Seeking or contributing? Evidence of knowledge sharing behaviours in promoting patients’ perceived value of online health communities

Abstract Background Health knowledge, as an important resource of online health communities (OHCs), attracts users to engage in OHCs and improve the traffics within OHCs, thereby promoting the development of OHCs. Seeking and contributing health knowledge are basic activities in OHCs and are helpful for users to solve their health‐related problems, improve their health conditions and thus influence their evaluation of OHCs (ie perceived value of OHCs). However, how do patients’ health knowledge seeking and health knowledge contributing behaviours together with other factors influence their perceived value of OHCs? We still have little knowledge. Objective In order to address the above gap, we root the current study in social cognitive theory and prior related literature on health knowledge sharing in OHCs and patients’ perceived value. We treat health knowledge seeking and health knowledge contributing behaviours as behavioural factors and structural social capital as an environmental factor and explore their impacts on patients’ perceived value of OHCs. Design We have built a theoretical model composed of five hypotheses. We have designed a questionnaire composed of four key constructs and then collected data via an online survey. Setting and participants We have distributed the questionnaire in two Chinese OHCs. We obtained a sample of 352 valid responses that were completed by patients having a variety of conditions. Results The empirical results indicate that health knowledge seeking and health knowledge contributing have positive impacts on patients’ perceived value of OHCs. The impact of health knowledge seeking on patients’ perceived value of OHCs is greater than the impact of health knowledge contributing. In addition, structural social capital moderates the effects of health knowledge seeking and health knowledge contributing on patients’ perceived value of OHCs. It weakens the effect of health knowledge seeking but enhances the effect of health knowledge contributing on patients’ perceived value of OHCs. Conclusions These findings contribute to the literature on patients’ perceived value of OHCs and on the role of structural social capital in OHCs. For OHC managers, they should provide their users more opportunities to seek or contribute health knowledge in their communities.


| INTRODUC TI ON
As information and communication technologies (ICTs) have become ubiquitous, people have become increasingly active in online health-related applications such as online health communities (OHCs). OHCs are a type of health-related virtual community (VC) designed particularly for different health-related stakeholders, for example health insurance, pharmaceutical companies, hospitals, health professionals, patients, and patients' relatives or friends. [1][2][3][4][5][6][7][8][9] There are different types of OHCs in which users can conduct different health-related activities, including transactions, appointment scheduling, counselling, social networking and health-related Q&As. 1,[10][11][12][13][14][15][16] In this study, we particularly focus on problem-solving communities where both health professionals and patients can participate and collaborate for health knowledge exchange, for example Q&A forums on health conditions, 13,17 mental health-focused Q&A forums, 18,19 pregnancy forums such as Babytree.com, 20 and cancer-focused communities. 3,21,22 In this type of OHCs, health professionals can provide professional health knowledge by contributing to the community and responding to patients' health-related questions. 20,23 Patients can disclose their personal health conditions, make new social ties, and seek or contribute health knowledge. 1,6,24,25 Using OHCs can help health professionals build their reputations and earn material rewards 12,13,19,26 and help patients improve their health outcomes, such as their e-health literacy and feeling of well-being. [27][28][29] These advantages make OHCs an effective way to alleviate the pressures on medical resources. 30,31 Exchanging health knowledge and information is a kind of basic activity in OHCs. 3,8,9,20,21,[32][33][34] Health knowledge in OHCs is a public good, and contributors lose their control over the knowledge they shared. [35][36][37] Scholars thus are curious about the reasons why people contribute health knowledge in OHCs. For example, some studies on health professionals have examined the impacts of factors such as professional capability, reputation and economic rewards. 7,8,19,20,31,38,39 Other studies, focused on patient users, have examined the impacts of extrinsic and intrinsic motivations [4][5][6]20,33,34,40 and potential hindering factors such as trust and privacy protection. 6,24,30 As social networking is an important feature of OHCs and users in OHCs also pursue social interactions, 1,3,22 some studies have also explored the impacts of users' social capital in OHCs. 13,28,29 In addition to exploring its antecedent factors, scholars recently began to explore the health outcomes of exchanging health knowledge in OHCs. For example, knowledge seekers can obtain health knowledge for their health issues and then use it to improve their health conditions. 9,41,42 Knowledge contributors also can obtain new knowledge because they have to understand other questions and then combine different knowledge to address those questions 9,29 ; this process could help them to create new knowledge.
OHC use therefore positively promotes users' health outcomes such as health conditions, health attitude and e-health literacy. 28,29,43 As a kind of health outcome, users' perceived value is a crucial antecedent for users' satisfaction with OHCs and also their continuous use of OHCs. 44,45 In this paper, we will clarify how users' perceived value of OHCs is a crucial antecedent of their satisfaction with OHCs, their continuous use of OHCs and related health outcomes in section Literature Review. As discussed above and in section Literature Review, few studies have explored how health knowledge exchanging behaviours and other factors influence users' perceived value of OHCs.
In order to address the above gap, we adopted social cognitive theory (SCT) as theoretical foundation. We treated health knowledge seeking and health knowledge contributing as behavioural factors and structural social capital as a social environmental factor, and, finally, built a model composed of five hypotheses. We tested our hypotheses with a sample of 352 valid responses.

