 Good afternoon. So I'll be presenting on implementing or the ongoing implementation of an electronic medical record system in Amman Reconstructive Surgery Project. So I give you a quick overview of the Reconstructive Surgery Project, which is providing multidisciplinary care, which includes also a psychosocial and rehabilitation component to victims of ongoing violence and instability in neighboring Middle Eastern countries. This project primarily focuses on orthopedic plastic and maxillofacial specialties. The project is currently housed in a 148-bed facility and generally admits between 50 and 55 patients per month with 200 patients present in Amman at any given moment. Generally, the length of stay in this project is four to five months, and the patients, once they move out of our surgical wards, they go to a rehabilitation center. That's the fourth and fifth floor of the hospital, these two top floors, or they can be housed in an off-site location. So since opening in 2006, MSF has treated more than 4,000 patients from Iraq, Syria, Yemen, and Palestine. So I just described a couple of challenges that have been experienced by the project since opening in 2006. The hospital team has experienced a number of challenges related to complex duplicative and person-dependent data collection processes, and a significant amount of this data was collected and inputted into over 100 Excel line lists and spreadsheets. Much of this data was related to patient movement and service utilization with very little medical data captured, such as procedure and diagnostic information and lab data. So patients were also followed using three different sequential identifiers during their closer treatment in Oman. Additionally, a new sequential number would be assigned if the patient came back for a subsequent phase of treatment, meaning that if a patient comes to Oman a second or third time, they could be followed with six to nine different identifiers, meaning also that we don't have a patient master index with a unique identifier to follow our patients. We had a lack of standard terminology and definitions. The diagnostic codes and procedures that were utilized by the hospital were also person-dependent. So the primary objective of this implementation was to streamline patient information flow and provide real-time feedback to the clinicians. So the solution proposed with BAMNI, which is an open source EMR system that's been developed by a company in India, and the overarching goals of this implementation of BAMNI was to provide a centralized platform, which can automatically generate a unique identifier, capture patient transition points and statuses, display key patient data, and then display key hospital indicators. So prior to the implementation process, there was an extensive hospital needs assessment that was performed in conjunction with ThoughtWorks, software and technology consulting, or a software and technology company, in India and key hospital stakeholders. At the end of the needs assessment, implementation timeline was proposed, identifying key requirements specified by the hospital team. During this time period, additionally, we managed to improve the hospital infrastructure, which included the recruitment of a system admin, purchasing and placement of three dedicated servers to manage the application. One was acting as our primary server, and then we have two that are redundancies. We also expanded the Wi-Fi network access in the entire hospital. Prior to this step, having a good area where we had good coverage was next to impossible. A governor's structure was also implemented, which was for the formation of a project steering committee. The project steering committee was providing overall direction for the implementation process, prioritizing these high-level features, making sure that the hospital team and the infrastructure for the hospital was ready for the rollout of the application. We did an assessment of the patient flow and information flow. This is an ongoing process that's relevant for each of the releases, which I will discuss in two seconds. And it was to identify any inefficient processes or duplicative processes. The biggest portion of the whole entire implementation process was the creation and definition of the medical data and content, which Kathy has touched upon. This is basically the backbone and maybe the ban of my existence at this point was this medical data definition. So the hospital, in conjunction with ThoughtWorks, is defining any and all medical content that will be incorporated into the EMR. The feature development talks about developing features that such as OPD scheduling, OT scheduling, bed management, importing of microlab results into the application. These are slightly easier to manage for the hospital because we simply relay our requirements to ThoughtWorks and then they build something that we need. So this is the proposed timeline. This is rather the adjusted timeline because the proposed timeline was over a six-month period and as you can see, it's more than six months. So release one was completed in November of 2016. I arrived in the project in July of 2016, so it took us from July until November to complete release one. Release one really involved improvement of the infrastructure and placement of these servers. The rest of the medical content was much easier to handle than in the subsequent releases. So release one involved registering our patients into the application, so we received medical data from the medical liaison officers in the patient's home countries and then this information is inputted into the application. The medical liaison officers in the home countries of Iraq, Syria and Yemen do not have access to the application, they simply send us all of the medical content. Afterwards, we upload all of the medical history forms, the videos, the photos, the x-rays and so on that we're receiving from the MLOs in the home country and then this information is ready to be presented to the surgeons and the anesthetists in the validation committee. Once the patient is registered in the application and a unique identifier is automatically generated once the nationality and the sex of the patient is inputted inside. So whoever's registering the patient has to be very careful when they're coding the nationality and the sex as these IDs at a later point they cannot be changed once they are assigned. Release two was completed in April 2017 and included the initial assessments for the nursing, medical, doctor, anesthesia, physio, surgical and mental health. In encoding these initial assessments into the application we did a lot of extensive work with mapping all the diagnoses and procedures to ICD-10, SNOMED and the international classification of health interventions. So we also have the ability to manage the movement of patients in beds throughout different locations inside the hospital. So we can move patients in beds in the surgical ward to the rehabilitation center, to the off-site location, which is a hotel. And then in the end, once all of this information is inputted for release two, we can generate a patient summary that summarizes or collates all of the key medical information that's necessary for the consultation, for the treatment of the patient. Release three is a really, really big release. It includes the rest of the clinical data that is collected by the hospital and we utilize over 100 protocols and forms and we've managed to accomplish six of them as of release two. The clinical data will include ward, OT, OPD follow-up, physio and psychosocial follow-up. Once all of this clinical data is inputted into the application, we'll be able to generate the discharge summary for the patient and discharge meaning to discharge from the hospital rather than discharge from the surgical ward. Additionally, the key features that will be developed in conjunction with ThoughtWorks will be the ability to generate an OT list, OPD appointments, physio and psychosocial outpatient appointments and then to utilize the medication ordering tab. So the impact of, the implementation, sorry. The impact of a release one was really quite minimal. We had maybe six users that were going