Search

Objective Research has convincingly demonstrated that Internet interventions are effective for anxiety and depression and it is suggested that these interventions can reduce therapist time and thus save costs. Therefore, many Dutch mental health organizations have started to implement these treatments into their services. They usually apply the Internet interventions in combination with the face-to-face treatments, i.e. blended treatments. To date, no empirical work has examined if implementing blended treatments actually contributes to treatment effects and cost savings. The objective of this naturalistic study was to examine this issue, using routine mental health care data for patients with depression or anxiety. Methods All 4448 records of patients with depression or anxiety seeking help at a large mental health service between October 2009 and December 2012 were obtained. Data on GAF-index, therapist time, and therapist costs of patients who received face-to-face treatments were compared with those who received blended care, using propensity score matching. Results Blended care was applied in 3.6% of all patients and was given more frequently to younger patients and patients with an anxiety disorder. There were no significant differences in symptom severity at baseline between the two treatment groups, nor were there significant differences between the two groups in symptom improvement. Patients who received blended care had more treatment sessions (face-to-face plus online sessions) than patients who received standard treatment, which resulted in significantly more treatment time and higher costs. Conclusion Findings suggest that in this sample blended care is more expensive compared to established face-to-face treatments and doesn't lead to additional benefits in terms of general functioning. This might be caused by suboptimal implementation strateges.

An important notion for representation formalisms of natural language semantics, is a subsumption hierarchy. Therefore a precise definition of subsumption is necessary. We shall argue that the usual solution of providing an extensional semantics and mapping subsumption onto set-inclusion, is not satisfactory. The problem is that extensions lose track of the structure. A better solution is used for conceptual graphs [13], where derivation rules define generalization.