| Social cognitive theory
Social cognitive theory (SCT) is a classical theory on individual behaviours. According to SCT, personal behaviours are shaped by the factors from three domains (ie environment, cognition and behaviour); the factors from any two domains can interact with each other and then influence the factors in the third domain. 46 For example, interactions between environmental and behavioural factors, which can be treated as parts of social environments, influences an individual's cognitions and, in turn, reshapes their behaviours and external environment. 46 In addition to being used to explain personal knowledge sharing behaviours in VCs, 37,40,[47][48][49] SCT also has been used to analyse change in personal cognition. For example, environmental factors such as trust and interaction positively influence personal cognitive Conclusions: These findings contribute to the literature on patients' perceived value of OHCs and on the role of structural social capital in OHCs. For OHC managers, they should provide their users more opportunities to seek or contribute health knowledge in their communities.

K E Y W O R D S
health knowledge sharing, health services, online health communities, social interaction, social support, telemedicine factors such as outcome expectation. 37,48 Personal health knowledge seeking behaviours together with environmental factors (eg structural social capital) positively influences cancer survivors' e-health literacy. 28 The above studies indicate that the change in personal cognitive factors could be explained by SCT. Since we focus on how behavioural factors and environmental factors influence patients' perceived value of OHCs, we therefore adopt SCT as a theoretical foundation. We propose that patients' knowledge sharing behaviours (ie behavioural factors) together with their structural social capital (ie social environmental factor) influence patients' perceived value of OHCs.

| Health knowledge sharing in OHCs
Following prior studies, 37,49-51 we define health knowledge sharing as a process composed of two aspects: health knowledge seeking and health knowledge contributing. Health knowledge in OHCs includes people's physical health, mental health, diseases and nutrition, such as hospital or doctor information, healthy life and behaviours, medicine information, personal health conditions, medical treatments and medical experiences. 34,52 Health knowledge seeking refers to the search, acquisition or consumption of health knowledge in OHCs. 51 Health knowledge contributing refers to the generation or provision of health knowledge in OHCs. 53 We reviewed prior studies on health knowledge sharing in OHCs and summarize the results in Table 1.

| Patient social capital
Social capital is defined as the sum of the actual and potential resources that an individual obtains from the network of relationships. 57 Social capital can be divided into three dimensions: structural social capital, relational social capital and cognitive social capital. 57 OHCs are online health-related social networks in which users with common interests, goals or practices interact to contribute and seek health knowledge and engage in social interactions. 1,26,47 It is the nature of social interactions and the resources embedded in social interaction networks that sustain the OHCs. 26,47 Therefore, in addition to health knowledge resources, users' structural social

| Patients' perceived value
Perceived value of OHCs is defined as patients' perception of the overall utility based on a trade-off between perceived benefits and costs of using OHCs. 63 Studies considering the direct studies on patients' perceived value of OHCs are few, so we summarized prior related studies on user-perceived value for our reference in this study (see Table 3).

| RE S E ARCH MODEL AND HYP OTHE S IS DE VELOPMENT
This study aimed to examine how patients' knowledge sharing be-  Figure 1. Note: Relationships between independent variables and dependent variables are shown after each independent variable (ns: not significant; +: positive; -: negative).

| Main effects
Health knowledge seeking is defined as users' search, acquisition or consumption of health knowledge in OHCs. 51 Under the user-generated content mechanism, health professionals and pa- Such a process enhances the contributors' understanding of health knowledge 29 and supports their learning of new knowledge in this collaborative consumption process. 9 Second, health knowledge contributing behaviours enriches health knowledge in OHCs and meets seekers' needs of health knowledge that is useful to solve their health-related issues. 41 Health knowledge seekers in turn are more likely to express their gratitude to the contributors. 29 In above process, contributors could develop close relationships with other users and obtain a sense of self-worth from other users' gratitude. 20,23,34 Namely, health knowledge contributing behaviour is beneficial for users to improve their evaluation of the utility of OHC use (ie perceived value). We thus hypothesized that,

