Primary Care Physicians ’ Use of Antipsychotics for the Treatment of Behavioural and Psychological Symptoms of Dementia in Japan

Objective: We aimed to investigate how primary care physicians (PCPs) in Japan use antipsychotics for treating the behavioural and psychological symptoms of dementia (BPSD). Materials and Methods: A nationwide cross-sectional online survey was conducted of PCPs who annually treated patients aged over 65 years with dementia. Responses from 509 PCPs were analysed by binominal logistic regression analysis. Results: Approximately one-third of the PCPs had treated excitatory BPSD (delusion, hallucination, agitation and violence), with half of them prescribing antipsychotics for this. Some PCPs still prescribed antipsychotics for non-excitatory BPSD, such as wandering. More than half of the PCPs had opportunities to learn about the appropriate use of antipsychotics and understood the increased mortality risk in elderly people with dementia. Referring to the Japanese Governmentʼs BPSD guideline for PCPs was negatively associated with antipsychotic dosage (odds ratio=0.491, 95% confidence intervals 0.32-0.75, p-value=0.001) and positively associated with a greater awareness of increased mortality with antipsychotics (odds ratio = 2.149, 95% confidence intervals 1.41-3.27, p-value=0.0004). Conclusion: PCPs continue to prescribe antipsychotics for excitatory BPSD in clinical practice despite official information about mortality risks. Educational material about the appropriate use of the antipsychotics should include specific mention of the risks of using antipsychotics for dementia.


Introduction
It is estimated that 44 million people worldwide suffer from dementia, including Alzheimerʼ s disease, vascular dementia, Parkinsonʼs disease dementia, frontotemporal dementia and Lewy Body dementia, and this number continues to increase 1) .Japan has one of the highest life expectancies in the world, with elderly people with dementia comprising an increasing proportion of the total population 2) .In 2015, Japanʼ s Ministry of Health, Labour and Welfare (MHLW) published its Comprehensive Strategy to Accelerate Dementia Measures (known as ʻthe New Orange Planʼ) to promote dementia measures in a holistic way 3) .This includes establishing specialised facilities, such as dementiarelated medical centres, in each local community.However, Japan has only a limited number of experts who can treat patients with dementia; they mainly work in psychiatry, neurology and neurosurgery.Primary care physicians (PCPs) are the doctors who offer general medical care in local communities separate from their specialty departments, such as internal medicine or surgery.PCPs often have opportunities to treat elderly people with dementia before they visit dementia care experts.If a PCP find it difficult to look after a patient with dementia, he or she can refer the patient to a dementia-related medical centre for more specialised treatment.MHLW expects PCPs to have an important role in dementia medical treatment 4) .
Among the symptoms that require the most care in the provision of dementia medical treatment are the behavioural and psychological symptoms of dementia (BPSD) 5) .These include depression, delusions and hallucinations, agitation, violence, wandering, sleep disorder and other manifestations 6) .Some guidelines indicate that physicians should take a psychosocial approach as the first-line treatment for BPSD, with pharmacological treatment as the second-line approach 7) .However, in a clinical setting, the situation often arises where antipsychotics are prescribed to treat intense symptoms out of necessity 8) .In particular, antipsychotics are often considered to manage excitatory BPSD (hallucination, delusion, agitation and violence).Recently, there have been reports of increased morbidity and mortality when patients with dementia take antipsychotics 9) -11) .Following a report in 2005 that reviewed antipsychotic use 12) , the US Food and Drug Administration announced caution about prescribing conventional and atypical antipsychotic medication to elderly people with dementia-related mental disease because of the increased mortality risk 13) 14) .A Japanese prospective cohort study also reported increased mortality in patients with dementia who had newly received antipsychotic drugs 15) .In addition, the cost of care is a serious problem in dementia medical treatment, with the total worldwide societal costs of elderly with dementia estimated to be US$604 billion in 2010, imposing a tremendous economic burden on society 16) ; the inappropriate use of antipsychotics for BPSD may contribute to this burden.It is therefore necessary for PCPs as well as dementia experts to understand the risks of prescribing antipsychotics to people with dementia.The aim of this study, therefore, was to investigate the practice of PCPs in Japan with regard to prescribing antipsychotics to treat BPSD.

