IOTIZED business in India .The question is how

   IOTIZED Sensor Based Framework for Inclusive Development in Health Insurance

Ankita Gupta*1, Gurmeet Kaur#2, Lakhwinder Kaur*3

*Computer Engineering Department, #Electronics Engineering Department, Punjabi University Patiala, *[email protected], #[email protected], *[email protected]

Abstract:

Present scenario of health insurance sector in India depicts that the cost of acquiring medical health has been scaled up enormously. Certain segments of the society remain in the state of affliction, whereas rest remains healthy. The root cause analysis of above problem in maximum cases is not just the unaffordable health services but a financial system that keep many segments of society out of the main economy. Recently, the Indian government has been pushing for digital operations and the corporate houses responded the same by building interoperable capabilities in technology infrastructure and APIs that plug into the unified payments systems. But, the bottleneck remains how to include such people not having any sort of legal documents or inclination to join the mainstream.  This paper proposes a solution to use IOT sensors and mobile technologies to build the financial trust of the unprivileged sections of society.  In this data mining methods are used to determine the inclusiveness factor score from the mobile and health sensor data. The expected outcome is increased adoption rate of digital services along with health insurances services of people who do not have credit history or trustworthiness in terms of material or collateral grantees.  All these factors collaboratively would lead to the inclusive development and aid the Indian economy in mounting the better position with respect to inclusive development index inspite of the fact that India has large sized informal economy. 

Keywords : Inclusive Technology , Health Insurance Inclusion , Financial Inclusion

Introduction:  

The size of informal economy is huge in India. Therefore, the opportunities in India are double for expanding one’s business in India .The question is how to bring the informal economy in the main fold. The informal economy not only include people that do not have accessible banking services, but also people who are differently disabled physically or mentally and have no or little access to medical services. People who live on fringes of the civil society or are being marginalized to due current economic order. It is also about people who live in their own self-sustaining shells and do not need intervention from the main stream civil society or economy. The governments need to design, implement and oversee the measures to remove barrier to their growth. It would act as a promoter for inclusive health, banking, social order and help in reducing the economic and technological gaps between these groups.

            Technology, innovation and social justice can help us achieve inclusive growth. But, for this new model of growth, revenue and social orders need to be adopted. These models can be assimilated with the most basic iotized “smarter technology” such as mobile phones which is already a part of our daily lives. Mobile technology infrastructure is already spread across India and adoption rate is one of highest for poor also. The need of the hour is to increase the participation of such population for inclusive growth and increase the innovation in development of inclusive technologies. This would directly or indirectly impact the use of services used by main stream economies such health insurance. Consequently, would generate interest of financial institutions in supporting capital and resources to inclusive advancements. Hence, the next section, describes one real life scenario that needs attention and technological intrusion for inclusive growth using method called “Scenario analysis”. It is a   process of analyzing possible future events based on some hypothesis by considering alternative possible outcomes (sometimes called “alternative worlds”). Thus, scenario analysis, which is one of the main forms of projection, does not try to show one exact picture of the future. Instead, it presents several alternatives for future developments. Consequently, a scope of possible future outcomes can be observed along with the development paths leading to those outcomes. In contrast to prognoses, the scenario analysis is not based on extrapolation of the past or the extension of past trends. It does not rely on historical data and does not expect past observations to remain valid in the future. Instead, it tries to consider possible developments and turning points, which may only be connected to the past. In short, several scenarios are fleshed out in a scenario analysis to show possible future outcomes. Each scenario normally combines optimistic, pessimistic, and more and less probable developments. However, all aspects of scenarios should be plausible. Although highly discussed, experience has shown that around three scenarios are most appropriate for further discussion and selection. More number of scenario’s leads to risks of making the analysis overly complicated.

The outcome of such analysis would help us to foresee the possible way by which the inclusive growth may happen in context of inclusive banking, inclusive finance policies, inclusive health care, health insurance and overall inclusive development.

Challenges and Gaps:

The current technological disruptions are creating renewed factors for acceleration of next industrial revolution. The issues, however, still remains how to organize to potential employment and other distribution effects.

1)     There are no single policy mix or ideal technology matrix that can be considered for inclusive growth.

