DEVELOPMENT AND VALIDATION OF A MOBILE APPLICATION-BASED MONTREAL COGNITIVE ASSESSMENT WITH A NOMOGRAM PREDICTION MODEL FOR EARLY DETECTION OF MILD COGNITIVE IMPAIRMENT IN TYPE 2 DIABETES MELLITUS

Abstract

Background; Cognitive impairment in patients with type 2 diabetes mellitus is far more common than most clinicians recognize, and it is almost never screened systematically in routine diabetes care. The Montreal Cognitive Assessment is the standard brief screening tool for this purpose, but delivering it in a busy outpatient clinic with printed forms, trained personnel, and enough time is not always realistic. A mobile application that digitally delivers the same test could change this; however, it needs to be properly validated before it can be trusted clinically. Separately, identifying which patients with diabetes are heading toward cognitive impairment before it becomes obvious requires more than a single blood test; it needs a prediction tool built from the variables that actually matter, presented in a format that a nurse or general practitioner can use without specialist support. Methods ;  One hundred patients with confirmed type 2 diabetes mellitus were enrolled at Saveetha Medical College and Hospital, Chennai. All participants completed both the paper-based and mobile application versions of MoCA during the same visit. The agreement between the two was measured using intraclass correlation coefficients. Age, sex, HbA1c level, diabetes duration, BMI, waist circumference, fasting glucose level, lipid profile, and treatment type were recorded for all participants. Univariate analysis followed by multivariate logistic regression was used to identify the variables that independently predicted cognitive impairment, which were used to build a nomogram. A p-value below 0.05 was used as the significance cutoff. Results; The app and paper test results were closely correlated (intraclass correlation of 0.94), which is strong enough to justify the app as a primary screening tool rather than just a backup option. Sixty-two out of 100 participants had cognitive impairment. It was more common in patients with diabetic complications (67.3%) than in those without (56.3%), but the difference was not significant (p=0.255). MoCA scores in both groups were below the normal cutoff of 26. After multivariate adjustment, only one variable independently predicted cognitive impairment: the combination of obesity and diabetes, with an odds ratio of 14.7 (p=0.038). Age was the only demographic variable that significantly predicted the prevalence of complications (P =0.027). HbA1c levels and diabetes duration did not differ significantly between the groups. Conclusion ; The mobile application was functional. It produces results that closely match those of the paper test, which is important because it removes most of the logistical barriers that currently prevent cognitive screening in diabetes clinics. The nomogram built from this data gives clinicians something they do not currently have a simple, point-based tool to estimate cognitive risk in a diabetic patient using information that is already available at every routine review. The broader finding is that cognitive impairment in type 2 diabetes is not something that appears only after complications develop; it is already present across the diabetic population, and the tools to find it earlier already exist. This gap is evident in their usage.