Darcheeni

Approximately 8 million individuals die in Low- and Middle-Income Countries (LMICs) every year due to inefficient healthcare practices, with 60% of these deaths stemming from poor quality care.[1] Pakistan, the fifth most populous country of the world attributes this poor quality of care to a number of factors. One of them is variance in care standards due to gaps in knowledge of the physicians. This variance can be explained by the fact that 71% of primary care needs are met by general practitioners (GPs) who are not required to attain any formal training before practicing independently and no regulatory mandates exist for their continuing medical education.[2], which makes it difficult to inculcate latest evidence into practices. 

Another contributing factor is that the physicians who manage to provide a substantial level of care are overworked as the doctor-to-patient ratio in Pakistan stands at 1:1300, notably below the WHO's suggested ratio of 1:1000 [3]. This undermines the quality of healthcare by increasing the workload for existing practitioners and resulting in overburdened and burnt-out healthcare professionals. As a consequence, on average, the overburdened physicians manage to spend just 1.79 minutes per patient, bringing down the quality of care significantly [4].  To add fuel to the fire, modern day’s physician role has become two-fold: 

1. attending to the patient's needs, and
2. entering diagnoses, orders, visit notes, and additional administrative data into the EHR. 

The cumulative effect of these factors widens the gap in healthcare quality, moving it further away from the ideal healthcare delivery system. This ideal system would have one physician for every 1000 people, with each physician dedicating 7.4 to 10.6 hours per workday to provide the recommended level of care to a typical group of primary care patients [5].

To solve these problems, we have built Darcheeni, an AI-driven healthcare framework that leverages artificial intelligence to assist and supplement physicians, streamline healthcare processes, and prioritize patient-centered care. 

Darcheeni begins working in the doctors room. It listens to and records the doctor-patient conservation as it takes place. The AI model in the meantime pulls patient data from various sources, including lab and imaging reports, as well as clinical notes from previous encounters from a HIPAA-compliant database. While the encounter is ongoing, the doctor-facing web application displays real-time processing of the conversation into discrete problems, a set of pertinent questions to prompt the physician in case of any overlooked aspects, and the distribution of important sections into bins and packets. This is a categorizing methodology, where bins refer to the general topics of discussion, such as medications, past medical history, and current problems, while packets refer to each instance of a bin. 

Once the conversation is over, the doctor stops the recording, and a personalized and editable diagnosis, assessment, and plan is created. The doctor verifies and signs off on the plan, leading to the creation of three pathways: billing, storage in the EMR via an API, and forwarding to the patient-facing mobile application. The personalized care plan on the patient-facing application is presented in the form of a smart to-do list, where each actionable item is sorted by priority. At home, the patient records and updates the patient-facing mobile application. Most phone app entries go to the medical records and will be reviewed at the next visit. However, some critical entries notify the doctor or clinic administration for timely intervention. 

Over time, Darcheeni builds an Digital Health Profile of each patient which can be accessed from anywhere from their phone and with their consent. This would help the healthcare providers with their areas of strengths, weaknesses, and error patterns guiding their care of the patient, while also generating periodic analytics for optimal health outcomes and population health strategies. 

The full solution is presented in the diagrams below:

Darcheeni solution serves a diverse range of stakeholders within the healthcare ecosystem, each benefiting in unique ways:

For Healthcare Providers
• Reduced Workload: By incorporating advanced transcription capabilities for three languages and integrating local slang and shorthand, our solution significantly reduces the time and effort required for documentation. This allows physicians to allocate more time to patient care, ultimately enhancing the quality of care provided.
• Enhanced Efficiency: The inclusion of trained and assessment models streamlines diagnostic processes, treatment planning, and data analysis. This results in quicker and more accurate decision-making, ultimately improving patient outcomes and satisfaction.
• Mitigated Burnout: With a user-first approach and features designed to alleviate administrative burdens, our solution aims to reduce physician burnout. By optimizing workflows and improving overall job satisfaction, healthcare providers can maintain a higher quality of care while experiencing less fatigue and stress.