2)     Limited work has been reported where the financial inclusion technology is integrated with the health insurance policies and technologies.

3)     The level of digitization does not reflect the true inclusiveness of the all sections of the society. Most the digital platforms do not accept the people with no credit history or government compliances such as Aadhar card.

4)     Open source technologies for inclusiveness of all sections of the society need more attention as well as encouragement from both public and private sector.

5)     The full advantage of penetration of mobiles phones has not been taken for fighting poverty or for increasing financial inclusiveness. There are few references where this technology has been used to aid people in financial distress, but no reference of using such technology for inclusive health care insurance.

Proposed Technological Model:

 A new industrial progress is underway. The world economies are in the era of rapid transformation that can not only service the elite but all the people who are under-privileged. The Figure gives how the communication can happen using different communication media, protocols and sensors.  

Communications Model: The first step will consist of setting up of infrastructure for communication and observing the subjects . It will consist of mix of multiple technologies as illustrated in Fig . This concept is similar to the mobile computing paradigm, where useful medical data is accessed from the user’s smartphone or wearable medical sensor. No data is stored on the smartphone nor any changes are allowed locally during accumulation of sensor data stream, but the data is accessible through cloud storage interface. This is done with the help of the integrity protocol such as Block Chain.

Data Model:  This approach needs two level of data modeling: first data model belongs to the financial inclusion aspect and the second will be for the health insurance inclusiveness.  The data pertaining to first model will be based on the GPRS locations trajectories (collected using mobile operator services) of the subjects registered with implementation agency (NGO, Social entrepreneur or Government) is collected. The data will consist of the location points and the distance travelled by the subject. Then it will be later on sampled at interval Table 1 of 10 minutes. Typically, the data set will consist of four fields: subject id, Date/ Time, longitude, latitude. The second level will of health or medical sensor data will be including data on temperature, heart rate, physical walking distance and sleep data. For insurance demographic data such as age, gender, ethnicity and behaviour traits such as smoking etc will be also be considered.

                              Table 1:  Example GPRS data showing position of a Subject. 

IoT Sensors:

 The main key sensor for financial inclusion is the mobile based GPRS sensor data. Other than this, it is assumed the subject will either voluntarily or in an organized manner will be get registered with some non-government    agency or the government running this program. The registration process will consist of registering the mobile phone number with the mobile operator with whom the non- government agency or the government has legal understanding. Other than this the person will also need wear certain medical wearable devices given below based on which the health care policy can be designated with the subject.

Sensor No

Sensor Type  

Main Use in Insurance Industry

1

Step Counter (Pedometer) or Waking Distance Counter or Calorie Counter against physical exercise sensor .

More physical activity like yoga, running, jogging, jumping etc., reflects on the person health status. An active person must be rewarded by wellness insurance system. This would help to measure physical activity and other factors like muscle mass etc. The computations of muscle mass, lean mass become important when a person has opted for supervised weight loss program initiated by insurer.

2

Sleep/ Rest  Patterns Analyzers

Its main use is in formulation of wellness insurance package, as parameter to determine the cost of insurance. Data Patterns or Sequences showing erratic sleep or rest patterns reflects that the subject is unable to have normal life due to either stress or some other factors. The reward systems can also validate the sleep patterns with help of ECG signal patterns correlation. The subject’s health is less prone to go astray in case the subject sleep right on time and gets up at right with full deep sleep. 

3

Blood Pressure Analyzers

Blood Pressure is a vital statistics of the body for estimating the health status of the subject. Today, in this stressful life, if a person is able to maintain normal range of blood pressure pattern. He or she must be rewarded as it poses less risk for the company and individual.

4

Sugar/Glucometer Analyzers

Maintaining right range of HDL (high-density lipoprotein), LDL (low-density lipoprotein), TL (Total Cholesterol) and Triglycerides and fatty acids is critical for person to remain healthy. Hence, insurance companies use this data to compute the risk.

5

Body temperature Sensor

Temperature becomes important factor for observing and computing risks especially in case when insurance policy holders are old, fragile or pregnant. 

