The Saudi Arabian adaptation of the Addenbrooke’s Cognitive Examination – Revised (Arabic ACE-R)

BACKROUND: The population of the Arab World is about 300 million and the Arabic language is one of the six official languages of the United Nations. As with the rest of the world, degenerative neurological conditions represent a major health problem in regions such as the Middle-East where Arab people are in the majority. However, clinical neuropsychology is still in its infancy in this region. Very few tools for the assessment of cognition have been developed for use with Arabic speakers in the Middle-East region. The Addenbrooke's Cognitive Examination – Revised is a brief cognitive assessment tool that has been well validated in its original English version as well as a number of other languages, but never been adapted for use with Arabic speakers. An important issue for the assessment of cognition in this region is the high level of illiteracy, particularly in older adults, making the development of tools that can be used with both literate and illiterate participants a priority. OBJECTIVES: The studies presented in this thesis involved the translation, adaptation and validation of an Arabic Addenbrookes Cognitive Examination- Revised (ACE-R) and involved data collection from both literate and iliterate participants. METHODS: The ACE-R was translated into Arabic and the process is described in Chapter 2. Critical to the process was the cultural adaption of the test items. Three parallel versions were developed. Data was collected from four participant samples, recruited in Riyadh, Saudi Arabia: (1) Healthy literate (N= 147); (2) Healthy illiterate (N= 283); (3) Literate with a diagnosis of Alzheimer’s disease (AD) or Mild Cognitive Impairment (MCI) (N= 54); (4) Illiterate with a diagnosis of AD or MCI (N= 169). Chapter 3 presents a study of the validity of the Arabic ACE-R in literate participants. Receiver operating curve (ROC) analyses were undertaken to determine the sensitivity and specificity of the Arabic ACE-R to MCI/dementia, as well as positive and negative predictive values. Optimal cut-off scores were determined. Chapter 4 presents a study of the reliability of the Arabic ACE-R with literate participants. Parallel forms of the Arabic ACE-R were administered on two occasions separated by approximately one week. Test-retest and internal reliability (Cronbach’s alpha) were examined. A version of the test was developed for use with non-literate participants and Chapter 5 presents a study of its validity with this population. Chapter 6 reports a study of the reliability of the tool with non-literate participants. Chapter 7 reports normative data for the Arabic ACE-R, identifying fifth percentile cut-off points. RESULTS: Literate participants: Amongst healthy controls Arabic ACE-R data were not normally distributed, hence non-parametric statistics used in analyses. Amongst healthy controls age was correlated with Arabic ACE-R performance (rho = -0.568, p<0.0001) and level of education was also correlated with Arabic ACE-R performance (rho = 0.559, p<0.0001). As there was a significant difference in age between healthy controls and patient groups, groups were matched for age by removal of young controls and participants also examined in three age bands. Groups were matched for level of education. There were significant differences between each of the three groups examined – Mild Cognitive Impairment, Dementia of the Alzheimer’s type and healthy controls. As the MCI group was small, Receiver Operating Curve (ROC) analyses were conducted on the combined MCI/DAT group compared with the healthy control group. Levels of sensitivity/specificity were high. For a cut-off point of 70, sensitivity was 1.000 and specificity was 0.946. The positive and negative predictive values (PPV and NPV) were also high, particular for base rates that are likely to be closer to those found in clinical practice. For literate participants, internal reliability was high (Cronbach’s alpha, 0.932) as was total score test-retest reliability (rho=0.944). Individual subscale reliability ranged from rho=0.685 (Fluency) to rho=0.865 (Memory). Illiterate participants: Amongst healthy controls Arabic ACE-R data was not normally distributed, hence non-parametric statistics were used again. Amongst healthy controls age was correlated with Arabic ACE-R performance (rho = -286, p