. Results The interaction between information source and group was significant (F 3,576=6.32, P<.01). Caregivers obtained more information than patients from the Internet. Caregivers and patients did not differ in the amount of information they obtained from doctors/nurses, interpersonal contacts or mass media. The interaction between information type and information source was significant (F 18,3456=6.38, P<.01). Participants obtained more information of all types from doctors/nurses than from the other three sources and obtained more information from interpersonal contacts than from mass media or the Internet. Conclusions The information sources of Chinese patients with cancer and their family caregivers were similar, with an important difference that caregivers obtained more online information than patients. These findings have important implications for patient care and education in China where the family typically plays a major role in the care and decision making.

million die of cancer annually. 1 Yet little is known about the health information sources Chinese patients with cancer and their family caregivers use to obtain different types of information. Research to date has focused primarily on the health information behaviours of cancer patients while understudying those of family caregivers, 2 despite the fact that cancer caregivers often carry heavy burdens of care and are in great need of various types of information to help themselves and their loved ones cope with and manage the condition. [3][4][5][6] The limited evidence on cancer caregivers suggests that caregivers' health information behaviours are similar to those of patients with cancer. These include that, first, medical professionals are the primary information source for patients with cancer and caregivers; 2,7-10 second, patients with cancer and caregivers typically are unable to obtain sufficient information from their health-care providers; [11][12][13][14][15] and third, other information sources (e.g. interpersonal contacts with friends and family, mass media, and the Internet) are commonly used by patients with cancer and caregivers to help meet their many unmet information needs. 2,[7][8][9][10]16,17 In terms of the types of health information wanted and sought by patients with cancer and caregivers, existing research tends to predominately examine a limited range of types of information, particularly those about the diagnosis/status of the illness and treatment plans and options, while understudying other types of information that are also important to patients and caregivers, for example information about how to cope with cancer psychologically and socially. 8,11,12,14,15 Our prior work on Health Information Wants (HIW) represents an important exception to this predominant approach.
HIW is defined as 'health information that one would like to have and use to make important health decisions that may or may not be directly related to diagnosis or standard treatment' 18 (p. 5.14). This concept was developed with the intention to cover a broad range of types of health information that patients typically would be interested in health-care contexts. This broad coverage differentiates the HIW concept from the previously predominant perspective that health-care providers typically have: from the provider's perspective, patients need to have only a limited range of information (which typically focus on compliance). [18][19][20] From the patient's perspective, however, there is often a wide range of types of information that patients want to have for reasons that may, or may not, be related to compliance, for example information about coping with cancer. The HIW concept therefore promotes an understanding of patient preferences from the perspective of the patient instead of the provider.
Our subsequently developed HIW Questionnaire (HIWQ) operationalizes patients' desire for health information to seven specific aspects of medical encounters: diagnosis, treatment, laboratory testing, self-care, complementary and alternative medicine (CAM), psychosocial aspects and health-care providers. The validity and reliability of the HIWQ are supported by empirical evidence from college students and community-dwelling older adults in the United States. 21,22 Our previous HIWQ-related work focused on patients only and did not take into account family caregivers. This study represents our first effort to translate and adapt the HIWQ to the Chinese context. In particular, in recognition of the important role Chinese families typically play in cancer patient's information seeking, decision making and coping, [23][24][25] we expanded the original HIWQ to also assess Chinese family caregivers' information wants and behaviours. The first paper derived from this study, focusing on comparing differences in types of information patients and family caregivers wanted vs obtained from medical professionals, was reported elsewhere. 26 The present paper focuses instead on assessing and comparing the health information sources used by Chinese patients with cancer and their family caregivers for different types of information. Advancing this knowledge is important because it might inform clinical practice and health education interventions, leading to improved patient care quality and outcomes (e.g. if evidence suggested patients tended to obtain one particular type of information from one particular source while family caregivers tended to obtain the same type of information from a different source, then health-care professionals might want to develop interventions targeting the different sources to ensure patients and their families obtaining the necessary information). The primary research questions guided the present paper were as follows: • What are the information sources used by Chinese patients with cancer and their family caregivers for different types of health information and what differences and/or similarities might there be between these two groups in the information sources they use for different types of health information?

| Research site
Participants were recruited from the Oncology Department of the Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital (referred to as the 'Hospital' thereafter). The Hospital is a top-rated large general hospital located in Chengdu, Sichuan, China.
It has with over 3000 beds serving patients primarily from southwest China, an economically less developed region. Its Oncology Department has 124 beds for patients diagnosed with a variety of cancer types and stages.