Materials and Methods
We conducted a cross-sectional questionnaire nationwide online survey of Japanese PCPs in 2015.This was administered by M3, Inc. (Tokyo, Japan) on their website (M3.com).The questionnaire excluded doctors working in specialty departments (psychiatry, psychosomatic medicine, paediatrics, obstetrics, radiology, anaesthesia and the emergency department), who were not considered to be PCPs.The study protocols were carried out under the Ethical Guidelines for Epidemiological Research (17 June 2002, Ministry of Education, Culture, Sports, Science and Technology and MHLW).After a complete description of the current study and the right to withdraw consent to participate, written informed consent was obtained by electronic signature from all respondents according to the principles of the Declaration of Helsinki (64th, 2013).
In the questionnaire (Table -1), we asked about how PCPs used antipsychotic drugs (regardless of subtypes and dosing regimens of antipsychotics) for BPSDs and how aware they were of the risks of using them.From the questionnaire responses, we investigated several conditions: BPSDs that PCPs targeted for treatment, BPSDs that PCPs prescribed antipsychotics for and the voluntary actions taken by PCPs with regard to BPSD care.Unadjusted and binominal logistic regression analyses were performed to identify factors associated with two dichotomous dependent variables: whether or not PCPs prescribed antipsychotics for elderly people with dementia and whether or not PCPs were aware of the MHLWʼ s warning about the increased risk of mortality from the use of antipsychotic drugs by elderly people with dementia.We simultaneously introduced groups of variables into the model.The potential independent predictors were age, department (internal medicine or surgery), total number of elderly people with dementia treated by the PCP, whether the PCP has participated in a dementia workshop for PCPs hosted by MHLW, the academy, or a pharmaceutical company, whether the PCP refers to MHLWʼs ʻBPSD treatment guidelines for PCPsʼ and whether the PCP is aware of the MHLWʼs warning of the increased mortality risk of prescribing antipsychotics for elderly people with dementia.So that multicollinearity did not occur, before performing binominal logistic regression analysis, we performed correlation analyses between dependent variable and each covariate, and confirmed that there was not extremely significant correlation.The results are presented as odds ratios (OR) with 95% confidence intervals (95% CI).The statistical analyses were performed with SPSS for Windows, version 23.0.For all the analyses, a p-value of < 0.05 was regarded as statistically significant.

Results
There were 509 respondents and we analysed them.Their characteristics of PCPs are shown in Table-2.Of these, 35 (7%) were women.The mean age was 53.1 ± 8.9 years, and 348 of the PCPs (68%) were engaged in the clinical department of internal medicine.The mean number of elderly people with dementia treated by each PCP was 72.5 ± 100.The binominal logistic regression analysis for prescribing antipsychotics revealed a significant negative correlation between workshops hosted by an academy (OR = 0.596, 95% CI 0.38-0.93,p= 0.022) and referring to the MHLWʼs BPSD guideline for PCPs (OR = 0.491, 95% CI 0.32-0.75,p = 0.001)(Table -3).The binominal logistic regression analysis for PCPsʼ awareness about the increased mortality risk of using antipsychotics revealed significant positive correlations with being engaged by the internal medicine department (OR = 1.811, 95% CI 1.17-2.81,p = 0.008), referring to MHLWʼ s online homepage (OR = 2.204, 95% CI 1.41-3.44,p = 0.0005) and referring to BPSD guidelines for PCPs (OR = 2.149, 95%CI 1.41-3.27,p = 0.0004) (Table-4).

