ID-PRO: Infectious Diseases Prescription Resource Optimization

We propose to establish an infectious diseases database leveraging the existing Sindh Intensive Care Surveillance System (SICSS) and develop an AI-enabled mobile application. This will enhance data-driven patient care and public health initiatives for empiric antibiotic prescriptions in all inpatients across the province of Sindh. 

Primary investigator; Prof Madiha Hashmi

We are addressing the lack of evidence-based practice in antimicrobial management due to insufficient data availability, leading to incorrect and unregulated empiric prescription of antibiotics. Low and Middle Income Countries (LMICs), such as Pakistan, bear a substantial disease burden, with over 60% attributed to communicable diseases and among the top causes of mortality are respiratory and gastrointestinal infections. 

The WHO AWaRe (Access, Watch, Reserve) classification emphasises responsible  antibiotic use. “Watch” has high potential for resistance and “Reserve” is to be used as “last-resort”. Despite Pakistan’s National Action Plan for antimicrobial resistance (AMR) of 2017, every 8 out of 10 prescribed antibiotics in hospitals are “Watch”, and there has been an alarming increase in the use of “Reserve’’ antibiotics. Therefore, reckless prescription practices, lack of infectious disease surveillance systems and substandard healthcare access contribute to rising AMR, nationally. An estimated 4.95 million deaths per year are attributed to AMR, globally. 

Combating this issue demands a robust approach of which we aim to undertake  accurate data collection for improved surveillance and implement effective antimicrobial prescribing practices.

The mobile application tool serves inexpert or unsupported healthcare professionals within hospital setups, addressing the need for appropriate empiric antimicrobial prescription. In Sindh, 53% of public hospitals have residency programmes. However, there are lack of mechanisms to audit or quality assure the training, leading to inappropriate antimicrobial administration by first responders. This is aggravated by only 8% primary care setups having existing antimicrobial stewardship practices, highlighting the need for widely implemented antimicrobial stewardship programmes. 

We piloted a project to develop a stakeholder-led antimicrobial stewardship programme in five PRICE collaborating ICUs. This involved (i) an audit to understand the current antimicrobial prescribing patterns using an ECRF, (ii) generating antimicrobial indicator reports and (iii) monthly focused group discussion with stakeholders (senior and junior ICU clinicians, clinical pharmacists, microbiologists and ICU nurses) to discuss the report and provide feedback. The aforementioned ECRF will be used to develop the infectious disease database. The stewardship programme allowed customization to enable collection of pertinent variables required for informed decision-making. A key finding of this study, to be published, was to prioritise the training of junior doctors and provide support through a user-friendly, point-of-care tool for judicious antimicrobial prescribing. 

Our solution serves the public with a low-cost, low-maintenance infectious disease database accessible even in remote areas. The indicator reports and surveillance dashboards enable informed decision-making by healthcare professionals and policymakers, including the ministries of health, drug regulatory authorities and hospital management. 

Our free-to-use clinical decision support system promotes judicious antimicrobial usage, furthering efforts against antimicrobial resistance. We will provide an open-source CDSS protocol for replication and adaptation. 

Additionally, the data from our infectious disease database will generate peer-reviewed publication, reducing knowledge gaps. These outputs contribute to global well-being by enhancing surveillance, guiding interventions, and building a network in the fight against antimicrobial resistance.

One of three individuals in a population of over 240 million lives in extreme poverty in Pakistan. However, in 2023 over 80% of the total population had an active cellular connection. Acknowledging this, we chose the power of technology to bring about change.

In 2022, healthcare costs consumed above 10% of individual household budgets, with over 60% paid out-of-pocket.

Amidst these challenges, LMICs share a disproportionate burden of communicable diseases and multidrug resistant organisms (MDROs). According to WHO, Pakistan ranks fifth in high tuberculosis rates, bearing 67% cases in the region with 510,000 cases/year. XDR-TB (extensively drug resistant tuberculosis) accounted for 3% cases in 2019 with 32% classified as pre-XDR-TB.  Over a three year period, of the 13,000 XDR typhoid cases in Karachi alone, 37% affected children under the age of 4.

In 2020, 34% of Gram Negative bacterial isolates showed Metallo-beta-lactamase producing Carbapenem-resistance (CRGN), nationally. Inappropriate empiric antibiotics are a major determinant of mortality in critical care patients with CRGN sepsis. 

