Penda Health’s Clinical Decision Support

While access to healthcare is a challenge in Kenya, it is increasingly recognized that quality of care is also a major problem. The primary care setting is uniquely suited to address a variety of public health problems, including the rapid detection of outbreaks and diseases of public health importance, and yet, primary care clinicians are insufficiently equipped with technology to improve quality and detection.

Penda’s solution involves frontline primary care clinicians using a suite of technology tools to screen, diagnose and treat common illnesses through point-of-care decision support and analytics tools to monitor trends.

Global scaling of this CDS system has the potential to substantially improve quality while maintaining affordability.  More people will be able to access high quality primary care, which further reduces healthcare costs by managing conditions in a more efficient setting than when diseases are more advanced.  

Clinicians working on the front lines of healthcare are responsible for treating a huge range of conditions each and every day.  A variety of factors makes it very difficult to consistently apply guideline-based treatment for so many conditions, including high patient volumes, lack of ongoing training, and inertia in how conditions are treated.  

The result is that patients don’t get the best treatment for their condition.  A common example in Kenya is the treatment of diarrheal illness in children -- the leading cause of death in children under 5.  There is strong medical evidence that affordable treatments such as ORS and zinc save lives and reduce future episodes of diarrhea, yet, according to quality of care studies in Kenya, fewer than 1% of children in Kenya receive zinc for this condition!  

This gap between what medical evidence tells us is the best treatment for a patient’s condition and what patients actually receive in the real world is known as the Implementation Gap, and it is the key problem that Penda is trying to address.

Penda’s solution is to integrate clinical practice guidelines at the point of care -- directly within our clinicians’ workflow.  The technology is a “Clinical Decision Support” (CDS) engine which is “queried” from our electronic health record via an API in real time.  The CDS returns clinical practice guidelines associated with patient complaints and diagnoses in order to prompt evidence-based suggestions within our EMR.  After a clinician selects a diagnosis, such as “Gastroenteritis with mild dehydration,” the CDS system returns several evidence-based suggestions such as zinc sulphate, ORS, and patient education materials.  The clinician must accept or reject these suggestions based on their knowledge of the condition and the patient.  With one click, the path of least resistance for end-users is to be in compliance with national guidelines.  

All of the data regarding this clinical decision-making is available to our internal clinical leadership.  This feedback loop is critical for improving the quality of the suggestions themselves and following up with clinicians who are routinely out of compliance with clinical guidelines.  This robust data allows us to detect outbreaks early and show real-world compliance with best practices for outbreaks, such as Covid-19

Our solution impacts nearly every patient who comes through the doors of a Penda Medical Centre.  As one of Nairobi’s largest healthcare providers serving low and middle-income areas of Nairobi and its environs, our solution will affect more than 250,000 people this year alone.  The potential market size within Nairobi itself is almost 9 million people, and since this is a technology-based solution, it has potential for rapid scale across the continent.  This population has suffered from the lack of affordable and high-quality primary healthcare, which increases the risk of poor disease outcome and population-level outbreaks. 

Penda’s vision is that “Everyone in Africa has Health Care they can Trust.”  As a social enterprise with a focus on trust, we are naturally extremely inclined toward understanding the experiences of our patients.  We have a robust customer experience team, complete with NPS surveys, exit interviews, and focus groups so that we truly understand the impact that mis-diagnosis or delayed diagnosis has on our patients.

We have seen first-hand over the last 12-months how detecting and responding to outbreaks in the community setting is critical for the health of our population.  Within days of Covid-19 being detected in Kenya, Penda was able to roll out Covid-19 clinical decision support tools to over 100 providers at 20 medical centers across Nairobi.  Our tools allowed us to close gaps in quality of care by showing where we were failing to record oxygen saturation and respiratory rate.  These tools are the technology infrastructure that is required to rapidly respond to outbreaks.

These tools have already been rolled out across 22 medical centers and over 100 clinical officers within Penda.  They are being used in over 500 patient encounters every single day.  Our challenge now is to continually expand our evidence-based guidelines, while also improving our existing system in response to feedback from frontline clinicians.  We also need to continually work to build buy-in and usability of the CDS tools among frontline staff.

Clinical decision support systems have been in use in high resource settings for years.  Unfortunately, there is a lot of pushback against them in highly resourced settings, in part because they tend to be used for a limited number of conditions and are intended to be used by expert or specialist clinicians who may not benefit as much from reminders of clinical guidelines.  What’s really innovative about our approach is that we have successfully matched these clinical tools with non-physician healthcare providers in a low-resource, primary care setting.  This leads to much higher levels of impact, because baseline adherence to such guidelines is poor, and higher levels of adoption, because clinicians in our setting struggle to stay on top of guidelines for the hundreds of conditions they routinely encounter.  We believe that many other primary care organizations in low-resource settings would also use this approach, but the bottleneck is that in most low-resource settings, clinicians are not routinely using a fully-functional state of the art EMR to provide care -- a problem we have already solved through years of work and investment.

