IGS University of Twente
Research Research services PhD & Education Press & Public Information
Thursday 22 June 2017, 14:30, Prof.dr. G. Berkhoff - Zaal

Phd Defence Sarah Janus

antipsychotic drug prescription in nursing home residents with dementia - perspectives of staff and proxies in a complete decision process 

Dementia is characterized by progressive deterioration in cognitive ability and capacity for independent living. Neuropsychiatric symptoms, like agitation/aggression, apathy and irritability, are known to contribute to the high disease burden. Often these symptoms lead to an admission in a nursing home. Additionally, these symptoms are highly prevalent and persistent and can be a serious burden to the patient, family and staff. In the pharmacotherapy of these symptoms a big gap exists between the available scientific evidence on its limited effectiveness and the reality of its frequent use in daily clinical practice. While several studies conclude a limited success of psychotropic drug treatment, especially with regards to antipsychotic drugs, and a high risk of adverse effects, these drugs are still widely used in patients with dementia. There is no indication that this practice is changing. This thesis aimed to reduce this gap and extend the knowledge on the reasons for prescribing antipsychotic drugs.


In the first part of this thesis we identified the magnitude of this gap. For long the treatment of neuropsychiatric symptoms with psychotropic drugs has been discussed in the scientific literature and treatment guidelines. Therefore a decrease in psychotropic drug use was expected. In Chapter 2 we looked at the psychotropic drug use over time (2003-2012) in the Netherlands. In this secondary data-analysis of seven studies we indeed found a small but significant decrease in the prescription of antipsychotic drugs 0.99 (95% CI 0.99-1.00) as well as an increase of anti-dementia drugs 1.03 (95% CI 1.02-1.03). These changes remain significant when we adjust for relevant demographic and clinical variables. No change was found for psychotropic, antipsychotic, anxiolytic and hypnotic drug use. The decrease of antipsychotic drug use is only minimal but very much needed.

In order to put our findings into perspective we also looked at the psychotropic drug use over time in Western European countries (Chapter 3). We looked at 37 studies on antipsychotic drug use and 27 studies on antidepressant drug use conducted in 12 different European countries. The antipsychotic use in nursing homes ranged from 12% to 59% and antidepressant use from 19% to 68%. The highest rates of antipsychotic drug prescription were found in Austria, Ireland, and Belgium while for antidepressants in Belgium, Sweden, and France. Due to heterogeneity between the studies as well as the use of antipsychotic and antidepressant drugs in Western European countries no time trend was found.

Psychotropic drugs and especially antipsychotic drugs are associated with modest benefits for some neuropsychiatric symptoms, such as agitation and psychosis. However, safety warnings regarding antipsychotics have been issued a long time ago because of the adverse events experienced by this specific patient population. For example, the use of psychotropic drugs is associated with an increased risk of falling. To create tailored fall prevention programmes, information on the magnitude of the association between fall incidents and the use of specific psychotropic drugs or drug classes is needed. In Chapter 4 we investigated these possible associations. In our retrospective cohort study we demonstrated that the use of antipsychotics (aHR 1.49; 95% CI 1.12-2.00) and hypnotics & sedatives (aHR 1.51; 95% CI 1.13-2.02) were associated with an increased risk of falling. Additionally we found that falls may be associated with individual drugs rather than drug classes. Within the drug classes clear differences are evident between individual drugs. While haloperidol and quetiapine are seen to have an association with falls, pipamperone and risperidone do not. It might be useful to take the fall risk into account when choosing specific drugs.

In the second part of this thesis we investigated the reasons for prescribing psychotropic drugs. Decision making by health professionals is a complex process and requires a diverse knowledge base with multiple variables and individuals involved. Many factors might influence the approach towards a specific treatment such as the patients’ preferences, family members’ preferences, nursing staff’s preferences, the physicians’ knowledge and beliefs as well as the culture within the nursing home organization. To elucidate the reasons for the high antipsychotic prescription rates among institutionalized patients with dementia we used the Theory of Planned Behavior (TPB). The experienced barriers of elderly care physicians towards the discontinuation of antipsychotic medication, as can be seen by the prolonged drug prescription periods, for the behavioral symptoms of dementia were investigated in a study in Chapter 5. Interestingly, half of the physicians responding to the questionnaire, agreed that antipsychotics have positive consequences for the patient, such as calming effects. Physicians who indicated that they discontinue antipsychotics less frequently believed more often that antipsychotics are associated with positive consequences for the staff. Physicians who tend to discontinue antipsychotics had a higher perceived behavioural control than the group who indicated having a low intention. This study shows that the relative positive attitude of physicians regarding the expected effects of antipsychotic treatment for the nursing staff and the expectation that known side effects will not occur, require confirmation and deeper insight.

