Served association.Strengths and Limitations of the AnalysisThe study used routine TA01 chemical information healthcare data which allowed the analysis of a long time series in a large dataset, but suffers the limitations that all such studies do in terms of the data potentially being incomplete because it was collected for another purpose. A particular issue is that dementia is known to be under-recorded historically (although Scottish recording is reasonably close to epidemiological predictions) [23]. The quarter 1 2011 dementia prevalence in this study was 4.2 in people aged 65 and over, compared to estimates of 6.6 and 6.4 from the largest UK study and an Europe-wide meta-analysis respectively [24]. However, the age-standardised prevalence of dementia in peopleComparison with Other StudiesThree North American studies have examined the impact of regulatory risk communications on antipsychotic prescribing [8,25,26]. In Canada, three regulatory risk communications in the period 2002?005 reduced the rate of growth of antipsychotic prescribing in people with dementia and caused some shift from risperidone and olanzapine to quetiapine [25], but total antipsychotic prescribing in older people continued to increase [27]. Two US studies of the impact of the 2005 FDA risk communications showed falls in antipsychotic use in older people with dementia [8,26], but there was little immediate impact on the scale observedRisk Communications and Antipsychotic PrescribingFigure 3. New antipsychotic prescribing and antipsychotic stopping in people aged 65 years with dementia. doi:10.1371/journal.pone.0068976.gin the 23148522 study reported here in association with the 2004 risk communication. To our knowledge, there are no published studies of subsequent regulatory risk communications in this field. Kales et al’s study in the Veterans’ Administration population also examined the use of other psychotropics, finding no change in hypnotic, anxiolytic or antidepressant use [8]. In contrast, our study shows that antidepressant prescribing rose considerably over the whole period. Although we found some evidence of transient substitution of other psychotropics for antipsychotics in 2004, the more striking finding was that prescribing of hypnotics, anxiolytics and antidepressants either flattened off or declined after the 2009 risk communication. This highlights that regulatory risk communications may have unexpected effects beyond the prescribing targeted, and evaluation should ideally seek to examine unintended as well as intended consequences [28,29]. The NHS England national prescribing audit published in July 2012 showed a reduction in the proportion of older people with recorded dementia prescribed an antipsychotic from 17.0 in 2006 to 6.8 in 2011, [30] compared with the observed reduction in this study from 16.9 in quarter 1 2006 to 13.5 in quarter 1 2011. In England, the 2009 risk communication was reinforced by a Department of Health commitment to reduce antipsychotic prescribing in older people with dementia by two-thirds over two years [13] [18]. In contrast, there was no such clear policy response in NHS Ion, in human genetic studies, IRAK-M has also been associated with Scotland. The greater observed fall in antipsychotic prescribing in England is consistent with there being an additional impact of the policy response over and above the risk communication directed at the whole UK. However, it is important to note that the number of people with recordeddementia in the English audit more than doubled since 2006, compared with an ,33 increas.Served association.Strengths and Limitations of the AnalysisThe study used routine healthcare data which allowed the analysis of a long time series in a large dataset, but suffers the limitations that all such studies do in terms of the data potentially being incomplete because it was collected for another purpose. A particular issue is that dementia is known to be under-recorded historically (although Scottish recording is reasonably close to epidemiological predictions) [23]. The quarter 1 2011 dementia prevalence in this study was 4.2 in people aged 65 and over, compared to estimates of 6.6 and 6.4 from the largest UK study and an Europe-wide meta-analysis respectively [24]. However, the age-standardised prevalence of dementia in peopleComparison with Other StudiesThree North American studies have examined the impact of regulatory risk communications on antipsychotic prescribing [8,25,26]. In Canada, three regulatory risk communications in the period 2002?005 reduced the rate of growth of antipsychotic prescribing in people with dementia and caused some shift from risperidone and olanzapine to quetiapine [25], but total antipsychotic prescribing in older people continued to increase [27]. Two US studies of the impact of the 2005 FDA risk communications showed falls in antipsychotic use in older people with dementia [8,26], but there was little immediate impact on the scale observedRisk Communications and Antipsychotic PrescribingFigure 3. New antipsychotic prescribing and antipsychotic stopping in people aged 65 years with dementia. doi:10.1371/journal.pone.0068976.gin the 23148522 study reported here in association with the 2004 risk communication. To our knowledge, there are no published studies of subsequent regulatory risk communications in this field. Kales et al’s study in the Veterans’ Administration population also examined the use of other psychotropics, finding no change in hypnotic, anxiolytic or antidepressant use [8]. In contrast, our study shows that antidepressant prescribing rose considerably over the whole period. Although we found some evidence of transient substitution of other psychotropics for antipsychotics in 2004, the more striking finding was that prescribing of hypnotics, anxiolytics and antidepressants either flattened off or declined after the 2009 risk communication. This highlights that regulatory risk communications may have unexpected effects beyond the prescribing targeted, and evaluation should ideally seek to examine unintended as well as intended consequences [28,29]. The NHS England national prescribing audit published in July 2012 showed a reduction in the proportion of older people with recorded dementia prescribed an antipsychotic from 17.0 in 2006 to 6.8 in 2011, [30] compared with the observed reduction in this study from 16.9 in quarter 1 2006 to 13.5 in quarter 1 2011. In England, the 2009 risk communication was reinforced by a Department of Health commitment to reduce antipsychotic prescribing in older people with dementia by two-thirds over two years [13] [18]. In contrast, there was no such clear policy response in NHS Scotland. The greater observed fall in antipsychotic prescribing in England is consistent with there being an additional impact of the policy response over and above the risk communication directed at the whole UK. However, it is important to note that the number of people with recordeddementia in the English audit more than doubled since 2006, compared with an ,33 increas.
