Can IOT and AI help fundamentally redesign India’s broken mental health system?
by Mihir Shahane
Citizen science research project under the guidance of Dr. Sukant Khurana.
ABSTRACT:
The main focus of World Health Day 2017 was to highlight the need of awareness in common mental disorders like DEPRESSION. It was held under a slogan ‘depression- Let’s talk’. Depression is the most common mental disorder along with anxiety, WHO studies show that ‘one in twenty person in India has gone under Depression at some stage of life’. Almost 322 million people are affected by depression at a given time worldwide and India is a home of almost 57 million people affected by depression (almost 18% of global burden)[1]. This large number contributes to a significant burden at a global level. This leads to a poor quality of life causing a huge social and economic impacts. Depression and suicides are closely related. Lack of awareness, stigma, insufficient supply of drugs and medicines, lack of mental healthcare system, and lack of mental health specialists (0.07 psychiatrists per 100 000 people and 0.12 psychiatric nurses per 100 000 people in India) have pushed the mental healthcare gap to 87.2% to 95.7%[1]. This traditional approach of current healthcare system can be updated or digitised to a smart mental healthcare system. All the phases of a good healthcare system such as detection, monitoring, data collection and analysis, and treatment can be made simple and accurate by using Internet of Things and Artificial Intelligence. Internet of Things for wearable sensors networks and Wireless Communication protocols for data sensing and sending to a server are used to store the data. Machine learning and data science opens opportunities to collect and analyse huge data without any human interference and any emotional approach.
Introduction :
National Mental Health Program (NHMS) and District Mental Health Program (DHMS) studies show that the average mental healthcare Gap in India is 90%. These figures depicts the inability of traditional mental healthcare system. The prevalence of Depression in women is almost two times the number of men in India. Also the working men and women found to have more prevalence than that of non working. The major causes for this include exposure to industrial pollution, environmental toxins, job insecurity, poor quality of sleep, poor nutrition, and substance use etcetera [2]. School and college students are more prone to a severe depression which cause many deaths in early age. Depression and anxiety are most common mental healthcare problems in pregnant women, it affects almost 10 to 15 % of pregnant women in India [3].
Considering the current situation of depression in India, troubles with our mental healthcare system have to be solved. Therefore, there is a huge need of digitization of traditional healthcare system. Modern technologies including Internet Of Things and Machine Learning or AI can play a vital role in detection, monitoring and treatment of common Mental diseases like Depression and Anxiety. Internet of Healthcare Things, Wearable Internet of things can improve the performance and reach of healthcare professional and it can capture and send real time data to the healthcare professional which can detect and analyse real time situation of the patient. Further this huge data can be processed and analysed by machine learning and artificial intelligence which is more capable of processing huge amount of data and giving evidence based decision and not any emotional influence. IoT and AI can overcome the unavailability of mental health workers and improve the capability of current mental health workers. This paper is based on digitization of current healthcare system by integration of IoT an AI with Traditional healthcare system to bridge the healthcare gap.
Literature Review:
Current research shows how various technologies can be implemented to improve the healthcare system. There are various methods for screening for depression such as Quick Inventory for Depression Symptomatology (QIDS), The Rating Scale Hamilton Depression, Montgomery Absurd Depression Inventory (MADRS), Beck Depression Inventory (BDI), The Geriatric Depression Scale (GDS)[4]. Studies employ movement patterns and physiological markers for detection of depression. Physiological markers, like face recognition, cardiac frequency markers like variations in heart rate and skin body temperature are frequently used. Study by Almeida C. Edwing, Ferruzca N. Marco , Guiterrez P. Ivan [4] includes three modules; two modules consisting of image processing with Raspberry Pi and Arch Linux as arm based module for sensing heart rate and skin body temperature, and third module is a web based data server for physician to collect and analyse the real time data. These physical markers can detect the depression in patient with real time monitoring as shown by their studies. The information obtained is processed on a web server through an application of viewing in a web page or mobile device. The system is designed to send alert message via MSN to physician and family in emergency situation such as extreme temperatures, irregular heartbeat, and absence or excess movement[4].
Another project named PSYCHE project [5] concentrates on wearable IoT with combination of biomedical technologies, micro techniques and communication technologies. It proposes micro non invasive biosensors and smartphones as a data collection system. The data collected by biosensors and smartphone is sent to a common server with internet connection. It also introduces PSYCHE garments for both men and women which includes fabric electrodes on patient’s thorax, a special garment with portable electronic device collects heart rate, breathing rate, amplitude and posture, activities etc. A smartphone is used for monitoring sleep and activity hours, mood agenda, detailed clinical assessment of mood state in the form of visual and analogue scales, voice samples for detection of depression and mood disorders. This will ultimately help to analyse mood disorders in patients [5].
