Developing a Novel Digital Tool for Personalised Antipsychotic Prescribing in People Living With Dementia: The Views of Australian Clinicians

Study Design

This qualitative study used semi-structured interviews and followed the Standards for Reporting Qualitative Research (SRQR) (Equator Network, 2023). The study was approved by the University of Sydney Human Research Ethics Committee (2024/HE000171). All participants provided informed consent before the interviews, confirming their full understanding of the study’s purpose, the confidentiality protocols in place, and their right to withdraw at any point. Assent was also obtained at the beginning of the interviews.

Participant Recruitment

Participants were purposively sampled across a variety of health professions who are directly involved in antipsychotic prescribing in dementia, including geriatricians, psychiatrists, neurologists, general practitioners (GPs), and nurse practitioners (NPs), to gain insights from a range of prescribing clinicians and healthcare settings. NPs in Australia are authorised to prescribe medicines; however, each state and territory has their own legislation which governs NP prescribing (Victoria State Government, 2019). Geographic variability and clinicians’ experience were also considered in the recruitment process. Recruitment of participants was facilitated through flyers, researcher contacts, research networks (Sydney Health Partners Geriatric Medicine Clinical Academic Group, Sydney Dementia Network) and key informants in the area. Furthermore, snowball sampling was used to gather more participants. Invitations were sent to potential participants by ET via email to confirm interest and consent to participate. For their time, participants were also provided a gift card.

Data Collection

Prior to interviews, participants received a study information sheet, an online consent form, and a brief online survey about demographic characteristics and clinical experience. Consenting clinicians were then interviewed using a semi-structured interview guide (Supplemental table 1). Interview questions were developed from an initial literature review (Tan et al., 2024) and discussions with the research team to elicit perspectives on the following areas: a clinical scenario to ‘set the scene’, factors associated with antipsychotic prescribing, outcomes, patient values and end use.

The interview guide was piloted with one clinician interviewee to ensure questions were appropriately developed to suit the intent of the interview. Online interviews were conducted by ET, an experienced qualitative researcher and pharmacy practice academic, and TT, a qualitative research-trained Pharmacy Honours student, via video call (Zoom Workplace), lasting approximately 30 min. This process was supervised by an experienced qualitative researcher, YJ, who also provided training and support. Interviews were conducted from August to October 2024 and were audio-recorded and transcribed verbatim using Microsoft Word auto transcribing function. Transcripts were verified by TT and imported into QSR Nvivo 14 (Lumivero, 2024) for data management and analysis. Interviews were conducted until data saturation was reached.

Data Analysis

Transcripts were analysed and initially coded for, by using a set of a priori categories structured according to the interview guide by TT. This was followed by thematic analysis using an inductive approach, following the steps outlined by Braun and Clarke (2006), and discussion between TT, ET and YJ. Following initial coding, thematic analysis and its interpretation was used to refine emerging themes by TT, under the supervision of ET and YJ. Any discrepancies were resolved by discussion. Analysis of all transcripts were then finalised in discussion with the wider group of investigators.

Participant Characteristics

Fourteen participants were individually interviewed, with six being males. Participants were clinicians that were involved in antipsychotic prescribing in people living with dementia, with the majority being either psychiatrists or geriatricians. Participants worked across a variety of work settings, mainly in hospital and/or residential aged care, with participants most commonly working within Northern Sydney. Participant characteristics are summarised in Table 1.

OPEN IN VIEWER

Table 1.