For Patients
• Improved Access to Healthcare: The inclusion of multilingual interfaces and low-literacy mobile apps ensures that patients from diverse linguistic and educational backgrounds can easily interact with their healthcare information. This enhances access to care and improves health outcomes for all individuals, regardless of their language proficiency. The system also understands local short hand and terms which allows better communication between patient and physician.
• Enhanced Engagement: By providing patients with a companion app we are able to provide a very personalized solution to them. We can gather vitals for each patient and provide healthcare solutions catered specifically to them. The app also serves as a personal healthcare assistant, sending timely reminders to patients. 

For Hospital Administration
• Documentation Improvement: Our solution addresses the major issue of incomplete documentation in Pakistan's hospitals, leading to logistical and administrative challenges. By facilitating comprehensive and accurate documentation, our solution improves the utilization of resources and enhances overall efficiency within healthcare institutions.
• Tracking and Analytics: With the influx of data generated through audio capture, detailed analytics can be generated to hold physicians accountable and provide insights into medical issues present in specific areas. This data-driven approach enables hospital administrators to make informed decisions and implement targeted interventions to improve healthcare delivery.

The Darcheeni team is based in Lahore, Pakistan, and consists of native technology experts, healthcare professionals, and community advocates. Having representatives from all three in our core team ensures that the decision-making prioritizes needs-led AI solutions tailored to local contexts.

Our proximity to the target community is a cornerstone of our strategy. Firstly, we have collaborative partnerships with local public and private health institutions all over Punjab, such as Shalamar Hospital and Indus North. Thus, the design and development of our solution are heavily guided by ongoing feedback from the local medical community. For instance, the input from doctors and nurses at one of our partner hospital (Shalamar Hospital) in Lahore has been instrumental in making the AI system culturally adept and user-centric. Additionally, linguistic diversity within our team prepares the solution to handle a composite language model that includes Urdu, English, Punjabi, various dialects, and medical jargon unique to the region. As such, Darcheeni is a community-focused initiative designed to democratize healthcare in LMICs, with Pakistan acting as a blueprint for future expansion. Eventually, as we scale to other regions, we will leverage local insights and expertise for similar healthcare challenges in other LMICs.

Darcheeni is in the pilot stage because we have developed a functional prototype that is currently deployed in clinical settings within a tertiary care hospital, established under the approval of a joint Institutional Review Board (IRB). Within this framework, we constructed a Clinical Innovation Lab (CIL) in the outpatient department (OPD) of the hospital to facilitate a controlled environment for system deployment. Starting in November 2023, the system’s foundational capabilities, multi-lingual note-taking and storing patient records, were implemented and tested in a real-world clinical setting. This initial phase focused on integrating these core functionalities into the daily workflows of two doctors operating within the CIL, who collectively attended to an average of 40 patients per day. Since then, we have been functional at two additional stations, one more in the OPD and the second one at the allied department of endocrinology. Our system has overall listened to, transcribed, translated, and created clinical notes for over 300 patients. 

We are applying to MIT Solve because we recognize the unique potential it offers to enhance and scale Darcheeni. Currently, we face several challenges including the need for large-scale data (125,000 cases) to train our models, recruitment of skilled personnel across medical and technological domains, and establishing a legal and regulatory framework suitable for international deployment. MIT Solve’s network of global partners and mentors could provide the critical support needed in these areas. Additionally, we are already discussing strategic partnerships for market entry and cultural adaptation in countries like Kenya, Turkey, Nepal, and Sri Lanka. Additional collaborative opportunities within the Solve community are vital for our vision to create an equitable and accessible health solution. Finally, like any large-scale solution, we also require financial support to expand our team and gain access to essential technological resources such as GPUs (Graphical Processing Units) to train larger and better AI models.

While the integration of AI in medical decision-making and note-taking isn't a novel concept by itself, our solution stands out due to its grassroots approach for health equity. We recognize that in low- and middle-income countries (LMICs) like Pakistan, where healthcare spending averages only USD 38 per person per year, there's a critical need to optimize the utilization of limited resources. Darcheeni addresses this challenge by leveraging AI to bridge the gap in healthcare disparities through personalized healthcare delivery.