One study shows the Internet based self help for detecting depression through social media on guided and unguided with therapists. The guided self help with internet based platform showed significant results in six months of case study. It includes pre treatment evaluation, a six month follow up with patients with emails and videocalls etc and a posttreatment evaluation. These studies suggest that depression and anxiety are detectable with constant follow up with patients for a certain duration of time with internet based guided treatment and monitoring [6].
Artificial intelligence (AI) deep learning protocols offer solution to complex data processing and analysis. Increasingly these solutions are being applied in the healthcare field, most commonly in processing complex medical imaging data used for diagnosis. Current models apply to AI to screening population of patients for markers of disease and report detection accuracy rates exceed those of human data screening. This paper discusses on exploration of an alternative model of AI deployment, that of monitoring and analysing an individual’s level of function over time. For adopting this approach, this paper proposes that AI may provide highly accurate and reliable detection of preclinical disease states of neurodegenerative diseases. AI based monitoring of an individual over time offers a potential for the early detection of change in function for the individual, rather than relying on comparing the performance based of population norms. Finally it explores an approach to developing AI platforms for individual monitoring and preclinical disease detection and examine the potential benefits to the stakeholders in this technological development[7].
Methodology:
The drawbacks of traditional mental healthcare system has created a huge gap between the system and patients. This system has to be redesigned in Integration with Internet of things and Artificial Intelligence. This paper proposes to redesign the whole system with Intelligent Diagnosis, Real time Monitoring, Data capturing and Management, and Smart treatment. This will enable to have a broad approach towards the huge population and will overcome the human limitations.
1)Intelligent Diagnosis:
The very first step in mental healthcare system is diagnosis of disease by physical symptoms and signs tests. Intelligent diagnosis with sensor nodes like heart rate variability, breathing rate, skin sweat sensor, image processing, speech will allow remote areas to diagnose their depression with healthcare specialists. This will allow to bridge the gap via Telemonitoring or Diagnosis in remote areas. Also working people in Banks, offices, factories, call centres, etc can be monitored during their working hours with image processing and speech processing etc. This data is further processed by AI and machine learning platforms which will detect the primary signs in coordination with Healthcare Professional. Mental Healthcare Workers and NGOs can be included to increase the reach of the project. This diagnosis system will increase the capability of these people. These observations will be made regularly and suspected individuals will be further consulted and checked manually. This will enable healthcare specialists to cover many patients in short time accurately. AI will allow the healthcare specialists to analyse the patterns easily with AI assistance for impactful assessment and detection in rural areas. Increased interactions with suspected patients with sensors will make easy to detect the common mental disorders.
2)Real Time Monitoring :
After the detection and diagnosis of common mental disorders based on the intensity of illness the next step is monitoring of patients. The traditional monitoring is based on manual check ups and tests which is not possible with huge number of patients. In real time monitoring wearable Internet of things is used with sensors nodes, actuators and communication devices which will collect real time data of patient’s health and will be able to send the data to a common data server with Bluetooth and internet connections. Wearable sensor nodes will collect real time data like Heart rate variability, Breathing rate, skin conductance, sleep variations, speech recognition and image processing which will be used to determine the patterns of common mental health disorders. For the people having mild depressive disorders will be allowed to wearable IoT sensor system without manual monitoring and people with severe depressive disorder will be assisted in their monitoring by mental healthcare workers along with wearable IoT. Machine learning and data science will be used to analyse the real time data and make them in patterns for unguided and guided monitoring. Any significant change in patient conditions will be captured and then medical assistance will be provided as per requirement.
3) Data collection, analysis and Treatment:
Diagnosis and real time monitoring of patients will generate a huge amount of complex data. This data has to e managed properly with security of data and confidentiality in information of individuals. A private server has to be built which can handle and analyse huge amount of data securely. Use of AI and machine learning will allow handling of huge data without any human interference. AI and Machine learning will study the data patterns as per the given instruction with real time analysis of data patterns. Healthcare professionals and workers will be directly benefited by the use of this technology, it will allow them to manage, handle and react to the huge data with less manpower. This server will sort the data in patterns based on many aspects such as Disease, age group, region , number of people affected without any human emotional interference. This will be beneficial to analyse and create more infrastructure and healthcare facilities based on meta-analysis.
Treatment to mental illness will be directly benefited by the machine learning and real time monitoring using IoT. The actual data available on server will enable to enhance mental healthcare infrastructure and facilities like medicines, drugs, and healthcare workers. Early detection by this task shifting process will enable to build a more comfortable, stigma free and secure environment for treatment of depressive disorders. Doctors and mental health professionals will directly get the biodata of patient from the server and all the data from his\her detection, monitoring of the illness.