Participant Characteristics

Characteristics	Number (%)
Gender
Male	6 (43)
Age group
30–39	4 (29)
40–49	3 (21)
50–59	4 (29)
60+	3 (21)
Specialty
Psychiatrist	4 (29)
Geriatrician	4 (29)
General practitioner	3 (21)
Nurse practitioner	2 (14)
Neurologist	1 (7)
Years of experience
1–9	3 (21)
10–19	5 (36)
20–29	3 (21)
30–39	2 (14)
40+	1 (7)
Work setting
Hospital	9 (64)
Public clinic	3 (21)
Private clinic	2 (14)
Residential aged care	4 (29)
Community outreach	2 (14)
General practice	3 (21)
Location of primary practice
Northern Sydney	6 (43)
Inner West	2 (14)
South Western Sydney	2 (14)
Central coast	2 (14)
Western Sydney	1 (7)
South Eastern Sydney	1 (7)

Themes were organised according to the topic areas about the development and use of the tool: factors for consideration, expected outcomes following antipsychotic use, benefits of using the tool, concerns with using the tool, potential utility, and suggestions for development. Illuminating or example quotes within key themes are presented below.

Factors for Consideration

Medical history. All clinicians recognised the importance of considering the patient’s current medical history prior to prescribing an antipsychotic. They particularly considered comorbidities and concomitant medications due to potential risks of adverse outcomes associated with antipsychotic use. Clinicians also described the importance of frailty when considering prescribing an antipsychotic. Furthermore, clinicians would consider treatment history prior to considering an antipsychotic, and the person’s response.

Dementia stage or symptom severity. Clinicians highlighted the importance of considering the stage of dementia or severity of symptoms when determining the need to prescribe an antipsychotic.

Patient-family considerations. Caregiver burden was a common factor that clinicians considered prior to prescribing an antipsychotic, either to help the family manage the patient or to prevent institutionalisation and allow the patient to remain at home. Clinicians also expressed the need for consent either by the patient or the carer before antipsychotic use. This would involve discussions of the risks and benefits associated with antipsychotic use as well as patient and/or carer treatment preferences.

Ultimately [it] comes down to a discussion with the family and the patient about risks and benefits, and weighing that up to see how much they’re able to manage the paranoia and non-pharmacological strategies. (Geriatrician 3)

Expected Outcomes following Antipsychotic Use

Adverse outcomes. When prescribing an antipsychotic, clinicians indicated numerous adverse outcomes that they would like to know the risk of when prescribing an antipsychotic to their patients. These could include risk of falls, extrapyramidal side effects, and mortality.

Things from the extreme of an increase in mortality to parkinsonism, extrapyramidal symptoms, falls, difficulties with speech and swallowing, metabolic effects from some antipsychotics. (Neurologist 1)

Treatment response. Clinicians highlighted the importance of reducing behaviours of concern and distressing symptoms that the patient is experiencing.

The effectiveness of the medication would be the primary outcome that I would be interested in. (GP 2)

Outcomes important to the patient and/or their carer. Clinicians identified that reducing behavioural changes would be an outcome important to the patient and/or their carer. This would allow the patient to continue to live at home independently where possible, ultimately alleviating the burden on the caregiver.

A greater focus on maintaining independence, preserving their autonomy and not being too sedated…that’s important for the patient so they feel they can function. (Psychiatrist 3)

Benefits of Using the Tool

Identifying risk. Clinicians expressed that using the tool could help with patient assessment by identifying potential risks prior to prescribing an antipsychotic. By more explicitly considering risk factors, participants suggested that prescribers might be more cautious prior to prescribing.

It does prompt clinicians to think about the risks of using antipsychotics in this group…if you put in the inputs and you go, OK, there’s quite a high risk if we are going to do it, it might prompt you [to] be much more cautious around the medications. (Psychiatrist 1)

Furthermore, some clinicians also communicated that the tool could be used to demonstrate response and risk over time. This could aid in developing a treatment plan based upon the initial output, whilst also assisting in deprescribing if the benefit of using the antipsychotic is minimal or if the risk of adverse events increases.

You could probably use it to see how risk changes with time. So, for example, depending on what the percentages are, as people become more frail, you would expect that that risk would increase with time. And so, it could actually be used as a tool to also deprescribe down the track. (Geriatrician 3)

Informed decision-making. Clinicians communicated that using the tool could help to inform prescribers as well as the patient and/or their carer about antipsychotics. It would also help them to consider more factors than they may usually think about when prescribing an antipsychotic.

