Definition of 1. network
The broad definition of "social network" is social interaction and personal relationships, such as networks with friends, school friends, non-school friends, family members, sexual partners, colleagues at work, neighbors and * * * who enjoy unhealthy behaviors (such as drug abuse and risky behaviors). Social relations need to be measured and must have been used to intervene in some aspects of design or implementation.
The measurement of relationship must go beyond the level of duality, including the connection between multiple participants (that is, intervention measures to exclude the interaction between spouses or a pair of friends). Interventions must change health behaviors or health outcomes (or their substitute indicators) in a targeted manner.
Social network intervention has successfully covered, retained and changed the behaviors of so-called hidden, inaccessible and high-risk groups, including men who have sex with men, injecting drug users and other priority groups (for example, low-income or ethnic minorities).
2. Network intervention
Network intervention is to use existing networks, establish new networks, destroy harmful networks or use social network data to educate participants about the potential impact of their health behaviors on their network members (including real and online social networks). Network intervention supports various types of health promotion (for example, health communication, family or organization methods), and expands the coverage of existing interventions or improves their effectiveness.
Personal network intervention includes those behaviors that specifically use network data to identify individuals to be recruited. These individuals act as supporters of behavior change according to certain network attributes. Interventions to subdivide the network include those aimed at people gathered in the network. Inductive network intervention includes stimulating and activating existing social connections in social networks to spread information or healthy behaviors. Interventions to change the network include changing the structure of the network by adding new members or cutting off contact with those who promote unhealthy and dangerous behaviors.
(1) Individuals selected according to network attributes participate in it, which may be opinion leaders. An informal school-based peer-guided adolescent smoking prevention intervention (ASSIST): a cluster randomized trial. The Lancet. 2008
(2) People interacting with specific groups were randomly tested to test the effect of peer education in increasing the intake of fruits and vegetables. National cancer institute. 1999.
(3) those processes that encourage or enhance peer-to-peer interaction and cascade information and influence to other network members (called induction process; Internet HIV prevention intervention by peer educators of injecting drug users: results of a randomized controlled trial in St. Petersburg, Russia. Aids behavior.
(4) It involves changing the network by adding or deleting members, adding or deleting specific social relations or changing the whole network. Internet support for changing drinking: preliminary findings of the network support project. Consulting clinical psychology.
3. Theoretical model
This kind of research has some social science implications, so there must be some theoretical models. To sum up, behavior change models include social cognition theory, rational behavior theory, planned behavior theory, health belief model, social norm theory, social influence, social ecological model, social learning theory and social identity theory. For network elements, the most cited theory is communication innovation theory.
4. Some explanations
(1) High participation rate and retention rate are crucial to these interventions, which aim to activate and stimulate the network and social environment mechanisms for promoting health. Theoretically, this intervention can avoid high-risk behaviors by redefining social norms. Usually, social norms are spread to others in the network by observing the behaviors of popular peers and then adopting and modeling these behaviors. A randomized, controlled and community-level HIV prevention intervention aimed at the sexual risk behaviors of urban gay men in the United States. Community AIDS prevention research cooperation organization. The Lancet 1997)
(2) Behavior communication in online social network experiments. Science 20 10/0 needs to be strengthened.
(3) I didn't expect the experimental design to be sent to the machine learning conference. Toulis P, Kao E. Estimation of causal peer influence effect [C]// International Conference on Machine Learning. PMLR，20 13: 1489- 1497。 Two other factors are also considered in the evaluation of confounding effect: the causal effect in the presence of interference. Statistical methods in med RES. 20 12// interference causality diagram. Statistics for 2000.
(4) Time effect has a great influence on this kind of research. The results of 6 months, 12 months and 24 months are quite different, which is also the main result of this meta.
(5) Personal intervention is the most effective.
(6) Social identification, social support, social communication and social learning process. Social contagion theory shows that health outcomes, behaviors and beliefs (such as obesity and happiness) "spread" through social networks. The application of network theory and social ecological model is based on individual model and how social factors affect individual biology, beliefs, decisions and behaviors. Traditionally, health behavior intervention can not successfully maintain new behaviors for a long time. The important role of social network in behavior maintenance is obvious in habit formation theory (emphasizing the importance of external clues) and behavior change theory (emphasizing the role of social norms in enhancing new behaviors).
(7) Social network with the goal of maximizing crowd behavior change: a cluster randomized controlled trial. Lancet.20 15 A method that can be experimented without knowing the whole network.
(8) In areas with limited resources and infrastructure, network intervention may be particularly useful. The network can be used to increase the number of participants or promote the dissemination and adoption of these interventions.
(9) The method of network science identifies peripheral nodes, bridge nodes and early adopters. Multi-stage optimization strategy (MOST) uses factorial experimental design to determine the unique effects and interactions of specific intervention components, which may include social networking methods or other effective components (for example, educational materials). It is helpful to provide a basis for evaluating the effectiveness and cost-effectiveness of network intervention and the best design. Limitations include dependence on self-reported results, inherent memory and expectation bias, and the results obtained in key population research may limit the universality of the research results to other populations or groups. Multi-stage optimization strategy (MOST) and sequential multi-task randomized trial (SMART): new methods for more effective electronic health intervention. Doctor of medicine. 2007
(10) Network phenomenon is inherent in our intervention measures, but it is ignored and not fully utilized at present. That is to say, the effect of the network should be considered when designing RCT, which may also be a kind of confusion.
summary
To consider the intervention time, online social network intervention is still a blank point, and the mixed correction and evaluation methods produced by network research are very important. MOST may be a better research and design method of network intervention at present.