Barriers for Implementation :
Will consumer wearable technology ever be adopted or accepted by the medical community? As the line between consumer health wearables and medical devices begins to blur, it is now possible for a single wearable device to monitor a range of medical risk factors. Potentially, these devices could give patients direct access to personal analytics that can contribute to their health, facilitate preventive care, and aid in the management of ongoing illness[8]. However, how this new wearable technology might best serve medicine remains unclear. For instance, it is possible to identify the severity of depressive symptoms based on the number of conversations, amount of physical activity, and sleep duration using a wearable wristband and smartphone app[9]. Sleep apnoea could be quickly diagnosed, and sleep quality improved, with a lightweight wearable that measures heart rate, breathing volume, and snoring (through tissue vibration) instead of a heavy polysomnography. Aspects such as quality of life, acceptability, and cost benefits are infrequently or incompletely reported in telemonitoring trials, and existing reviews of remote monitoring have frequently been criticised for their poor methodology[8].
This new technology raises additional questions concerning the impact on users’ health and wellbeing. Currently, wearables exist within a “grey area” regarding user safety. The potential issue of harm is largely absent from the current literature, but it is conceivable that people may become over-reliant on automated systems that provide a false sense of security or fuel a self-driven misdiagnosis. Patients could also suffer from negative consequences of excessive self-monitoring by finding it uncomfortable, intrusive, and unpleasant[10]. The reliability and validity of wearable devices is also concerning. Devices are marketed under the premise that they will help improve general health and fitness, but the majority of manufactures provide no empirical evidence to support the effectiveness of their products. Finally, for patients and medical practitioners, the privacy and security of personal data generated by consumer wearables remains problematic. Storing of personal data in private server is always a worry for patients. Misreading or misuse of personal data could lead to legal cases.
Conclusion:
India’s National Mental Health Survey and World Health Organization has given some threatening figures about India’s mental health and healthcare services in case of common mental disorders. Almost 90% population has not received any medical help in their lifetime which is the cause of most of the deaths every year. Common mental disorders like Depression and Anxiety is a major cause of suicides in India which is the second largest cause of death. The major reason behind this situation is the negligence of people, stigma, and failure of traditional healthcare system. Traditional approach needs more manpower, can cause uncomfortable situation in society. Unavailability of healthcare professionals and workers in India pulls back the traditional approach of healthcare system. To provide a quality healthcare service for common mental disorder a systematic approach in integration with Internet of Things and Artificial Intelligence. These technologies enable healthcare professionals to collect and analyse huge amount of data patterns with computer based analysis and without any human interference. Transformation of complete traditional healthcare system with smart IoT and AI based smart healthcare which includes smart diagnosis, real time monitoring and smart data collection analysis and treatment practices. A common server based data control system will create one roof for collection and analysis of huge data which in turn enable healthcare professionals to have a clear approach towards patients as provided by machine learning. Common data collection and analysis system will allow healthcare organizations to create a better healthcare environment including better infrastructure, Number of specialists and medical aids. Considering the advantages of modern healthcare system, it is also needed to look at the drawbacks of the system which include patient’s comfort in using wearable IoT devices, data security etc. Therefore, a complete modern healthcare system will be implementable when there is a good balance between use of technology and patient’s comfort.
References:
1) World Health Organization 2017, Depression in India, Let’s Talk.
2) National Mental Health Survey (NHMS), Depression in India and Healthcare Gap.
3) National Mental Health Survey (NHMS), Depression in women meta analysis.
4) Detection of episodes of major depression in older adults through physiological markers and movement patterns, Almeida C. Edwing, Ferruzca N. Marco , Guiterrez P. Ivan.
5) Telemonitoring with respect to mood disorders and Information and Communication Technologies: overview and Presentation of the PSYCHE Project. Harve Javelot, Anne Spedazzi, Luisa Weiner, Sonia Garcia, Claudio Gentili, Markus Kosel, and gilles Bertschy.
6) Internet-based treatment of depression: A randomized controlled trial comparing guided with unguided self-help, Thomas Berger , Gerhard Anderson.
7) Deep Machine Learning Application to the Detection of Preclinical Neurodegenerative Diseases of Aging, Mathew J. Summers, Tamas Madl, Alessandro E. Vercelli, Georg Aumayr, Doris M. Bleier, Ludovico Ciferri.
8) The Rise of Consumer Health Wearables: Promises and Barriers ;Lukasz Piwek, David A. Ellis, Sally Andrews, Adam Joinson.
9) Bravata DM, Smith-Spangler C, Sundaram V, Gienger AL, Lin N, Lewis R, et al. Using pedometers to increase physical activity and improve health: a systematic review. JAMA. 2007; 298(19):2296–304. PMID: 18029834
10) Krantz DS, Baum A, Wideman Mv. Assessment of Preferences for self-treatment and information in health care. Journal of personality and social psychology.