By empowering healthcare professionals to make empirically guided prescriptions and access to predictive data, we anticipate a shift in prescribing behaviour, with greater emphasis on improved practices and a reduction in inappropriate antimicrobial prescriptions.

Over the next year, we plan to expand the CDSS to all units of PRICE, building a national repository for infectious disease and antimicrobial data in critical care services. We will liaise with provincial health ministries and clinicians to ensure widespread adoption. In parallel, the free-to-use mobile app will be accessible to all hospital-associated healthcare professionals nationwide. 

The PI is a co-applicant for the grants of Critical Care Asia and Africa Network (CCAA) and International Severe Acute Respiratory and Emerging Infections Consortium (ISARIC). [Collaboration for Research, Training and Implementation in CCAA-Wellcome Invited - Innovations 224048/Z/21/Z; Foreign, Commonwealth & Development Office and Wellcome 222048/Z/20/Z]. Over the next three years, we can leverage these networks to transform the project into a global initiative. 

CCAA is a collaborative platform committed to advancing critical care research, creating networking opportunities, and addressing healthcare challenges in the regions of Asia and Africa. Engaging with ISARIC, will extend our solution to low-resource settings in Eastern Europe, Southeast Asia, and South America.

The CDSS will be an open-source implementation model. This will facilitate the development of evidence-based guidelines and policies tailored to the specific national contexts, ultimately addressing antimicrobial resistance on a broader scale.

Our impact monitoring for the infectious diseases database in SICSS focuses on achieving 100% data completion from ICUs. This involves collaborative efforts in data-sharing protocols and standardised collection methods.  We will conduct site risk feasibility, initiation training/visits, and ongoing online/telephone check-ins for Source Data Verification (SDV) to ensure data accuracy.  Weekly meetings will be held with five units at a time to review key variables of 1 patient within an 8-week cycle.

Our AI model will undergo rigorous model performance evaluation, employing various techniques used to assess effectiveness, accuracy and reliability. Key indicators include application usage metrics, user engagement, and a subsequent increase in user data. Continuous support through training and feedback sessions ensures usability.

Monitoring and evaluation of the application (web and mobile) involves monitoring active users, sign-up ratios, and user engagement with iterative improvements based on feedback.

Collaboration with SICSS ensures standardised data collection. Engaging a data scientist initiates data processing, validated for accuracy. Training sessions ensure real-time data access, monitored through site visits and online feedback. The plan is comprehensive, ensuring a systematic approach to measure and enhance project impact.

In the next two years, potential barriers include financial constraints for hiring personnel, AI-model designing, application development and user training efforts. We aim to secure funding through grants and government initiatives. In Pakistan, there is scarcity of clinical microbiologists and infectious disease consultants, with public hospitals having either no microbiology laboratory on-site or limited supplies to run cultures. We plan to collaborate with designated out-sourcing laboratories to ensure sustained support and validation.

In rural areas, limited access to reliable internet connectivity may hinder application usage. To overcome this, we will prioritise lightweight data transmission. 

During implementation, cultural acceptance of CDSS will require ongoing engagement with local authorities and clinicians. We will identify a champion within each healthcare facility to advocate for judicial antimicrobial use. Training sessions and workshops to familiarise clinicians with the application will emphasise positive impact on patient care. Ongoing support and regular feedback sessions will be provided to address any concerns.

Over the next three years, we will focus on building local partnerships with hospitals and government agencies to integrate the application into existing healthcare systems. Continuous user feedback and iterative improvements will be essential to enhance application usability and effectiveness in diverse healthcare settings.

We are applying for the unique focus that The Trinity Challenge has on addressing global health emergencies, with our proposed solution. A prominent barrier is the potential misuse of the mobile application by non-healthcare workers. The Trinity Challenge's expertise and network can provide valuable insights and resources to enhance our access control measures. Specifically, engaging with The Trinity Challenge will facilitate collaboration with professionals in data security, privacy, and ethical considerations. This will enable the development of frameworks to mitigate the identified risks effectively. The Trinity Challenge's support will be instrumental in improving our solution against misuse, ensuring its secure deployment for global infectious disease monitoring.

ISARIC from Nuffield Department of Medicine at oxford university, UK is a global federation of clinical research networks, providing a proficient, coordinated, and agile research response to outbreak-prone infectious diseases. Our solution can be scaled up through ISARIC during non-pandemic time to prepare for an outbreak

We can leverage the dedicated ADVANCE-ID network for implementation of our solution.

Tropical Diseases Foundation, Philippines to implemet our solution in another LMIC

Email from Home Institution