Technical Challenges in Providing Remote Health Consultancy Services for the Unreached Community

Transcription

1 Technical Challenges in Providing Remote Health Consultancy Services for the Unreached Community Eiko Kai Department of Electrical Engineering and Computer Science Kyushu University Fukuoka, Japan Ashir Ahmed Department of Advanced Information Technology Kyushu University Fukuoka, Japan Abstract Insufficient healthcare facilities and unavailability of medical experts in rural areas are the two major reasons that kept the people unreached to healthcare services. Recent penetration of mobile phone and the demand to basic healthcare services, remote health consultancy over mobile phone became popular in developing countries. In this paper, we introduce two such representative initiatives from Bangladesh and discuss the technical challenges they face to serve a remote patient. To solve these issues, we have prototyped a box with necessary diagnostic tools, we call it a portable clinic and a software tool, GramHealth for managing the patient information. We carried out experiments in three villages in Bangladesh to observe the usability of the portable clinic and verify the functionality of GramHealth. We display the qualitative analysis of the results obtained from the experiment. GramHealth DB has a unique combination of structured, semi-structured and un-structured data. We are currently looking at these data to see whether these can be treated as BigData and if yes, how to analyze the data and what to expect from these data to make a better clinical decision support. Keywords Portable Clinic, Personal Health Records (PHR), Remote Health Consultancy, BigData, CDSS (Clinical Decision Support System) I. INTRODUCTION There are 1 billion people are unreached in terms of accessing to quality healthcare service [1]. About four thousand children die of diarrhea in a day, one pregnant mother dies in every 90 seconds. This scenario can be dramatically changed if we can simply convey few simple medical tips to the target unreached community. Most of the unreached people are from rural areas in developing countries [2]. Healthcare service does not exist there for two major reasons: (1) Doctors do not want to stay in the village as they do not find their livelihood requirements fulfilled (2) Quality hospitals/clinics do not sustain without stable income. Recently, mobile phone became available in each corner of rural areas. Health consultancy over mobile phone became popular in Bangladesh as an alternative solution. One such service holder receives calls per day for health consultancy [3]. Consultancy over mobile phones brought many benefits to the people especially to the remote female patients who can talk to a male doctor keeping themselves anonymous for private diseases. People can call at any time of the day from anywhere in the country. A doctor can prescribe OTC (Over The Counter) medicine, can interpret clinical records and also can introduce a hospital or doctor near the patient s place. However, in order to diagnose a disease properly, doctors need to see the clinical records measured by diagnostic tools. Kyushu University in Japan and Grameen Communication s GCC (Grameen Communication Center) Project in Bangladesh have prototyped an affordable portable clinic [4] to be deployed in a community and to measure basic health data. The collected data will be made available at the doctor s side before the patient make a call. The portable clinic project started checking health data as a health check up service in the remote areas in Bangladesh. The health records were analyzed locally with a predefined logic and categorized them in four groups: green, yellow, orange, and red. Patients with orange and red have unusual clinical results and are selected to consult with the doctor over video conferencing tool equipped with the portable clinic. In 2011, our project proposed an efficient model to utilize doctors precious time and serve more patients [5]. We checked up 462 people in three villages. In 2012, our project has made a larger step to build a healthcare BigData by using our GramHealth tool and targeted to serve people in two rural areas and in two local organizations. In this work, we introduce the current status of two representative remote health consultancy systems, case study results obtained from one of the health consultancy service providers in Section II. We also describe the technical challenges. Section III describes the anatomy of our portable clinic and introduce a mechanism which can serve patients in a mass scale in rural community. GramHealth tool is also explained in this section which collects and archives personal health records. Section 4 summarizes our work with future directions. II. ADVANTAGES OF MOBILE-PHONE BASED REMOTE HEALTH CONSULTANCY SYSTEM AND TECHNICAL CHALLENGES This study focused on mobile-phone based remote health consultancy services.