| Moderating effects
According to SCT, individuals' behaviours together with environmental factors can reshape their cognitions. 46 OHCs enable knowledge seekers to obtain relevant knowledge to improve their health conditions. 42 High structural social capital usually means users have more social contacts. When seeking health knowledge in OHCs, patients with high structural social capital may receive massive replies and the useful knowledge, thus might be overwhelmed by those that are useless. 79 They need to devote a lot of time and energy to distinguish useful information from useless ones and are more likely to experience negative emotions such as anxiety and depression. 80 The added unnecessary costs would make them underestimate their perceived value of OHC use. In addition, higher structural social capital also means that patients have diversified channels to seek their needed health resources and obtain more value. 57 They can find what they need through other activities such as through personal channels instead of through public postings. When users use more personal channels to seek knowledge, they will rely less on health knowledge seeking in OHCs. Their per- contributors with a higher level of structural social capital, their knowledge could be exposed to more users and therefore receive more gratitude. The positive feedback and experience obtained in above process will enhance contributors' sense of self-worth. 20,23,34 Their perceived value of OHC use derived from health knowledge contributing behaviours thus will be enhanced. We thus hypothesized that,

H3b: Structural social capital enhances the impact of health knowledge contributing on patients' perceived value of OHCs: when structural social cap-
ital is high, the effect of knowledge contributing behaviours will be stronger.
In addition to the above variables, prior studies have found that women are more likely to continue participating in sharing, 81 age has a negative effect on users' participation behaviours in VCs, 21 education has positive effect on users' health knowledge contribution, 23 and tenure has positive effect on users' information-seeking behaviours in VCs. 82 We thus proposed that gender, age, education and tenure also might influence patients' perceived value of OHCs, and we treated them as control variables.

| ME THODOLOGY
We designed a questionnaire and an online survey for data collection and hypothesis test. This research was approved by the Shantou University Academic and Ethics Board.

| Constructs and scales
All scales for our four key constructs were adopted from prior research and adapted to the OHC context. We took the following pre- we also did a pilot study by inviting 12 undergraduate students who have OHC use experience to complete the questionnaire. During the process, we asked them to tell us any confusing issues and then modified them accordingly. The questionnaire was frozen when the back-and-forth translation and pilot test were completed. We used a 5-point Likert-type scale (note: 1 for completely disagree and 5 for completely agree). Table 4 shows the final items of all constructs.

| Results of measurement model assessment
We assessed the measurement model with explorative factor analysis using SPSS 20 (see Table 6) and confirmative factor analysis using Mplus 7.4 (see Table 7).
For convergent validity, as shown in Table 6 validity. In addition, as shown in Table 7, the AVE square root value of one variable is greater than the correlation value between this variable and the other three variables. These indices indicate a good discriminant validity.
We also checked the potential collinearity issues in three different ways. First, the eigenvalue of every single independent variable is not equal to 0 and the greatest conditional index value is 3.289 that is less than 20, 84 Second, the greatest variance inflation factor (VIF) value is 2.784 which is less than the suggested value 10. 85 Third, the correlation value between health knowledge contributing and structural social capital is 0.787 which is less than the cut-off value 0.8. 86 Therefore, the multicollinearity has no serious effect on the empirical results.
We also tested the model fitness (see Table 8). All indices are at or over the acceptable level, indicating the model fitness is good. 37

| Results of structural model assessment
Although the correlation value among different variables meets the cut-off value 0.8, 86 they are still slightly high. In such a situation, structural equation modelling using latent variables works better. 87,88 We therefore used the latent moderated structural equations (LMS) approach via Mplus 7.4 to test all hypotheses (see Figure 2).
As shown in Figure 2, the effects of health knowledge seeking

| Limitations
There are several limitations that may affect the findings in this study. First, our sample size is relatively small. Empirical findings might be more robust with a larger sample. Second, we built a concise model that includes three antecedents (ie health knowledge seeking, health knowledge contributing and structural social capital).
We did not include the factors such as types of health knowledge, 32 characteristics of health care, 26 type of patients' illnesses and characteristics of OHCs that might influence patients' perceived value of OHCs. Including these variables, especially the characteristics of health care and OHCs, could capture the impacts of contextual factors and therefore might have interesting findings. We address the lack of examining the impacts of these factors as a limitation of this study. We appeal to scholars to pay more attention to these factors and explore their impacts on patients' perceived value of OHCs in future studies.

| CON CLUS IONS
We posit that patients' perceived value of OHCs is influenced by

CO N FLI C T O F I NTE R E S T
The author declares that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.