1. Personalized Healthcare Delivery: Our solution actively listens to doctor-patient interactions, instantly generating electronic medical records without additional effort from physicians. This real-time data capture facilitates personalized healthcare delivery, empowering providers to offer timely interventions and manage chronic conditions effectively. Moreover, our accompanying app utilizes machine learning algorithms to analyze patient behaviors and patterns, enabling personalized healthcare for maximum effectiveness.

2. Localization Factor: Built with a deep understanding of local healthcare contexts, our solution incorporates local medicine terms, multilingual support, and integration of local slang and shorthand. This ensures effective communication and interaction between healthcare professionals and patients, overcoming language barriers and improving accessibility. Unlike other EMRs adapted from different contexts, our solution is specifically tailored to meet the unique needs of the Pakistani healthcare system.

3. Nationwide Accessibility of Health Records: Our solution facilitates nationwide accessibility of health records, ensuring every patient receives the care they need regardless of location or socioeconomic status. By democratizing access to healthcare data, we empower providers to deliver proactive measures, early interventions, and personalized care plans tailored to each patient's needs. This comprehensive approach not only improves patient outcomes but also promotes health equity across the nation.

4. Economic Growth and Prosperity: Our holistic approach promises a significant positive impact on the nation's GDP. By making standardized healthcare accessible to all at an affordable cost, enhancing health outcomes, and reducing workforce absenteeism, our solution contributes to economic growth and prosperity. Through increased productivity and improved workforce health, our solution lays the foundation for sustained economic development in LMICs like Pakistan.

5. Transformative Impact on Stakeholders: Beyond merely documenting medical records, our solution actively shapes healthcare delivery and policy at a national level. By providing actionable insights and real-time data, we empower healthcare providers and policymakers to make informed decisions that directly impact patient outcomes and population health. This transformative approach not only improves healthcare delivery but also drives systemic changes for better health outcomes and societal well-being.

We aim to develop our tool, Darcheeni, specifically tailored for the Pakistani population so that its implementation at a nationwide level solves the most pertinent problems prevailing in our healthcare system. 

Short Term Impact

Where on average a patient gets to be seen by a doctor for only 1.79 minutes, integration of our solution in the Pakistani healthcare system can expedite the diagnostic process and minimize human errors, ensuring patients receive timely and appropriate care. This impact is crucial in high-volume clinical settings where the speed and accuracy of diagnosis can significantly influence health outcomes.

At the same time, evidence suggests that AI algorithms can identify patterns and anomalies that are yet unknown to medical experts and thus can be overlooked by the human eye. Therefore increased diagnostic accuracy will play a pivotal role in reducing misdiagnoses.

In addition to serving as a physician's assistant in towns and cities, Darcheeni can also provide evidence-based guidance to nurses and lady health workers serving in remote areas with limited healthcare training and experience, benefitting the rural and marginalized communities. The initiative to train healthcare providers on using this system effectively will built considerable capacity within the local healthcare workforce. Training sessions will ensure that providers are not only adept at using the technology but are also more confident in their diagnostic decisions. This empowerment contributes to overall improvements in the healthcare system’s efficiency and service delivery.

Longer-term Societal and Health System Outcomes

Over the longer term, the use of our solution will lead to better patient health outcomes. According to the Harvard School of Public health, using AI in diagnostic medicine can improve health outcomes by up to 40%. Therefore, Darcheeni, through its timely and accurate diagnoses, reduction in the rates of misdiagnoses and early incorporation of latest guidelines in care plans, will directly improve recovery rates and reduce complications associated with advanced stages of undiagnosed conditions.

By making high-quality diagnostic capabilities accessible in previously underserved areas, Darcheeni addresses one of the fundamental challenges in Pakistani healthcare—disparity. This technology-driven approach contributes to health equity, bridging the urban-rural gaps and ensuring that geographic location no longer dictates the quality of healthcare received. With our solution providing actionable data and insights into patient care, healthcare policy makers will have a clearer view of the system’s performance and needs. This transparency can facilitate more informed decision-making, leading to policies that better serve the population’s health needs.