Patients, family, as well as the clinician, probably are better informed, especially if it’s not something that they deal with very often. (Geriatrician 3)

Some clinicians identified that the tool could be used as a reference point for prescribing or a potential way to validate their decision to use an antipsychotic.

I think it’s good. It’s like a kind of template, a starting point and something you can refer to and a validation tool for what they’re on. (GP 1)

Most clinicians expressed that the tool could be used to help in the discussion with the patient and/or their carer about using an antipsychotic. This in turn would assist the informed decision-making process, as the tool could also be used in an educational manner.

It’s a tool for education that you can educate the family, that this is what is, you know, advised. (GP 1)

Additionally, some clinicians suggested it could help in the consenting process for using an antipsychotic, and aid in shared decision-making for treatment goals.

It sort of helps, I guess, both the clinician and the carers and if the patient can participate to sort of understand where the risks sort of lie and whether the benefits are worth pursuing. (Psychiatrist 2)

Clinicians articulated that the tool could help guide the discussion of using antipsychotics by providing more accurate information and giving the patient and/or their carers actual numbers associated with certain outcomes. Clinicians expressed that quantifying the risk would better help inform the shared decision-making process, particularly for patients and their carers who are more data driven.

I think what would be useful is if you could give carers a kind of percentage risk of different outcomes. I think that would be really helpful because people often ask what the percentages are. (Neurologist 1)

In addition to this, some clinicians voiced that it could also give them more confidence discussing the risks that are associated with antipsychotic use by providing quantified information.

It would give me more confidence that when I’m discussing that with the family or discussing it with the GP or discussing with somebody else, I can kind of make the case to say ‘yes, of course we know there are risks but they’re pretty small’. (NP 1)

Potential tool for consistent prescribing practices. Some clinicians suggested that the tool could potentially help to standardise antipsychotic treatment, whilst promoting a consistent approach to assisting patients in these clinical situations.

It would help sort of standardise certain things as well because I guess psychiatry is done very interestingly, sometimes some people have a method, some people don’t quite have a method, and they tend to stick to things that they’re sort of used to…I think that would probably help them streamline and have a bit more consistent approach to patients rather than sort of relying on, in a way, it’s basically just gut feeling. (Psychiatrist 2)

Other clinicians viewed this tool to have a form of standardised language that allows clinicians and other healthcare professionals to communicate with each other based upon the output of the tool. This would specifically benefit patients who are consulting multiple healthcare professionals.

It can be a form of communication, a bit like, you know, the Heart Foundation, CVD risk factors [where] you plug in the numbers, and you get this result. And that’s something that everyone uses across the board. (GP 1)

In addition, some clinicians recognised the tool as a means to rationalise their decision to prescribe an antipsychotic to their patients. Similarly, it could justify antipsychotic use to prevent delaying treatment in situations where there is hesitation.

I can see why that clinician made that decision or they can see why I made that decision. They may not agree with it, but at least they see where I came from with my decision-making. (NP 1)

Concerns With Using the Tool

Complexities of antipsychotic prescribing. When asked about potential concerns about using the tool, clinicians raised the fact that they would still be guided by patient assessment despite what the output of the tool may be. The clinical decision of using an antipsychotic would be based on situational needs, and this clinical decision would override information from the tool.

I would say that in most cases where someone has bad dementia and there’s significant BPSD, the goals of care are often to focus on comfort and quality of life, and limiting their distressing symptoms. (Geriatrician 1)

Moreover, clinicians indicated that if the risks were not presented appropriately, then the output from the tool could adversely affect discussing these issues with the patient and/or their carer. This could potentially delay treatment involving antipsychotic use, where there is clear benefit of its use, as patients could refuse treatment.