to utilize the application at this point in time and the biggest change really was introducing this unique structured patient identifier. The users of release one were also using computers and laptops and their day-to-day management of information so they were quite comfortable with navigation and there was very minimal training involved aside from you giving them the test environment and asking them to click buttons. So release two involved, as I mentioned, standardization of terminology, which was pretty time consuming and labor intensive. We had an OT list Excel sheet that contains all of the surgical acts that have been performed on a patient since the beginning of the project in 2006 and we went through each of these acts and mapped them to ICD-10, SNOMED or the International Classification of Health Interventions. We also implemented the bed management tool prior to this implementation. The ward supervisors and the rehabilitation supervisor were managing with daily Excel spreadsheets so each month they would produce 30 different sheets that would in the end summarize their bed occupancy in each of these locations. And these spreadsheets would have to be constantly updated. We also tweaked the patient and information flow in release two and I just say tweaked because we made just a minor, minor change of the way patients are moving during the initial assessments in the OPD. So rather than the patients seeing any of the care providers during the initial assessments, we would start them with the nurse who's responsible for this set of information, then they move to the OPD doctor, then they see the anesthetist and after they see the surgeon. At the end of the day, we have everything nicely summarized in the patient summary. And then this touches on the visibility of information collected by different care providers. Okay, challenges. There were challenges, yes. So there were a lot of challenges that impacted the original timelines, the six month time period and then resulted in a rather extended transition period for the hospital. This involved the development of the medical content, implementation of information patient flow changes and then communication of detailed requirements to a third party. So the biggest, yeah, I talk about lessons about. Okay, so there were challenges. However, the user feedback has been overwhelming positive from the hospital and they're still surprisingly quite excited about the remaining implementation process. So as I said in release one, we had around six users and then in release two, it increased to 52 users and these users included surgeons, anesthetists, medical doctors and nursing, physio and psychosocial team members. So we did a really quick self-administered six item questionnaire to assess user satisfaction around 20 people responded and generally people were quite satisfied with the overall implementation process. So the question of EMR helps me to do my job better. I think would be more relevant once we have completed the implementation or phase three at the end of October. As right now there are certain people who it actually does help them with their job, meaning that they have for like, for example, the surgeons, they have the complete history of the patient available to them in the application where other departments are kind of managing with using both the paper system in parallel with the electronic system. So what are our lessons learned? There are several listed on this slide but for me, the main lesson I learned since the nine months that I've been working on this project is that we need to focus on the patient flow, information exchange on managing people and then their expectations rather than the technology itself. By managing expectations and everybody has come to the table saying, oh EMR is coming, we can do this and this and this and this and then to say slow down, like Kathy said, we have to prioritize what is the most important for this implementation process. So how do we go about doing this? The project steering committee has really been helpful in setting the direction of this project. They're good at managing the expectations and managing this implementation process. We've provided very early and constant commutation to the hospital team about the status of the implementation. We've placed posters inside most of the hallways on the different floors detailing the timelines which I have to print I think every one month, a revised timeline, but it's okay. And then really as I mentioned to focus on engagement and support of the clinical staff. In this hospital, we've been incredibly lucky. The team has been really involved with generating feedback, with being proactive and giving me advice and giving me support and encouragement even through the difficult times that again the focus needs to be for most implementation. It's the content that is going to be a setback in delaying the timelines. So what are our next steps? Onboarding the remaining hospital staff. I think there are about 50 or so more users who need to be exposed to the application. This includes the ward nurses and the ward doctors and then providing continuous training for the end users. The remaining content development, completing development of key features outside OPD and OT and then this knowledge transfer and ownership to the mission. Meaning we have to document how we've gone this far and document the requirements. My time is running short, sorry. So the last three points you guys couldn't read out. Well done. I'm impressed with everyone for managing to condense so much into so little time. And quite a journey and experience of learning that embedding yourself in that project for nine months. I was particularly struck by the data negotiations and we'll put your brains about that later. But clearly a project with considerable socio-technical challenges as well as heading in the right direction. So can I call up the panel while they're coming? Can we take any questions for clarification first of all? One down at the front there. So I'm not sure I got it completely right. Maybe I misunderstood, but I have the impression that when you did the evaluation, are you looking at how easy it is to actually introduce the data in the system or the use of the data already in the system by the user? Sort of the medical doctors, do they actually use the EMR to look back at their patient and to adapt the treatment for example, or to look at allergies or whatever side effects or social democratic stuff? This is one and the second one. How, I mean because Cathy explained in her presentation that it generated automatically analyzes of some sorts, 28 if I remember well, do you have something similar for the EMR in terms of surgical side infections for example or some other stuff? We're in the initial phases, I would say of the implementation process. So for the initial assessments, we are in the physicians, the nurses are directly inputting using the electronic medical record system and then they can view this information automatically on a patient's summary. We're not at the point where we can, what is output any of the information that we're inputting in the system yet, but this is in the pipeline in the following releases that we have customized reports for indicators and for patient management and so on. Always, another one down here. Catini, OCA, MSF. Thank you very much, really very nice presentation, a lot of work, I must have gone into it. Just a brief question. You mentioned the software, the open source software BAMI. Is that in fact the application, the application is open source software that you've been using and if so, who owns the data and how is the patient privacy being insured, thank you very much. So the software is open source, but I don't wanna say something that I'm not sure about. The servers are hosted inside of our, in our hospital with one that's being out, one that's based in coordination a couple of miles away. But this is not open in the cloud or in a network. It's, you can only access it if you're inside of our hospital and connected to our network. So this is the way it's, the data's not being transferred anywhere right now. It's simply hosted inside the hospital.