2 A. Remote Health Consyltancy over mobile phone: How it works? In a typical mobile phone based healthcare consultancy system, the doctor is located in an urban area in a call center or in his/her house. The doctor has a facility to receive phone calls, a computer based hospital database to support the patient. At the patient side, there is a mobile phone. The patient places a call to a hot-line number of a call center [Figure. 1]. The call is usually routed to a doctor in a roundrobin fashion. There are exceptions when a female patient prefers to consult with a female doctor or a follow-up patient looks for the previously consulted doctor. The consultancy has the following three major phases- (a) Introduction phase: the doctor introduces him/herself, and then asks for patient basic information (name, age, sex, location etc.). (b) Diagnosis phase: the patient explains the symptom and then the doctor interrogates the patient based on the symptoms to find out the cause of the symptom. (c) Advice phase: the doctor then either prescribes medicine (over the counter medicine only because of the medical policy issue), or suggests a nearby hospital for further checkup and consultancy. An advanced healthcare service provider keeps the patient-doctor conversation records in a CDR (call details record) and uses special software tool to keep the patient profile details including the list of medicines prescribed. amazing advantage of remote consultancy over mobile phone. The following table [Table I] has the summary results of our observations. Observed Item Results (n=400) (a) Caller Patient: 60%, Relatives: 40% (b) Age distribution of the patient 0-10 years: 29%, years: 15% years:24%, years:17% years: 9%, 50+ years: 7 % (c) Sex Male: 67%, Female: 33% (d) Location Rural: 30%, Urban: 70% (e) Call completion Complete: 68%, Incomplete: 32% (f) Time of call Day (8:00-15:30): 57 % Evening (15:30-23:00): 18% Night (23:00-8:00): 25% (g) Time occupancy of a single call (h) Consultancy about (i) Type of advices Introduction phase: 8%, Diagnosis phase: 27%, Advice phase: 67% Disease related: 79%, Preventive healthcare related: 21% Prescribed medicine: 54%, Advice: 28%, Referred to specialist/hospital: 17%, (j) Patients Follow up: 17%, New: 83% (k) Patients Fully satisfied: 71%, unsatisfied: satisfaction 21%, average: 8% (l) Major diseases Gastro-intestinal: 22%, Respiratory: consulted 17%, Reproductive:10%, skin: 10% CDR TABLE I. SURVEY RESULTS OF A TYPICAL REMOTE HEALTH CONSULTANCY SERVICE In the followings, we describe the major findings- GSM Net Center Figure 1. A Typical Mobole-phone based Remote Healthcare System B. Amazing Facts We have gathered the patient-doctor conversation records archived in December We have found that there were more than 10,000 audio call records. We have clustered the records in 100 groups and randomly selected 400 audio records for our case study. In the following table [Table I], we summarize our observation. There are a good number of female patients making calls (33%) by themselves. This is quite amazing to observe because the female patients are usually attended by her husband or their parents. In many cases, they feel shy to share their private diseases with a male doctor. However, over a mobile phone, the female patients are less hesitant to share their private diseases with a remote doctor. This is an (a) Caller: It is not always the case that a patient makes the call by him/herself. There are cases when someone else places the call on the patient s behalf. Low-literate patients do not feel confident to directly talk with a doctor. Obviously, the kids (<12 years) the older people or very sick patients always depend on a relative to make the call on their behalf (40%). Most of the calls are made by the patients themselves (60%). (d) Location: The service is not only targeted for the rural patients, the urban unreached and also the elite class people also makes calls during emergency period or at unusual time (after midnight) when other hospitals are not open. In our observation, 70% of the calls were made from urban areas. 30% calls from rural areas may seem smaller in compare to the urban areas. Patients are not yet aware of the service. (e) Call Completion Rate: Not all the calls are complete calls i.e. the doctor-patient conversation do not have always a happy ending. There are more than 32% incomplete calls. The potential reasons are: the poor voice quality, test calls made by curious people to check whether the service really work, the patients irritations when the doctor speaks in a very formal way which they are not really accustomed with and they drop the call from their side.

