Could Your Voice be the Key to A Diabetes Diagnosis?

A diabetes diagnosis usually is determined through a blood test of which there are several. The Glycated Hemoglobin (A1C )is the leading test used for diabetes diagnosis. It measures the average amount of glucose in your blood over the past two or three months.   A score less than 5.7% is considered normal. For the diabetes diagnosis criteria, anything between 5.7% and 6.4% is considered prediabetes. An A1C of equal or greater than 6.5% is diagnosed as diabetes.

A Type 2 diabetes diagnosis also may be determined by the Fasting Blood Glucose (FBG) test. It checks blood glucose levels after fasting the night before. Diabetes is diagnosed at a fasting blood glucose of equal to or greater than 126 mg/dl (milligram/deciliter) on two separate tests.[1]

The third most common diabetes diagnosis test is the Oral Glucose Tolerance Test (OGTT). With this test, you fast overnight. Then, the fasting blood sugar level is measured. Then you drink a sugary liquid, and blood sugar levels are tested regularly for the next two hours. The test is designed to indicate how someone processes sugar. A reading of greater than 200 mg/dL after two hours means you have diabetes.[2]

But now a new study[3] suggest that a Type 2 diabetes diagnosis may be determined by analyzing the sound of your voice using artificial intelligence (AI).

Short voice clips are clues for diabetes diagnosis

The study conducted by Klick Labs and published in May Clinic Proceedings: Digital Health, combined voice technology with artificial intelligence.  The researchers used six to ten seconds of people’s voice, along with health data, such as age, height, sex and weight, to create an AI model to determine a Type 2 diabetes diagnosis.  The model has 89 percent accuracy for women and 86 percent for men.

The research involved 267 people, who were diagnosed as either non or Type 2 diabetic. Six times a day for two weeks, participants recorded a phrase into their smart phones. From more than 18,000 recordings, scientists analyzed 14 acoustic features for differences between non-diabetic and Type 2 diabetic individuals.

The Klick Labs team analyzed several vocal features, like changes in pitch and intensity that cannot be perceived by the human ear.  Using signal processing, scientists were able to detect changes in the voice caused by Type 2 diabetes. Researchers found vocal changes manifested in different ways for males and females

Klick said its non-intrusive and accessible diabetes diagnosis approach offers the potential to screen vast numbers of people and help identify the large percentage of undiagnosed people with Type 2 diabetes.

Diabetes effects on voice

There have been other studies on the effect of diabetes on voice. A study conducted in 2012 found that patients with Type 2 diabetes with poor glycemic control and with neuropathy had a significant difference in the grade (degree of hoarseness) of their voice compared to the control group.[4]

A systematic review in 2019 found that 12.5 percent of people (or 1 in 8) have voice problems that exceed those of the general population. The studies pointed to a higher prevalence (12.5%) of voice problems among individuals with diabetes mellitus as compared to the general population, and higher gastro-esophageal reflux disorder (GERD) related symptoms. The other voice related changes reported across the studies include the presence of laryngeal involvement, hoarseness, and increased strained voice, pointing to the presence of voice changes among those with diabetes. [5]

The research conducted by Klick Labs and others points to the considerable potential of the use of voice technology in Type 2 diabetes diagnosis and potentially other health issues.

Hashtags:

#diabetes #diabetesdiagnosis #diabetestype

[1] “Diabetes – testing,” Mayo Clinic, Accessed August 31, 2024. https://www.mayoclinic.org/diseases-conditions/diabetes/diagnosis-treatment/drc-20371451

[2] Ibid

[3] Press release: AI and 10 seconds of voice can screen for diabetes, new study in Mayo Clinic Journal, reveals,” Klick News, Ocober 18, 2023. https://www.klick.com/news/ai-and-10-seconds-of-voice-can-screen-for-diabetes-new-study-in-mayo-clinic

[4] Hamdan AL, Jabbour J, Nassar J, Dahouk I, Azar ST. Vocal characteristics in patients with type 2 diabetes mellitus. Eur Arch Otorhinolaryngol. 2012 May;269(5):1489-95. doi: 10.1007/s00405-012-1933-7. PMID: 22302159. https://pubmed.ncbi.nlm.nih.gov/22302159/

[5] Ravi, R. and Gunjawate, D.R. (2019), Effect of diabetes mellitus on voice: a systematic review. Pract Diab, 36: 177-180. https://doi.org/10.1002/pdi.2240