I wouldn’t want it to particularly highlight risks of side effects or adverse outcomes at the expense of the patient necessarily agreeing to have the prescription. You know, patients should be fully informed, but particularly with paranoia or psychosis, the more severe cases that I see, there’s obviously very real risks of doing nothing as well. (Psychiatrist 3)

All clinicians were also asked whether the tool could factor in patient and/or carer values. Some clinicians believed that the tool could not, due to the complexity arising from a potential lack of capacity to consent to treatment.

It would be very difficult to really incorporate that as part of your tool. I mean, there may be factors around limited insight challenges with consent and capacity and that’s difficult to quantify within a tool for sure. (Psychiatrist 1)

However, some clinicians did express that, despite the tool’s limitations, it could help guide the discussion surrounding patient values.

I would say that this instrument would be a tool for the clinician to kind of guide that discussion. It wouldn’t necessarily be able to do that role in and of itself in terms of adapting or discussing of values. (Psychiatrist 3)

Limitations in existing evidence. The data surrounding the target patient population was a concern to some clinicians, whereby the evidence regarding the use of antipsychotics in people living with dementia may be limited.

We’re not really looking at a choice of antipsychotics. So, any kind of other antipsychotic use is more off label use. There’s probably not a lot of evidence around it. So, I do wonder about the utility of this tool in everyday practice. (Psychiatrist 1)

Clinicians further elaborated that the evidence regarding antipsychotic use in this patient population may be poor, as they would not be included in the research of effectiveness.

I would imagine that in terms of research in this population, it would be somewhat limited in the sense that the really aggressive patients probably wouldn’t be enrolled in clinical trials. (Geriatrician 1)

Similarly, there were concerns about how the tool could overestimate antipsychotic risk as the target patient population may already be at high risk of adverse outcomes prior to antipsychotic use.

We are looking at a fairly old and frail population who are already going to be at high risk of a lot of these outcomes…I would wonder whether a lot of it would be attributed to the antipsychotic, when in reality it was the pneumonia or something else not entirely related…it’s hard to say how much of it is just their baseline risk and how much of it is the antipsychotic. (Geriatrician 1)

There may also be conflicting evidence between tools used in clinical practice, where there may be differences in how they quantify certain outcomes that lead to varying outputs.

There are tools to look at anticholinergic burden. However, the tools don’t agree with each other. Different drugs are rated differently because anticholinergic burden is assessed differently. (Psychiatrist 4)

Potential concerns with technology-based tools. Many clinicians stressed the importance of utilising clinical judgement in situations where antipsychotics may be clinically indicated, as they communicated that prescribers could become over reliant on using a tool such as this.

The potential concerns are really people’s brains being switched off in trying to make the difficult decisions. (Psychiatrist 4)

Although, other clinicians did not recognise this as a concern, as the tool only provides additional information that should be considered in the clinical decision-making process.

Obviously, you do your own assessment and like any other standardised tool, you still make your own decision. (NP 1)

Another issue that clinicians identified was that the tool may be too time consuming, particularly if there were too many inputs needed. This is attributed to the fact that clinicians may be time poor and lack the capacity to use a tool that may be too complex.

I just wonder…in terms of actually getting clinicians to use it. I think clinicians are time poor...I don’t know that they’ll have enough time to go to a website to enter all the details and to get these numbers. (Geriatrician 1)

Potential Utility

How the clinician would use it. Clinicians suggested a variety of ways that the tool could be used in practice. It could be used at the point of care with the patient when discussing the use of the antipsychotic, although it would be time dependent.

I would personally probably think about using it at point of care when you’ve got the patient, the family there to both go through it with them there. But I suppose it comes down to time. (Geriatrician 3)

Others proposed that it would be used prior to suggesting an antipsychotic.