3 (f) Time of call: 57% of the calls were made in the daytime (8:00-15:30). Normally the clinics in the urban areas start their services in the evening hours. Female patients find this time suitable when they have less household works. (g) Time occupancy of a single call: We measured the conversation phases to see the pattern how much time is spent for which purpose in a single conversation. The result shows that 67% of the conversation time is spent to advice a medicine and explain the usage. (h) Consultancy about: 21% people make calls for healthcare issues i.e. to know about the preventive healthcare or seeking for explanation of clinical records. (j) Follow-up Patients: There was no simple way to identify a follow-up patient from the call audio call records. Only one identifier was the caller phone number. We made assumption from the content of the conversation and found that 17% of the total consulting patients were follow up patients who consulted for the same disease case. (k) Patients satisfaction: 71% people are satisfied with the service. 21% of the patients were not satisfied. The reasons assumed due to the voice quality, some people feel irritated when the doctor asks very common or repeating questions-people think that they do it for making more money as the fees increases with the duration of the call. C. Technical Challenges Although our study shows that 71% people are satisfied with the present consultancy service. There is however, a big room for improving the service by introducing simple additional functions into the present system without making any substantial changes in the infrastructure. In this section, we discuss the technical challenges followed by our ideas to address these issues. (1) Maintaining a patient ID: A patient ID is a key element to keep and maintain individual healthcare records. The present system does not offer a unique ID to their patients. A CDR keeps the mobile phone number of the caller, however there are cases when a patient calls from relatives phone or uses a family-owned share phone. Therefore, the phone number cannot be a unique ID. (2) Disease diagnosis process: In the present system, there is no diagnostic tool at the patient side. The doctors are afraid of making inaccurate assumptions from the symptoms expressed by the caller. A physical measurement is necessary to better understand the degree of a symptom and to make a better clinical decision. (3) Patient profile archive: The doctors at the call center are offered and trained to insert the patient profile during the conversation. Many doctors do not feel comfortable to use a computer during the conversation. Also it will take extra time to insert the patient profile which will again irritate the patient. As a result, the patient profile never gets sufficiently stored. Without past records, it is difficult to take care of the folow-up patients. (4) Patient s location: Currently the call center has to ask a series of questions to identify the geographical location of a patient. A doctor cannot accurately refer a patient to a hospital or to doctor if patient s location is not known. (5) Prescription: The medicines in developing countries have English names. The low-literate patients have difficulties to understand the names prescribed by the doctor and take a memo. Some providers use SMS service to send the medicine names. There is a policy that the doctor can only prescribe OTC medicine. Therefore, the doctors have limitations to treat the patients (6) Health Data Portability: Some patients have the past clinical records in hard paper format. It is difficult to read out the clinical data for the remote doctor. Some hospitals keep the past records in digital format. Currently there is no scheme to transfer the digital data from one hospital to another. The same is true for the developed countries too. III. OUR PREVENTIVE HEALTHCARE APPROACH: PORTAVLE CLINIC AND BIG-DATA TO SUPPORT CDSS We came up with a two-phased approach to serve the remote patients- in the first phase we collect villagers personal health records (PHR) as a preventive measure. The PHR for each member is built and made accessible by the remote doctor in this phase. The second phase will be service delivery phase. In this section, we describe our portable clinic concept and explain how to collect and archive PHR of the villagers in an efficient way. A. Portable clinic: A proposal to efficiently serve the remote patients We considered disease diagnosis issue (mentioned in section II(C)) as the primary missing item in the current remote health-consultancy system and proposed an affordable, usable and sustainable concept portable clinic to be added in the current initiative for preventive healthcare. A portable clinic is a device equipped with essential diagnostic tools (for temperature, blood, blood pressure, ECG, urine, etc). The clinic is designed to be affordable (<US$300) and can be carried by village female health assistants. A prototype of the concept has already been developed and is in the field for our experiment. The portable clinic box will be owned and operated by a village health assistant. In an ideal situation, she visits the patients door step for regular and on-demand physical checkup. The physical healthcare data will be stored in the portable clinic and a copy will be sent to the GramHealth health records database. GramHealth is a software tool

4 developed by our department considering the needs of the villagers. The call center doctor can access GramHealth through the Internet or have a copy of the database in their call center server. Upon receiving a call from a patient, the doctor asks patients ID number and finds patient s previous records with personal and family profile. This way, the doctor doesn t need to repeatedly ask questions about the patients personal profile. This way, the doctor time is saved and also the cost burden of the patient will be less. It also provides a good psychological impression to the patient with a feeling that the patient is known to the doctor. A past record contains previous prescribed medicine and the doctor can easily ask the status for the follow-up patients. B. Experimental Environment We carried out three series of health camps in three remote villages (basundia village in south-west, rampal village in the southern part and Vomradoho village in northern part) in Bangladesh by using our prototype version of portable clinic and GramHealth healthcare software tool. In order to save doctor s precious time, we use our developed triage system to consult only with the patients at risk. Our experiment environment consists of the