Parkinson's Disease: Risk Assessment of Persistent Delirium and Its Management

Background and Rationale Parkinson's disease is the second most common neurodegenerative condition. Delirium is an acute neuropsychiatric syndrome characterised by an altered level of consciousness, confusion and impaired attention. Usually, delirium is transient if the aetiological factors are corrected or if its effects are minimised by medical treatment. Delirium is potentially preventable in 30-40% of cases, but delirium can result in increased mortality and an increased risk of dementia.

Persistent delirium is when delirium continues to affect patients over weeks to months. However, the exact duration for delirium to be defined as being persistent remains controversial. Previous studies have reported the duration of persistent delirium to be between 7 days to >4 weeks. Persistent delirium in this study will be defined as delirium lasting ≥14 days, which is the average reported by the previous studies. Delirium in older people at the point of hospital discharge is associated with worse outcomes regarding function, cognition, mortality and are more likely to need nursing home placement.

People with Parkinson's could be five times more likely to develop delirium and often delirium in Parkinson's is under-recognised in secondary care.

The use of risk stratification of delirium in people with Parkinson's could highlight the most vulnerable patients so that preventive care can be implemented to minimise their risk of adverse outcomes and complications.

Currently, management strategies for delirium in Parkinson's disease are predominantly based on expert opinion and there is a need for research to determine what are the optimal management strategies for this patient group. Moreover, there are concerns that current community support services do not adequately address the psychosocial issues in older people with advanced Parkinson's disease, which may provide an even greater challenge for patients recovering from persistent delirium.

In summary, older people with Parkinson's are particularly susceptible to delirium, but there is also a paucity of research in this area, which needs to be urgently addressed, to not only prevent the adverse outcomes at an individual level but also the detrimental socio-economic impacts on the wider community.

Participant Recruitment

Eligible potential participants will be invited to the study by their usual medical team during an outpatient Parkinson's clinic at a rural, district general hospital. Eligible potential participants with their permission will be contacted by researchers. Carers of people with Parkinson's will be invited to a qualitative interview. Participants will be recruited consecutively in the order that consent forms were completed till the target sample size has been achieved for the quantitative part of the study. For the qualitative part of the study, recruitment will cease once data saturation has been reached.

Methods

There are 3 main parts to this study:

1. Development and testing of an assessment tool for persistent delirium in people with Parkinson's: Participants will be recruited from outpatient clinics for people with Parkinson's this may take up to 8 months. Initially, the potential participant's usual doctor will invite them to take part in the study, before the research team goes through the details of the study with the potential participant. Written informed consent will be needed. If the potential participant cannot consent for themselves, then a consultee will be contacted. A consultee is someone that is well known to the potential participant and is willing to advise on whether the potential participant, should take part in the study. In tandem with participant recruitment for at least 14 months, there will be data collection from medical records. Medical records will be reviewed to identify the characteristics and events that have affected people who then go on to develop persistent delirium. Participants' characteristics that will be analysed will be broadly related to their general level of physical fitness, the severity of their Parkinson's disease, visual or hearing difficulties and any conditions affecting their mind, such as dementia. Events that the medical notes will be checked for include: the medical care, blood tests and imaging that the participant received, any complications during their hospital stay and the support that they received in the community. Participants will be divided into 3 groups: those with persistent delirium, which is where participants become unwell, which has led to the participant becoming confused or have changes in their behaviour for 14 days or more. Another group with non-persistent delirium, this is when the delirium has lasted less than 14 days and a control group who have not had delirium at all. The Chief Investigator will write a summary of each participant but identifiable details will be excluded. These summaries will be reviewed by at least one other medical specialist involved in the care of older people, to double-check that the study findings are as accurate as possible. Potentially, another medical colleague may be required if there is any initial disagreement between specialists. To further confirm the accuracy of the findings, healthcare staff entries in the notes and information collected from relatives/carers can also be used. 8 months will be needed to develop and test the assessment tool. This will be done by using statistical methods, such as logistic regression to determine which of these factors are most strongly linked to persistent delirium. A scoring system can be developed, which allocates an increasing number of points to participants with a risk factor that is more strongly linked to persistent delirium. The accuracy of the tool will be determined by its ability to detect people with delirium (sensitivity) and its ability to detect people without delirium (specificity). It is estimated that 207 participants will be required. This is assuming that persistent delirium is present in approximately 20% of patients. The estimate of persistent delirium is based on local audit data.
2. Development of a management plan From the information collected in 1), there will be a review of the care that a participant received to identify factors that helped and hindered their recovery from delirium. This information will allow the Chief Investigator to begin to develop a care plan, which can then be tested in future studies. The majority of the data needed will be collected during the development of the assessment tool. The writing up of the study will begin 1 year after the beginning of the study and conclusions will be drawn from these data. Data Analysis of Quantitative Data The extracted data is expected to be non-parametric. Descriptive statistics will describe participants' demographics and clinical characteristics, using mean, median, standard deviation, interquartile range and percentages. The Kruskal-Wallis test will be used to establish if there is a statistically significant difference between two or more groups (e.g. those without delirium, those with non-persistent delirium and those with persistent delirium). Post hoc tests will elucidate which groups differ from each other. Chi-squared tests will compare between-group distributions of proportions. Repeated measures ANOVA or Friedman's will determine changes in repeated measure assessments (e.g. full blood count, urea and electrolytes, c-reactive protein). Hierarchical logistic regression analysis, adjusting for covariates (e.g. demographics, clinical factors, infection), will ascertain the significant risk factors for delirium and provide odds ratios. Multiple comparisons will be corrected using Bonferroni's correction. Survival analysis will be utilised to determine the time to the outcomes of interest, such as persistent delirium, mortality and increase in community support. Survival curves will be compared using the log-rank test. It is expected that participants' medical records will have missing data. Missing data will be recorded to determine if there are any missing value patterns, associated with certain outcomes. To develop the assessment tool, the "regularised" regression techniques, such as LASSO or Elastic Net will be used. A scoring system will be devised which allocates an increasing number of points to risk factors that have a stronger association for predicting persistent delirium. The sensitivity and specificity of the risk assessment tool will be calculated and these characteristics will be displayed as ROC (Receiver Operator Characteristic) curves. AUC (Area Under the ROC Curve) can summarise the overall accuracy of the test and this can be computed by the trapezoidal rule. Adjusted odds ratios calculated using logistic regression will compare this study's secondary objective: associations between the management of persistent delirium and its outcomes. These findings will be assessed for their suitability for a larger study to be conducted by the Chief Investigator and these data can inform researchers about the sample size requirements for future studies.
3. Interviews

Written informed consent will be sought from people with Parkinson's and their carer when they attend their routine hospital follow-up appointment. A single researcher will then meet with the participant for a one-to-one interview, lasting for about 1 hour. The interview will be conducted at either the participant's own home or another pre-arranged venue. Questions will focus on the participant's experiences of hospital, the care that they received when they returned home and the challenges that they faced in the community. It is expected that up to 20 participants in total are needed for this study, consisting of up to 10 people with Parkinson's and up to 10 carers. The interviews will begin 2 years into the study and participants will be re-consented to ensure that they still agree to the interview.

Data Analysis of Qualitative Data

The qualitative study will be analysed using thematic analysis. Phrases or expressions from the transcripts will be coded to identify interesting features. Open coding will first be used, which allows the early formation of themes and concepts. These early themes will be reviewed by comparing them to other transcripts to refine and develop recurrent themes. From this, the themes will be defined and a thematic coding frame will be produced which will include sub-themes. This process will be data-driven coding and these data will undergo open, axial and selective coding. The constant comparative method will be used to analyse data to develop a grounded theory, to explain some of the participants' behaviours and reactions to persistent delirium. To achieve this, there will simultaneous analysing and coding of data; each new set of data will be compared to previous extracts, new codes and themes will be added as necessary whilst comparing participants' responses, so that the themes are refined. Integration of these findings will begin the development of an explanatory model. Findings will be reported as recommended by COREQ (consolidated criteria for reporting qualitative research) guidance. The Chief Investigator will be the main data coder, with assistance from Co-Investigators.