6

Medication Adherence

The logic of Wellness Insurance is based on assumption that the person does not want to become sick and if he or she is sick he/she will try to come out of the health problems. But if a person does not follow up medical treatment and misses to adhere to the medicines, he or she need not to be rewarded.

7

Pulse and Oxygen Sensor

This would help to find the oxygen levels in our blood and help insurance companies to find risk associated with conditions like asthma.

8

Body Composition Analyzers

It is a crucial parameter to determine the risk involved in case the person is overweight. The overweight person is more prone to health problems. Hence, only that person should be rewarded in wellness insurance plan who maintain right level of weight.

9

Heart Rate/ Beat Analysis

This sensor reading gives information on the state of heart healthiness. It is one that the ailment for which insurance coverage is also denied sometimes. Using Sensor data, the insurance companies may identify the degree of risk associated and compute the premium accordingly.

Algorithms:  The objective of this research work is to model financial inclusion and to construct the health insurance inclusiveness model for informal economy persons. Hence, the algorithm can be divided into two steps. Qualifying the first step is necessary for the subject to get qualified for health insurance inclusiveness.

a)     Financial: This step consists conducting a Geospatial Analysis of the GPRS Data and then visualization of the routes taken by the subject. The person itself is considered as a “sensor” the physical location of the person can be captured in three ways for this achieving the aforementioned goal. First from the mobile sensor that takes GPRS data and second from the wearable medical device. In both the cases a frequency analysis can be done for spotting the route the subject follows and the spot where he or she spends maximum time such as Home or temporary selling space. The call log analysis of the subject along with his or her most frequently called mobile number can also help us find the stability and soundness of the socio-economic circle of the person.  Therefore, all this data is going to be primarily a space vector model that may be further transformed into a network data model. The mobility data and call log data study of the subject can be modeled for building Trust worthiness of the subject that determines his or her qualification to receive benefits from the incumbent agency .     

Trust Matrix Qualifying = Call Qualifying Score1, GPRS Qualifying Score3

Trust Matrix Subject = Call Log Data Score, GPRS Log Data Score;

            The Trust matrix can be defined as matrix that gives three kinds of score, that are computed on the basis of frequency analysis of the call log data and geospatial data. The frequency analysis of all these datasets is done using apriori algorithm.  It proceeds by identifying the frequent individual items in the database and extending them to larger and larger item sets as long as those item sets appear sufficiently often in the database. Apriori uses a “bottom up” approach, where frequent subsets are extended to one item at a time (a step known as candidate generation), and groups of candidates are tested against the data. The algorithm terminates when no further successful extensions are found. By using this way we are able to find how does the subject frequent at particular spots and who all are normally with him in call most of the time.   As an illustration , we can check the example below :

Subject

 

 

 

982856489

982856500

98285600

 

982856489

982856500

982856700

  A

982856489

982856500

982856600

  B

982856489

982856500

982856700

  C

Table Illustration for Frequency analysis of call data

For this call log data table, it is clear that most of the  calls are  initiated by the subject 982856489, and in 50% calls are directed to the number 982856500. Then, 50% of the calls have 98285600 and 982856700 in common.  This set are found as candidate for grouping by the apriori algorithm. And it shows that subject communicates most frequency with these people. And if this process continues for over six months or so. It can be inferred that the subject has stable socio-economic relationship social network score with people and GPRS coordinates will tell about the Physical presence stability GPRS Coordinates Data Score.This way the actual score of the person can be compared with Qualifying score matrix to arrive at a decision of including the person in main stream financial industry on and later on for the giving health care befits through the instrument of insurance .

b)     Health Insurance:  Most of the health insurance companies during their risk estimation process take into account two basic metrices “Age” and “gender” parameter along with that behaviour traits will also be considered. But, in context to solution of our problem of inclusiveness of the subjects. The sensor data need to be considered as well.  It also should be noted that not all kinds of sensor data can be used for this purpose. A simple non-intrusive   medical sensor would be appropriate for this purpose. The sensor that can check the weight, body temperature, heart-rate, walking steps will be useful.  The question is how to model the “Qualification” of the person for health insurance benefits from the incumbent. The table gives an illustration how the sensor data can be linked for health insurance using Empirical Rules method as illustration, if the person qualifies for the inclusiveness for Health Insurance.