Lastly, initial testing and deployment of Darcheeni has garnered positive feedback from a subsection of its target population and key stakeholders - the healthcare providers and administrators. This underscores the immense potential of it holds for improved physician experiences, smoother clinical workflow, and enhanced health outcomes.

Darcheeni aims to make an impact by building its infrastructure on the quadruple aim of health. Following are the major impact goals we aim to achieve by nationwide deployment of this solution:

Impact Goal One
Ensure access to quality health-care services which are safe, effective, and affordable for all.
Indicator: Percentage of the physicians using Darcheeni and patient surveys that assess satisfaction.

Impact Goal Two
Shift the focus of the healthcare system from tertiary to primary care.
Indicator: The percentage of health issues resolved at the primary care level where Darcheeni is deployed, without referrals to specialists or tertiary care facilities.

Impact Goal Three
Reduce the burden of complications caused by the diseases of lifestyle (hypertension, diabetes, hypercholesterolemia and obesity).
Indicator: Incidence of the complications of these diseases (amputations, blindness, stroke and ischemic heart disease) in patients treated with Darcheeni.

Measuring progress towards these goals: 
We are initiating a trial in a diabetes clinic at a tertiary care hospital in Lahore which serves as our primary collaboration site. With its fifth largest population in the world, Pakistan is among the top 3 countries with a high prevalence of diabetes. Due to this, we believe that this trial will effectively highlight the capabilities of our solution against a complex and widespread disease. The trial, which shall run for a period of six months, aims at tracking and measuring the longitudinal health outcomes (such as HbA1C levels, urine albumin creatinine ratio (ACR) and wound healing times) in two sets of patients - the first one managed with Darcheeni and the second one managed by the traditional method. 

Darcheeni integrates machine learning (ML) and rules-based technologies within an Amazon Web Services (AWS) cloud infrastructure. At the heart of this system are our own fine-tuned models for Automatic Speech Recognition (ASR), clinical note automation, and assessment and plan generation. These models are specifically adapted to handle multiple dialects and medical terminologies to enable the generation of assessment and plans. The machine learning models are designed to improve over time through both supervised and self-adaptive learning methods. In addition to this, our patient-facing mobile application includes a smart to-do list (which uses AI to determine which tasks should be displayed on a need-to-know basis) and scheduling features, while the physician-facing web app supports clinical note and plan generation, editing, and scheduling functionalities.

Over time, as our database grows and we gain more clients, Darcheeni will incorporate elements of behavioral technology and big data analytics to enhance its healthcare solutions. The former will be used to understand patient behaviors and preferences, allowing Darcheeni to tailor its interventions and recommendations accordingly. Similarly, it will use big data analytics to derive actionable insights, identify trends, and predict future healthcare needs by analyzing vast amounts of patient data collected through its platform.

• Full-time Staff: 6
• Part-time Staff (includes undergraduate researchers as well): 8
• Contractors and other workers (includes volunteers as well): 45

Darcheeni was conceptualized in 2021 and research/development began officially in March 2022.

At Darcheeni, we are committed to fostering diversity, equity, and inclusion within our team and throughout our operations. Our leadership team is local and women-led and we actively encourage women (especially female students) to apply for development and technical roles, such as machine learning engineers and web developers. As an organization based in Lahore, Pakistan, within a research university with a rich diversity of students of various ethnicities, socio-economic backgrounds, personal identities, and regions of Pakistan, we engage with a diverse talent pool and prioritize inclusivity in our recruitment efforts to ensure equal access to opportunities for individuals from all backgrounds. Our team also actively seeks to promote student researchers, providing them with mentorship opportunities, internships, and exposure to real-world projects.

In addition to recruitment, we have also implemented rigorous harassment policies and procedures to ensure a safe and respectful work environment for all team members. Additionally, we recognize and have accommodations set in-place for team members with mental health issues and chronic diseases to to ensure their well-being and productivity. The core team has an open-door policy, where any team member can voice their concerns regarding the project, and each members opinion, whether they are full-time or contractual, is taken into full consideration before any decision regarding the project is made.

Darcheeni's business model aims to ensure it's sustainability and scalability while remaining closely aligned with the needs of the communities it serves. 

