Page - 156 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

than once. The primary goal of the system was to encourage developing a written Asthma Action Plan, and at the end of the study only 18% of intervention group participants had one (compared to 22% of the control group). Based on a follow-up survey, the authors cite unrecognized and unaddressed barriers as the primary reason for the poor effect of the system: for example, believing that a written plan was not necessary in their situation, being discouraged from getting one by a health care professional or by a previous negative experience with trying to create a plan, or feeling that it was unimportant compared to other priorities in their life. Research using these theories, or theories in general, falls into two categories: studies which test a theory in an applied setting to determine if the theory applies to that setting, or studies which aim to solve a practical problem and make use of the theory to formulate a better solution [26]. The attempt to apply Fogg's theory to the NHS organ donation website can be viewed as a study of the first type. However, the outcome might be considered predictable: the target behaviour was very different than the behaviours in the studies from which the theory was derived. Once the decision has been made to become an organ donor, the user may not be deterred by being required to use an unprofessional-looking website – unlike the decision to purchase a book or music, which can always be purchased later or from another vendor. This also explains why self- efficacy needed to be added to the Health Belief Model; it was originally developed to explain simple behaviours like getting a vaccination, and needed an additional construct to be applied to more complex behaviours where the person might doubt their own ability to engage in the behaviour. Likewise, a theory may not apply if the target group is very different from the original target group or if the target behaviour is influenced by forces not accounted for in the model. For example, clinicians' behaviour is usually regulated in part by law and reimbursement policy, so it is unlikely that any behaviour change intervention would cause clinicians to behave in a way that substantially contradicted these policies. Likewise, interventions to change a health behaviour in school children or in military personnel might be quite different, since many choices in these groups are not made by the individuals themselves. A thorough review of the literature may be informative to determine if a theory has been successfully applied for that behaviour, type of person and context of interest. Several of the models discussed above have the shortcoming that they consider the behaviour of a person in isolation, disregarding social influences. These models are unlikely to apply in situations where social interaction plays a large role, or when the decision is made by a team rather than an individual. Finally, as mentioned above, correct operationalization of the theory is critical. This is likely the reason for underuse of the asthma website mentioned above: although barriers are a construct in the Health Belief Model, the investigation of barriers prior to developing the system failed to identify the barriers, which ultimately led to the system not being used. Use of a guide such as that proposed by Kok et al. can help assure that the theory is applied correctly [18]. 4. Discussion Health behaviour and behaviour change theories are widely applied in the field of psychology and have now been classified and translated into a taxonomy of practical techniques. Many health informatics interventions, particularly those directed toward healthcare professionals and patients, aim to change user behaviour in some way, and use of a behaviour change theory in intervention design or participant selection is S.MedlockandJ.C.Wyatt /HealthBehaviourTheory inHealth Informatics156