following facilities: (a) a small call center in Dhaka with one female doctor and two male doctors (b) A portable clinic with 9 diagnostic tools. The tools are not in a single box but are dismantled and operated by two assistants in the villages. (c) GramHealth software tool to obtain and maintain the patient health records. GramHealth does not automatically capture data from the diagnostic tools, we manually inserted the health checkup data into our GramHealth database (d) considering the network infrastructure facility in the village, we have developed a off-line version of GramHealth to store the obtained health profile locally and send to the central server when network is available. In this experiment, we aim to (1) confirm the whether he technical challenges mentioned in section II (C) could be efficiently served, (2) monitor whether the system can work with the compromised infrastructure- where unstable bandwidth and regular power-outage is common (3) see whether doctor s precious time can be saved without compromising the healthcare quality. Portable Clinic GramHealth C. Results and Discussion In the following table, [Table II], we discuss the technical challenges outlined earlier, our ideas to solve them and the results obtained from our experiment. Challenges (1) Maintaining a patient ID (2) Disease diagnosis process (3) Patient profile archive Our ideas, test results and comments During our experiment, we generated sequential 8-digits numbers. We considered national ID, passport number, voter ID, telephone number etc. None of them are available for every patient. In future, we have the plan to develop an FHR (Family Health Record). Therefore the ID needs to be compatible when a member of a family moves from one family to another after their marriage or any other social causes. This is an open issue to be researched. We introduced a low cost Portable Clinic to be owned and operated by a village health assistant. We tried with 9 different diagnostic tools, however, we believe that the number and the type of the tools will depend on community. PHR data can be collected and shared with a remote doctor by using this portable clinic box. At the doctor s side in the call center, we assigned an assistant for doc to assist the doctor in recording and digitizing the content of patientdoctor conversation (4) Patient s location GPS can be used. Our portable clinic does not have this function yet. During our experiment, our staff inserted the location for the patient. But this is will not be applicable when the patient calls from a different place. (5) Prescription Only OTC medicines are prescribed due to the medication policy. Government is considering to extend the OTC prescription if the doctor can see the patient through video. (6) Health Data Portability This is listed as a future work for us. Unreached Community Village Health Assistant Center Figure 2. Experimental Environment with Portable Clinic and GramHealth software tool TABLE II. TECHNICAL CHALLENGES AND EXPERIMENTAL RESULTS We setup our portable clinic in a community space in a village. To our surprise there were hundreds of people came to pre-register for our portable clinic based healthcare checkup service. In order to handle the masses, we came up with the concept of triage to categorize the people in following four different classes.

5 Category Indication Action Red High risk Re-check the risky items and Orange Medium risk consult with doctor for symptom analysis and prescriptions Yellow Low risk Provide Health care manual Green Healthy No action TABLE III. PHR AND TRIAGE We served all the people showed up in the camp. In a three day camp in three villages, we served 462 patients and archived their records. Among them, 33(7%) were identified red, 146(32%) orange, 203(44%) yellow and 80(17%) green. It was observed that not only general people came to take the checkup but people with previous health complaints also visited. Almost 40% people needed consultancy with the remote doctor. At this data collection phase (first phase), we have started collecting another patients to be finished by March, The second phase (remote healthcare) has just started. The collected data will be archived as Big-Data to find out location based, age based, family based disease pattern and the effectiveness of treatment. Once the treatment pattern gets mature, GramHealth can be a good source of knowledge base for health education to common people, patients, researchers and family members. IV. CONCLUSION AND FUTURE WORKS In this work, we introduced the technical anatomy of mobile phone based remote consultancy services in developing countries. We carried out a case study to analyze one-month long patient-doctor voice records logged by a healthcare service center. We reported our findings which reflected the demand and the adoption of technology based on the socioeconomic culture of the country. We explained the technical challenges in the highly compromised infrastructure and proposed the affordable and usable "portable clinic" to collect health care data from the patients' door in an efficient way. We also developed a software tool "GramHealth" to collect and store the data for the remote doctor in the call center. The health records are producing a BigData of villagers to be populated by end of March As a future work, we will analyze the collected BigData to turn our Database into a knowledgebase so that the patients, researchers and common people find the system more useful as a source of info-medicine. REFERENCES [1] Global Health Issues accessed on Sep 25, 2012 [2] The Remote and Rural Steering Group, Delivering for Remote and Rural Healthcare, The Scottish government, Edinburg, Nov 30, 2007 [3] Ashir Ahmed and Takuzo Osugi, ICT to change BOP: Case Study: Bangladesh, Shukosha, Fukuoka, pp , Sep 2009 [4] Ashir Ahmed, Koji Ishida, Masaharu Okada, Hiroto Yasuura, Poor- Friendly Technology Initiative in Japan: Grameen Technology Lab, The journal of social business, Volume 1, No 1, January 2011 [5] Shinya Kato, Study on Implementing a Portable Clinic based on Social Needs, Undergraduate Thesis, Kyushu University, March, 2012 [6] Ahsananun Nessa, Moshaddique Al Ameen, Sana Ullah, Kyung Kwak, Applicability of Telemedicine in Bangladesh: Current Status and Future Prospects, The International Arab Journal of Information Technology Vol. 7, pp , 2010 [7] Eiko Kai and Ashir Ahmed, Remote health consultancy service for unreached community: amazing facts and technical challenges, Proceedings of the first MJIIT-JUC Joint Symposium, November 21-23, 2012, MJIIT, UTM, Kulalumpur, Malaysia. Figure 3. Remote Consultation over Skype