Exploring the measurement properties of the Montreal Cognitive Assessment in a population of people with cancer.

BACKGROUND

Cancer and cancer-related treatments are associated with a constellation of physical and psychological changes. Treatments associated with noncentral nervous system neoplasms can have short- and long-term effects on cognition, affecting quality of life in people with cancer. Clinical measurement tools specific to cancer-related mild cognitive impairment (MCI) are lacking. The Montreal Cognitive Assessment (MoCA) has been validated in a geriatric population and used in studies assessing MCI in persons with cancer, but no studies have yet shown its psychometric properties when used with this population.

PURPOSE

The purpose of this study is to explore the psychometric properties of the MoCA within a population of persons with noncentral nervous system cancer.

METHODS

A total of 74 participants were included from persons attending a Cancer Nutrition-Rehabilitation Program at the McGill University Health Centre. Rasch analyses were conducted.

RESULTS

The MoCA data fit all the properties of the Rasch model with a person separation index of 1.04 and person reliability of 0.52. The MoCA items were found to measure a unidimensional construct and spanned 6.57 logits, with item difficulty levels between 2.49 and -4.08 logits. However, the MoCA presented a lack of items of higher difficulty, as person cognitive ability levels ranged from -0.51 to 5.17 logits.

CONCLUSION

Within the limits of a small sample size, the results of this exploratory study suggest the possibility that the MoCA, when used within a population of persons with cancer, may meet criteria for unidimensionality and adequate item fit but may present weaknesses when used with participants of higher cognitive abilities.