Experience Alz Test App

Over a two year period we conducted two face-to-face research studies with groups of 65 and 61 participants. In our first research (n=65) we determined the specificity of our unique picture-questions in comparison to the Mini-Cog test* for cognitive impairment detection and chose the best 15 for ALz Test. In our second research (n=61) we compared cognitive impairment detection ability of our 15 picture-questions with the Mini-Cog:

Objectives: 

To develop a digital test (ALz test) designed to detect signs of cognitive impairment as accurately as the Mini-Cog test.

Null Hypothesis: 

There is no statistically significant difference between passing/failing scores on the Mini-Cog and ALz tests.

Design: 

Case control study.

Setting: 

Testing was conducted at the participants' homes, who were assessed face-to-face using the Mini-Cog and ALz tests. Mini-Cog testing involved a three-word recall and clock drawing. ALz test consisted of fifteen multiple choice picture-questions.

Participants: 

Volunteers who wanted to be tested (n=61) for signs of cognitive impairment.

Measurements: 

Rejected a null hypothesis because of the statistically significant difference detected between the Mini-Cog and ALz test scores.

Calculated statistical significance of every ALz test's question related to the correct identification of cognitive impairment in comparison to Mini-Cog.

Calculated the most effective scoring criteria of the ALz test responsible for a greatest precision in detection of cognitive impairment in comparison to the Mini-Cog test and the least amount of false negative results.

Limitations: Research participants (n=61) comprised two unbalanced groups where 36 passed the Mini-Cog test and 25 failed.

Results: Two-tailed T-test rejected null hypothesis (p<0.0001) concluding that there is statistically significant difference between the Mini-Cog pass/fail and ALz test scores. Confidence interval of 95% was detected for the ALz test score at and above 69%. Correct identification of cognitive impairment in comparison to the Mini-Cog was 92% with false positive results occurring in 28% of cases.

Conclusions: 

Digital cognitive testing using the ALz test appears to be as effective as face-to-face assessment with the Mini-Cog evidenced by only 8% error in correct identification of cognitive impairment. Ongoing research is necessary to further confirm these findings and re-evaluate the use of this method in a more extensive face-to-face assessment.

In a nutshell: When we calculated data from the second research (n=61), there were three options to choose from for the pass/fail score (following respectively) in comparison to the Mini-Cog test:

• False+ (11%, 2.8%, 28%)
• False- (28%, 48%, 8%)
• Correct symptom ID (72%, 52%, 92%)
• Correct ID of no symptoms (89%, 97%, 72%)

We chose the third pass/fail score because:

• Correct identification of cognitive impairment in 92% of the cases
• Probability of the false negative result is only 8%

* Mini-Cog is a rapid test that is used around the world by healthcare professionals to detect signs of memory impairment. We are using it in the clinical trial from a generous permission of its creator, Dr Soo Borson.