Our key resources encompass our AI-driven platform, team expertise in technology and healthcare, and partnerships with medical institutions and tech communities. Our core activities include the continual refinement of our AI algorithms, platform maintenance and updates, oversight of patient and physician interfaces, close collaboration with local health providers for implementation and feedback, and evaluation of the social impact of the system - such as improvements in patient outcomes, increased health accountability, and increased healthcare accessibility.

As a technological intervention, Darcheeni focuses on enhancing healthcare delivery through improved diagnostics and patient management, catering specifically to healthcare providers in LMICs, including hospitals and clinics with resource constraints, and the patient populations they serve. Our value proposition revolves around offering reliable, culturally sensitive medical diagnostics and patient care planning accessible on resource-constrained devices, tailored for diverse linguistic groups. Collaborations with local healthcare providers, academic institutions for research and development, and tech-forward organizations for infrastructure support are integral to our success. 

Our system deployment occurs through direct integration with hospital systems, distribution via health networks, and outreach through medical conferences and digital platforms. Our cost structure primarily encompasses development expenses, operational costs, and staff salaries (including full-time and contractual staff members) which we have so far managed through efficient resource allocation and strategic partnerships. Any surplus generated is reinvested into further research and development to test different AI models and ensure the continuous evolution of our AI platform to address emerging healthcare challenges. Our main source of revenue will be a tiered pricing model, involving a small fee per patient record for individual doctors and a monthly subscription for healthcare institutions. Other revenue pathways include subsidiary ventures such as consulting and data analytics services (offering insights into disease trends, treatment outcomes, and population health management) and investments.

Darcheeni's plan for achieving financial sustainability includes a "free-for-service" model and targeting low-income clients to ensure equitable access to our healthcare solutions. 

Under the free-for-service model, we plan to provide our services directly to clients or third-party payers, such as healthcare providers, insurance companies, and government bodies. This involves charging a nominal fee per patient record or a monthly subscription fee per physician. For example, we offer a pricing structure of Rs. 10 per patient record or Rs. 3000 per month per physician, although this value may change based on the specific needs and preferences of that particular area. Similarly, insurance companies may choose to subscribe to Darcheeni's services on behalf of their policyholders and pay a monthly fee based on the number of physicians or patients covered under their policies, while government bodies may implement our services in public hospitals and clinics.

Additionally, to reach low-income clients, we aim to establish partnerships with community health workers (CHWs) or local health volunteers who possess established trust and rapport within these communities. These CHWs can serve as intermediaries, facilitate access to Darcheeni's services, and encourage the local population to employ these services. We also intend to provide free or subsidized screenings in remote or rural areas where access to traditional healthcare services is limited, medical expertise is low, and health education is poor. Additionally, as a localized solution,we will be carrying out targeted budgeting and allocation of resources, which would decrease the otherwise high expenses had Darcheeni been a generalized, blanket solution.

Currently, our major revenue stream comes from investments and external funding, as our project is in the pilot stage of development and we are not charging any fees from the users. Aside from this, our revenue-generating efforts have shown promising signs of success. We showcased different versions of our work to a diverse audience of over a hundred investors, potential collaborators, physicians, and policymakers, eliciting an overwhelmingly positive response. This resulted in a substantial contribution of approximately PKR 6 Crores (over 215,00 USD) from Syed Babar Ali, former caretaker Finance Minister of Pakistan. Additionally, we have established collaborative partnerships with Shalamar Hospital, Indus North, and King Edward Medical University to establish Centers for Innovation and Learning (CIL). Aside from serving as basecamps to test our model's efficiency, these labs are also slated to become the initial deployment sites for our model by the end of this year. Moreover, discussions are ongoing with the Governor of Punjab and the Secretary of Primary and Secondary Health regarding the potential deployment of our model across Punjab upon its completion. Outside of Pakistan, too, we are currently negotiating with representatives from UAE, Turkey, and Rwanda.

Moving forward, we remain focused on optimizing our fee structures, expanding our client base, and formally deploying our system to work in production. By adhering to our core principles of affordability, accessibility, and quality, we are confident in our ability to achieve financial sustainability while making a meaningful impact on global health equity.