Discussion
This study aimed to establish an overview of the practice followed by Japanese PCPsʼ in the prescription of antipsychotics for BPSD for elderly people with dementia, through the use of an online survey.The results showed that from one-third to half of the PCPs treated excitatory BPSD and approximately half of them prescribed antipsychotics for this.These results were similar to the questionary survey for the GP in Australia 17) .From half to two-thirds had taken opportunities to learn about BPSD in one way or another (through workshops, guidelines, or reference to the MHLW homepage).However, despite the Governmentʼ s promotion of their strategy about dementia, more than onequarter of the PCPs still prescribed antipsychotics for wandering (142, 27.9%).Needless to say, when elderly people with dementia are taking antipsychotics, physicians and other health care providers need to pay attention to the risk of the appearance of extrapyramidal symptoms 18) 19) .For wandering, caregivers should guess the motive why elderly with dementia perform that action and a nonpharmacologic therapy, such as a coordinating environment, should be given priority.It is undesirable to use antipsychotics for the purpose of sedation to suppress wandering; there is a risk of side effects as well 20) 21) .It seems that the MHLW needs to provide more learning opportunities for PCPs to allow a more appropriate use of antipsychotics.
In England, over half of the care home residents are in residential homes with no nursing staff and general practitioners (who play a similar role to Japanese PCPs) prescribe and are responsible for monitoring medication in care homes 22) .In 2009, the UK Department of Health reported that the use of antipsychotics for dementia was unacceptably high and recommended a target of a two-thirds reduction over a period of 3 years 23) .The same year, similar recommendations were incorporated as guidelines in the National Dementia Strategy (NDS) for England launched in February 2009 24) .However, Szczepura et al. 25) later reported that NDS was not associated with a reduction in prescribing antipsychotics in England or in changes in the types of antipsychotic prescribed in care homes.The authors supposed that the reason why reductions in the prescribing of antipsychotics driven by the NDS were not sustained in care homes was that the length of treatment became excessive because of the lack of regular review by general practitioners.Our statistical analysis indicated that one of the factors that resulted in PCPs becoming more cautious about prescribing antipsychotics for BPSD was referring to MHLWʼs guidelines.This may be explained by the guidelines describing the risk associated with using antipsychotics; PCPs referred to this and became more careful about using antipsychotics.Another statistically significant factor was age; however, this was considered not to have clinical usefulness because the odds ratio was extremely close to 1.
The binominal logistic regression analysis suggested that the factors associated with PCPs being aware of the MHLWʼs warning about the increased mortality risk with using antipsychotics for elderly people with dementia were being engaged by the internal medicine department, referring to MHLWʼs online homepage and referring to the BPSD guideline for PCPs.The reason why referring to the BPSD guideline raised the awareness of the increased risk of mortality for PCPs was because the guideline had mentioned this risk.Similarly, referring to MHLWʼ s online homepage of MHLW gave access to a link to the guideline, which PCP could then refer to.
Several limitations of the present study should be acknowledged.First, the definition of the PCP may have been indistinct.Results of characteristics of PCPs were based on the respondentsʼ self-assessment, so we were unable to perform enough inspection about years of working as PCP.Whether respondents actually treated elderly with dementia in each local area was unclear.The final sample may be not fully representative of the population of PCPs in Japan.Second, we asked PCPs in the questionnaire whether they had participated in workshops hosted by an academy or pharmaceutical company, but we did not identify which organisations these hosts were.We should have excluded PCPs who had participated in dementia workshops that did not provide information about antipsychotic medication.Finally, the definition of special terms regarding BPSD was unclear.For example, it may be difficult for PCPs to correctly distinguish violence from agitation.Regarding the term, ʻwanderingʼ, some PCPs might understand that ʻwanderingʼ contains ʻroamingʼ that means moving repeatedly in the same space following their pattern, which is often done by patients with frontotemporal dementia.Following this understanding, it might be understandable to prescribe antidepressants.Also, we should have explained especially about agitation, violence and wandering for PCP in detail in the questionnaire.
In conclusion, despite its limitations, this exploratory study provided an overview of the practice of PCPs in prescribing antipsychotics for BPSD and of their behaviour and their opportunities to learn about BPSD.The results suggested that PCPs may come to use antipsychotics more appropriately by referring MHLWʼs guidelines that included specific mention about the risks of using antipsychotics for dementia.We hope this study will contribute to an improvement in the quality of primary care for dementia and that it may also result in a reduction in the economic burden.
5. BPSDs targeted by PCPs for treatment are shown in Figure-1.The most common BPSD was sleep disorder (357; 70.1%).From one-third to half of the PCPs treated excitatory BPSD delusion, 191

Table - 2
Characteristics of primary care physicians

Table - 3
Total number of elderly people with dementia treated by PCP.＊ p < 0.05 Binominal logistic regression analysis of factors potentially associated with prescribing antipsychotics †OR; odds ratio, CI; confidence interval, MHLW; Japanʼs Ministry of Health; Labour and Welfare; BPSD; behavioural and psychological symptoms of dementia, PCP; primary care physician.† Total number of elderly people with dementia treated by PCP.＊ p < 0.05

Table - 4
Binominal logistic regression analysis of factors associated with the awareness of risks of antipsychotic use From Questionnaire No.9 (No opportunity to conscious, N = 178; Conscious, N = 331)