| Design and materials
This study was a cross-sectional, paper-based survey study of Chinese patients with cancer and their family caregivers. The validated 21item HIWQ, 21,22 developed based on the HIW concept as reviewed above, 18 was translated and adapted for this study. The HIWQ includes two parallel scales, with one measuring preferences for different types of information (the Information Preference Scale/IPS) and the other measuring preferences for participation in the corresponding types of decision making (the Decision-making Preference Scale).
Due to the specific scope of the present study, we adapted only the IPS in this study. The IPS contains seven subscales with items in each subscale measuring one specific type of health information with a 5point Likert-type scale ranging from 1=None to 5=All. These include information about: diagnosis (items 1-4), treatment (items 5-7), laboratory testing (items 8-10), self-care (items 11-13), CAM (items [14][15][16], psychosocial aspects (items [17][18][19] and health-care providers (items 20-21). The original English version of the 21-item HIWQ is freely available online. 21 A bilingual translator translated these original HIWQ items into Chinese, which was then verified and revised by the first author for accuracy and consistency. To answer the research questions of this study, the following new survey questions were added to each of the original 21 items: How much of this type of the information have you already obtained from (i) a doctor/nurse; (ii) a family/friend/ neighbour/acquaintance; (iii) newspaper/magazine/television/radio; and (iv) the Internet? The format of the instrument was also revised to accommodate these new questions. Table 1 below illustrates a sample item with the original formatting from the original English instrument, while Table 2 illustrates the same sample item from the revised Chinese instrument.
To determine any problematic items of the Chinese instrument, we conducted multiple rounds of testing using the cognitive interviewing technique, which is commonly used in fields such as health sciences and psychology to validate questionnaire items among intended populations. 27,28 Our cognitive interviewing was done with, first, a total of nine graduate students and visiting scholars that were all native Chinese speakers studying in the United States. Next, we conducted cognitive interviewing with two older Chinese women living in Beijing, followed by testing with three medical students interning at the Hospital. Finally, we conducted two rounds of cognitive interviewing with seven participants from the Hospital (four patients with cancer and three family caregivers; data from these participants were excluded from our final data set). We revised the instrument based on the results from each round of the cognitive interviewing. This Chinese instrument has excellent validity and reliability, as reported in detail in our first paper from this study. 26 A full copy of the questionnaire used in this study can be obtained upon request to the first author.

| Participants
Participants were recruited from patients of our research site, and family caregivers accompanying these patients during the time of data collection, March 2014. Inclusion criteria included the following: (i) at least 18 years of age; (ii) able to read, write and speak Chinese; and (iii) normal hearing and vision, corrected or uncorrected. One of the authors (MZ), an oncology physician with 20 years of experience working at the Hospital, recruited participants through word-of-mouth. A total of 79 patients with cancer and 119 caregivers completed the instrument (N=198; some patients were accompanied by more than one family caregiver; we allowed multiple caregivers of a patient to participate if they all consented to). gave each a small hand towel as our token of appreciation.

| Data analysis
Three-way mixed ANOVA was conducted in the SPSS 22.0 software, (IBM Corp., Armonk, NY, USA), with one between-subject factor (group: patents vs caregivers) and two within-subject factors (type of information; source of information). Consistent with our previous research, 22 participants' original responses to all HIWQ items were rescaled to range from 0 (corresponding to the least amount of information obtained) to 100 (corresponding to the most amount of information obtained) with a mid-point of 50. The formula used for the rescaling was as follows: rescaled score=(raw score−1)*25. Then, we averaged the rescaled items measuring the same type of health information and created a scale score for each type of health information obtained from each source (i.e. 28 scale scores were created for each participant, representing the seven types of health information obtained from each of the four sources of study).
Additionally, to further examine the relationship between general Internet use and the amount of information patients and caregivers project. 26 The amount of the seven types of information obtained by patients and by caregivers from the four sources is reported in Table 4 below.
Our three-way ANOVA found a significant interaction between source of information and group (F 3,576 =6. 32 15). Therefore, patients and their family caregivers did not differ in the total amount of information they obtained (Table 5) These findings are illustrated below in Fig. 1 (Table 6).
Pair comparison tests showed that the significant effects were mainly driven by using the Internet or not, as indicated by the signif-  even among those who also seek health information online. 9,29 Recent evidence suggests that in developed countries such as the United States the Internet has become the second most prominent source of information (after medical professionals, and more prominent than mass media or interpersonal contacts). 7,9 However, among our T A B L E 5 Summary of the three-way ANOVA results

F-value/D P-value Effect size
Major effects Chinese participants, the Internet was the least used source of information. Plausible reasons include, first, compared with the vast amount of high-quality online information available to English speakers, much less such information is available in Chinese. [30][31][32] Second, compared with their Western counterparts, Chinese participants may have lower levels of computer/Internet access and literacy. 33 In fact, over 90% of our participants reported having very low-to-medium household income, and more than half of the patients and roughly one-third of the  Our findings also suggest that Chinese patients with cancer and their family caregivers differed in their use of one of the health information sources: the Internet, with caregivers obtaining more information from the Internet than did patients (these two groups did not differ in the amount of information they were able to obtain from doctors/nurses, interpersonal contacts, or mass media). This finding is not surprising given that, compared with their caregivers, the patients in this study were significantly older and had lower levels of overall health condition, education, household income, and Internet use history and frequency, with these factors being shown to be negatively correlated to use of the Internet for health information. 16 This finding is also in line with those reported in the literature, 36 supporting its generalizability across populations and settings.

| CONCLUSIONS
Our findings suggest that the health information sources of Chinese patients with cancer and their family caregivers were similar, with an important difference that, compared with patients, caregivers obtained more health information from the Internet. Still, for both the patients and caregivers groups, the Internet was the least used source of information compared with the other three sources of information.
Importantly, while the amount of treatment and laboratory testing information participants obtained from mass media did not differ from that obtained from the Internet, they obtained more information of the other five types from mass media than from the Internet. These findings have important implications for patient care and education in China particularly as Chinese families typically play a major role in the care and decision making. 'D'=mean difference across conditions; For Internet use history, 1='Never' or 'Less than a year', 2='1-3 y' or '3-5 y', and 3='5-10 y' or 'more than 10 y'; For Internet use frequency, 1='Never', 2='Less than once per month', 'More than once per month', 'Once per month', or 'Once per 2-3 days', and 3='Everyday'. *P<.05, **P<.01.