Empirical Rule: This rule applies to generally a factor ‘X’ having normal or bell-shaped distribution with mean mu “µ” and standard deviation denoted by sigma “?”. Assuming that the time series data of count of walking steps   of the subject normally behaves like normal distribution. Typically, an adult ‘s ability to walk certain distance per day remains in particular tight normal range, but eating disorders and other factors like sedentary life style may impact this healthy habit. He or she may become weak or may become underweight with the passage of time due to some malnutrition problems. Hence, there is a need for point or credit system that can help the incumbent agency to provide benefits to the subject. According to the empirical rule Ref the observations (in our case “Physical steps taken in day of the person over a time period) fall within the second and third standard deviation of the mean Figure

 

Figure   Deviation Limits  as per Empirical Rule

 

An average person has a stride length of approximately 2.1 to 2.5 feet. That means that it takes over 2,000 steps to walk one mile; and 10,000 steps would be almost 5 miles. A sedentary person may only average 1,000 to 3,000 steps a day.  Hence, for example If the values of the “physical steps taken in a day” remain in this second or third zone ((µ – ?) and (µ + ?)), there is not much to worry as the normal ability of an adult considering the age and gender also. But, if the values with in the first Deviation, there is some serious problem with the subject as he or she seems to be not active physically. Hence, some measures are needed  to be taken in terms of his or her health.   

Expected Outcomes: The outcome of such research can be easily judged from observing the adoption rate of such people, corporates and governments. Usage of such technology would have double impact as it would make the financial inclusiveness easy and at the time would help people suffering from poverty, low income, inequality, disability, climate change issues and disasters. It would aid to leverage the health risk. Such technologies would provide acceleration to social entrepreneurship as many products and services can be designed with similar technology in micro finance once the trust and credit scoring system is in place for such people. The work is important especially from a health insurance perspective for poor, by using this technology the insurance companies can have more accurate picture of the exposures, and risks of the person being insured even if he is economically weak but is trust worthy.

Conclusion: This work can be understood by bifurcating the concept of inclusive development enabling technology.  The First part is about Financial Inclusion using mobile technology and second part is about giving affordable health care through the health insurance inclusion.  The proposed model is based on the trust and credibility. All thought the concept is intangible but they can be quantified by using statistical and data mining methods.  The proposed algorithm in the first place identifies the frequency between the most visited locations using GPRS co-ordinates and using the same concept computes the social network score based on the frequency of the calls to inner circle and outer circle of the person which need to be given the privilege or facility. These computations are only possible due to advent on GPS technology. And in this case, it is tracking the position of the person for the good reason. i.e financial inclusion in the main stream economy. The location histories of a person is basically in the form of spatio-temporal data and such real-world location histories imply, to some extent, mobile users’ home or office location.  Thus both by using the spatial as well as temporal aspects to group the GPS data we were able to construct a trust matrix for qualifying the financial inclusion. This is done by computing the stay points and then grouping them temporally and determining the location visited most number of times within a time period. Which reflects the stability of a person. Hence, is trust worthy and if the person is talking to same social or business circle since long that reflect the stability of his or her business or social ties.

The second part, is all about providing affordable health care using insurance  as a vehicle of risk and affordability . In this case also, we can see by simply wearing few medical sensor devices that would communicate his daily vital health statistics would simply enable the person to quality for affordable health care. This would even offer a chance for the health insurance syndicates to design products that would be highly customized, justified in costing as per the health risk of the individual and group.  

Future Direction:

This paper gives a technological solution and illustration on how to a group of people who do not belong to the main stream economics can be included into the main fold of the banking and have affordable health care. The implication of affordable health does not only apply to the risk coverage but can also help in maintaining the wellness of the person as he or she will be able to monitor his or her health constantly. But, the main bottle neck remains is the investment from the government and the private sector. The new age entrepreneurs need to look into the feasibility aspect of such concepts and fine tune their social entrepreneurship skills to add value for every stake holder. Hence, for future directions, It is suggested integration with mobile wallets, payment gateways and crypto currency based enterprise must be done, only then it would be become an enabling technology for the poor, underprivileged or non-included sections of the society.