Background Stroke is the foremost medical condition responsible for acquired disability and dependency. The initial psychological and physical deficits should arguably be identified early to allow interventions to be put in place if potential long-term sequelae are to be minimised. Physical impairments within stroke cohorts have been extensively researched and reported. We have access to numerous scales that describe general physical functioning during daily activities and its effects on independence and quality of life. Deficits in movement are easily identified and attributes of physical movements can be associated with such obvious measurement terms as strength, speed, co-ordination and task completion. As these attributes can be graded, it is simple to compare patients over time, within stroke cohorts and against the general population: controlled studies are straightforward, even if natural biological variation demands large samples. However, though physical impacts on a patient may be easily observed and measured, corresponding deficits in cognition and mood are less easily detected and quantified. Psychological problems are common within stroke populations, and exert both short and long term effects throughout all stages of rehabilitation. Despite our awareness of the potentially critical effects of psychological factors on patient outcome, there is a dearth of high quality research in this area. Although many cognitive and mood assessments are available, including some that were developed specifically for use in stroke, these are neither regularly administered nor have been convincingly shown to be accurate and reliable in identifying specific deficits. Thus, it is understandable that research available to us that describes prevalence and effects of cognitive and mood problems post-stroke is sparse in comparison to our knowledge of physical deficits. Methodology In order to begin to understand the effects that cognition and mood have on patient outcome post-stroke, a way of identifying prominent issues is required. Knowing that their effects can have both short and long term impact, reliable screening beginning early after stroke onset could offer opportunities to improve patient outcome through early implementation of interventions. Before an appropriate screening tool can be selected, evidence is required to support its feasibility and accuracy within acute stroke cohorts. In this thesis, I investigate cognitive and mood-screening assessments in stroke, across a series of linked projects. I carried out a review of the current research and I surveyed stroke units to identify which cognitive and mood assessments are commonly implemented. I collated and offer synthesis of the published data on accuracy of cognitive assessment instruments. I used these results to inform a diagnostic test accuracy study, examining selected measures that are commonly used in UK practice to screen for cognitive and mood problems. Based on these results, I designed (and was awarded grant funding for) a clinical study to assess test properties of cognitive and mood screening instruments in a rehabilitation setting and to describe potential obstacles affecting patient assessment. Findings There is heterogeneity in the choice of cognitive and mood tests employed across research and clinical practice. There was some overlap in assessment choice within these domains but no clear consensus on a preferred assessment tool. This is in part explained by the substantial number of tests available, it is telling that the most popular assessments accounted for only a fraction of the tool assessments employed. My literature based work also points to a relative lack of published science employing a cognitive or mood assessment tool. My review of diagnostic test accuracy found that properties of cognitive tools commonly used in practice and research (Folstein’s Mini Mental State Examination: MMSE, the Montreal Cognitive Assessment: MoCA, the Addenbrookes’ Cognitive Examination Revised: ACE-R and the Cambridge Cognitive Examination revised: R-CAMCOG) were susceptible to changing populations and purpose of assessment, with test properties differing when screening tools are used in acute and chronic stage of stroke. Depending on the cut-offs that are used to define “screen positive” cases, these tools would have varying ability to identify multi=domain cognitive impairment or dementia. Generally when applying standard (i.e. the traditional cut-off described for test use in an unselected population) cut-offs, sensitivity was good but specificity was low. Specificity could be improved when the cut-offs were altered while maintaining reasonable sensitivity and this suggests that screen positive thresholds may need to be altered to suit a stroke population The need for lowering our standard cut-offs suggests that there may be factors present in typical acute stroke patients which affect assessment accuracy compared to the populations and purpose for which these scales have been developed. Using the MOCA in the acute setting of my clinical study, confirmed that stroke cohorts require altered cut-offs to improve accuracy in cognitive impairment detection. A stroke cognitive assessment that can be derived from a standard neurological examination *the Cog4) has been described. Cross sectional comparison of MoCA and Cog 4 suggest that Cog4 has questionable validity and stroke specific cognitive measures are required since scores derived from other types of measures are not necessarily testing the most appropriate domains for stroke deficits. A lack of published data on cognitive and mood screening in the first days post-stroke suggested that describing the feasibility of assessing stroke patients in an acute setting would be a useful topic for research. My subsequent clinical study incorporated verbal and non-verbal assessments for mood and the MOCA. As well as usual test accuracy outcomes I considered feasibility issues such as proportion of patients suitable for initial approach, acceptance of assessment, prevalence of common stroke related impairments that mandate assistance or cause difficulty in completing assessments, or that preclude assessment altogether. A moderate proportion of patients who were approached declined to take part and several others required external assistance to complete the assessments. Shorter, less cognitively demanding assessments required less assistance and appeared to offer higher accuracy for predicting mood problems at follow-up. These results suggest that delaying cognitive and mood assessments until later during the post-stroke period may reduce the interference from acute stroke deficits. The final piece of work generated from my PhD studies, and that is ongoing, continues the theme of feasibility of cognitive and mood assessments. Cognitive and mood assessments are performed in stroke rehabilitation centres. The rehabilitation setting was chosen, as it will include varying patterns of physical and cognitive impairment. By comparing brief assessments and more lengthy measures of cognition, I hope to identify the most appropriate testing scheme that minimises patient burden. As part of this work I will describe the impact of stroke deficits on assessment and quantification of the patient’s psychological capabilities. Conclusions In conclusion, these studies have demonstrated a lack of guidance and of protocols for cognitive and mood assessment post-stroke. The evident heterogeneity in choices of assessment in research and usual practice indicated a need for evidence based accuracy studies. In conducting these I found that usual measures are susceptible to the population, timing, and cut-off used to define test positive cases, together indicating undesirable sources of variation. Transient stroke-related problems may lead to overestimation of persistent impairments. Although acute screening of cognition and mood would be possible, such screening may not be widely acceptable to patients and would require a high level of assistance from health professionals. Acute screening should only be performed if there are potential benefits that could impact on the patient from identification of cognitive or mood problems at this early stage. With the transient changes in cognition and mood that the majority of stroke survivors experience, screening is best left until later in the patient journey. However, there may still be potential feasibility issues of administration and assessment completion during later stages. Therefore, I suggest that studies that investigate what assessments are feasibly administered to stroke patients in later stages are required. This will inform future trial recruitment for complete data requirements as well as provide clearer picture of stroke survivors’ affected cognitive domains and or mood problems.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:681871 |
Date | January 2016 |
Creators | Lees, Rosalind A. |
Publisher | University of Glasgow |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://theses.gla.ac.uk/7128/ |
Page generated in 0.0017 seconds