I would probably say if I was going out to see somebody and I thought they may need an antipsychotic and I had their medical information and that, I would probably start using it there to give me a broad idea. (NP 1)

Potential users. Many clinicians indicated that there could be multiple users, particularly younger practitioners and GPs, as specialists would have less need for the tool, being more familiar with handling these clinical situations.

Look, I’m pretty experienced. And I probably won’t…but I think a lot of GPs would use it and a lot of younger practitioners and non-geriatricians would use it. (Geriatrician 2)

Likewise, participants proposed its use for nurses, to assist in the clinical decision-making process for discussions with either the clinician or the family. It could also be used in situations where antipsychotics are prescribed on an as needed basis, reducing unwarranted antipsychotic use.

It would be a useful tool for those nurses to use to then present to the general practitioner and the family. (Geriatrician 2)

Other clinicians advocated the tool’s use for pharmacists who could assist in the safety aspects of using antipsychotics.

Pharmacists can say ‘oh, you know you put in this but according to my calculation’…they can use it as well to help with safety. (GP 1)

Suggestions for Development

For the development of the tool, clinicians would like to see increased accessibility through a cloud-based service or an app that could be used on their mobile phone. Similarly, they advised that the tool should be visually aesthetic, whilst also smooth to use. Additionally, they would like to have its data continuously updated as evidence and guidelines change over time to ensure its validity.

I think you just want to make it relatively frictionless and easy to use, so it’s picked up. I think if it becomes challenging…then people will just stop using it. (Psychiatrist 3)

Most clinicians suggested that prompts to aid in decision-making would be helpful, including the type of antipsychotic used and comparative information in terms of efficacy and safety. This may include an explanation of what the output of the tool means, whilst also having understandable information displays to discuss use of antipsychotics with the patient and/or their carer.

You could think just in terms of data display, how best to display that for the benefit of the patient or the relative so they could kind of better understand it…[to] augment or help the clinician with their usual kind of verbal explanation of the risks and benefits. (Psychiatrist 3)

Clinical Implications

The need to reduce inappropriate prescribing of antipsychotics in people living with dementia is imperative (Lau et al., 2024), with only 10% of antipsychotics prescribed to this population deemed appropriate (Macfarlane & Cunningham, 2021). The Royal Commission into Aged Care Quality and Safety found evidence of overprescribing of psychotropics and lack of efficacy in older adults (Aged Care Quality and Safety Commission, 2022). Thus, the findings of this study can help inform the development of a tool that has potential to reduce inappropriate prescribing of antipsychotics. Additionally, some clinician barriers to optimising prescribing in people living with dementia, including a lack of data availability, difficulty in assessing efficacy and safety, and patient/caregiver concerns (Green et al., 2019), can be addressed. By culminating the evidence on antipsychotic use into one accessible tool, better informed decision-making for clinicians, patients and caregivers can be achieved.

Strengths and Limitations

This study has several strengths and limitations. This study was strengthened by the inclusion of a variety of clinicians that prescribe antipsychotics to people living with dementia, providing insights on the potential benefit and concerns of the tool. Whilst the sample size of this study is small, we ensured data saturation was reached by recruiting two more participants. This study also used a thematic analysis approach which allows researchers to summarise and interpret multiple data sets despite a small sample size, whilst enabling an understanding of thoughts and experiences of an individual (Clarke & Braun, 2017; Kiger & Varpio, 2020), aligning with this study’s aims.

However, there are some limitations. Firstly, participants were predominantly sampled through direct contacts of researchers and snowballing. This may have introduced selection bias towards more knowledgeable and interested clinicians. To minimise selection bias, we purposively sampled clinicians from both metropolitan and regional areas of New South Wales, while also considering clinician experience, with participants ranging from junior to senior clinicians. Moreover, with only one researcher verifying audio recordings of transcriptions produced by technology, there may be potential inaccuracies in the transcripts that were analysed. While this study focused on the views of prescribers, it would be important for future research to also investigate the perspectives of patients, carers and non-prescribers such as pharmacists and nurses.