In Chapter 6 we looked at reasons for nurses to ask physicians to prescribe antipsychotics to elderly nursing home patients with dementia. In this study, nurses and nursing assistants also frequently agreed on items related to the positive effects of antipsychotics for the resident and for the staff. Nurses and nursing assistants with a lower job satisfaction were more likely to call for antipsychotics. Having more positive beliefs about treatment effects and feeling more in control toward asking for antipsychotics were positively associated with the intention to call. All variables explained 59% of the variance of intention. The current position (nurse/nursing assistant) was associated with actual behavior to call. In order to decrease the nurses’ requests for prescribing antipsychotics, one needs to address the nurses’ and nursing assistants’ belief in positive effects of antipsychotics for the resident, which is not in line with available evidence. Nurses and nursing assistants need to get more education about the limited effectiveness of antipsychotics.

The third section of this thesis focused on the preferences of physicians prescribing treatments for neuropsychiatric symptoms. By measuring benefit and value, preference elicitation methods are used to increase patient-centered medical decision making. In Chapter 7 we looked at studies that used a preference elicitation method, representing the patient’s view for different health care contexts. The three decision contexts have different requirements for use and elicitation of preferences. In the review we included 379 preference elicitation methods. These methods were categorized in 1) matching methods [MM] (N=71, 18.7%) i.e. time tradeoff, 2) discrete choice experiments [DCEs] (N=96, 25.3%) i.e. Best Worst Scaling (BWS), 3) multi-criteria decision analysis (N=12, 3.2%) i.e. analytic hierarchy process and 4) other methods (N=200, 52.8%) i.e. rating scales. Most publications of preference elicitation methods had an intended use in clinical decisions (N=134, 40%). Fewer preference studies had an intended use in health technology assessment (HTA, N=68, 20%) or benefit–risk assessment (BRA, N=12, 4%). In clinical decisions, rating, ranking, visual analogue scales and direct choice are used most often. This review indicates that relatively simple preference methods are often useful in clinical decisions because they are easy to administer and have a low cognitive burden. MM and DCE fulfil the requirements of HTA and BRA but are complex for respondents. However, no preference elicitation methods with a low cognitive burden could adequately inform HTA and BRA decisions.

In Chapter 8 we then applied a preference elicitation method to compare the preferences of health care professionals and proxies in making treatment decisions using the Best Worst Scaling (BWS) approach. The respondents were asked to choose between an antipsychotic or non-pharmaceutical regimen for a hypothetical patient with dementia demonstrating neuropsychiatric symptoms and to select the best and worst attribute out of 16 treatment attributes. The non-pharmacological treatment option was chosen by 52% of the proxies and 71% of the physicians and nurses. The respondents who chose antipsychotics rated the factors “fastest result” and “most effective” as important. Physicians ranked “experience with antipsychotics” as an important factor for prescribing antipsychotics. Only the proxies rated the factor “having a low negative effect on the patient” as important. The nurses and elderly care physicians who chose the non-pharmaceutical treatment ranked “appropriateness” and “of little burden to the patient” as important factors. This study shows that while doctors and nurses prefer non-pharmacological interventions, proxies indicated a preference for pharmacological treatment because of the immediate effect. However physicians follow treatment guidelines and nurses and proxies rely on the physician’s recommendations. Physicians therefore need be sensitive to these differences.

In Chapter 9 the results of the different studies are discussed. The methodologies for understanding decision making used in this research are discussed. We used the TBP, which tries to explain and predict the individual’s behavior and the BWS, which values different options (i.e. aspects of treatments) to quantify the most preferred one. Both approaches may make different assumptions about how people make choices. Therefore using the two approaches for understanding the prescription process of antipsychotic drugs may lead to a better approximation of reality. The overlap in results verified the most important factors for these treatment decisions. Both approaches combined could imply that most nursing staff and physicians feel that pharmacological treatment is needed in certain situations. If they consider antipsychotic treatment, their main motive is to decrease problems in the environment (for staff, other residents) and to decrease these problems quickly. Therefore more participation of the proxy, the responsible family member, is needed to defend the interests of the patient who is no longer capable to do so on their own.

We conclude that the three stakeholders, physicians, nurses and proxies, should discuss antipsychotic drug prescriptions in multidisciplinary consultations. This would lead to more awareness by the physicians of the different parties involved in the prescription process, more realistic expectations for the treatment by the nurses and proxies and would lead to more individualized non-pharmacological treatments since proxies have a better understanding of the patient’s identity. This can lead to more appropriate antipsychotic drug prescriptions for these patients.