How many people does your solution currently serve?250,000 a year 

How many will it serve in one year?500,000 a year  

In five years?1,250,000 

The aforementioned SDG-driven indicator will be measured in real-time, as opposed to retroactively. Data collected from all digital platforms, including our EMR is pushed to our cloud-based data warehouse. This includes demographic and medical data, lab tests, results, diagnoses, prescription, call/chat logs, phone numbers, and clinical notes. Because this data is collected proactively, we are easily able to provide visibility into all aspects of the project. We will continue to use Microsoft’s Power BI for our analytics. 

To gather patient feedback, we will continue to collect data via three media: follow-up calls, exit interviews, and SMS surveys. We will continually monitor our Net Promoter Score for each tool to understand where gaps/bugs lie and analyze the results of our exit interviews in order to ensure that we are providing a high quality of patient care and patient satisfaction. We will follow up on all complaints to ensure that patients have their needs met.

Full Time Staff- 472 

Consultancy- 5

Part Time-3

We hire passionate, innovative and talented people who can deliver results for the company. We then provide continuous training, and digital tools to enable them to provide quality care. 

• Chief Medical Officer, Dr. Rob Korom, MD: Dr. Korom is a Harvard-trained primary care physician experienced in healthcare quality management in Kenya.  He has worked with Penda in Nairobi since 2016.

• Head of Quality Improvement, Dr. Sarah Kiptinness, MBChB, MMed:  Dr. Kiptinness is a Kenyan-born family medicine consultant physician with expertise in antenatal and palliative care.  She leads the design and execution of the tech interventions to ensure they achieve the patient-centered and public health outcomes we expect.

• Chief Product Officer, Rama Siva, MBA: Mr. Siva brings robust technology and product management experience from both large international companies such as Apple in the U.S., and start-up companies in both India and Kenya.  He has direct entrepreneurial experience with e-learning platforms directed at employee growth and development.  Mr. Siva is responsible for implementation of new technology tools at Penda, including the EMR vendor relationship, our e-learning platform, and patient-facing technology tools.

• Head of Business Intelligence, Arnold Matzdorff, MS:  Mr. Matzdorff is a South African business intelligence leader who has been leading Penda’s business intelligence team since 2018.  He is responsible for the end-to-end data stack and ensuring that data metrics are accurate and up to date to close the loop with frontline staff.

Penda is deeply committed to diversity as we believe that a diverse team will lead to the best outcomes for the organization.  More than 50% of our leadership team is women, including our CEO, CFO, and General Manager.  Our leadership team includes four Kenyans, one Ugandan, one Indian, and two Americans.  We know from experience that deep knowledge of our patients and our market is critical for gaining traction and building a profitable business.  We are therefore intrinsically motivated to cultivate a diverse workforce with a variety of backgrounds rooted in East Africa.

If Penda were selected as a Solver, we would be most excited to engage with digital health technology mentors within the network.  The field of Clinical Decision Support in global primary care is still very young, and therefore very exciting!  We are eager to connect with others who have experience at any level of implementing CDS in a real-world setting.  We could benefit from learning from people on the technical side of CDS as well as from people with on-the-ground implementation experience and driving culture and behavioral change in large healthcare organizations.

As described above, we are eager to engage with experts in the implementation of healthcare technology.  This includes on the developer side, and enhancing the user experience as well as the nuanced targeting of CDS recommendations for particular conditions.  This also includes experience in driving technology uptake throughout a large organization, specifically in healthcare.

We would love to partner  with 

• Partners HealthCare Digital Health innovation team: Experts in using digital health tools to improve population health
• Iora Health: Real-world practitioners of using outstanding frontline primary care tools to improve health outcomes.
• MIT Data scientists: How can we use our massive databases to further improve population health, provide earlier detection of disease outbreaks and integrate cutting edge tools like machine learning to further enhance patient care?

Healthcare in Kenya is fragmented and patients describe it as frustrating.

Penda Health is a women-led company (CEO, CFO, General Manager) that also caters primarily to women and children as customers.  The technology described in our MIT Solver project improves the quality of primary healthcare for the tens of thousands of women who access our services every month.  The care we provide to women improves access to family planning, safe antenatal care, and post-partum care.  The Clinical decision support system described above makes it more likely that women will receive evidence-based and patient centered care for the common outpatient conditions that they come in with.  There is no question that Penda is promoting a world that empowers women and their families.

At Penda, we are just starting to use machine learning to manage our drug supply chain by creating automated orders to our suppliers based on historical trends and usage across our network of 22 medical centres.  We are now at the point when we can start applying machine learning and AI to our clinical care given the vast array of structured and unstructured clinical data we have in our data warehouse.  We believe an early target of our AI work could be to create automated suggestions for lab tests, prescriptions, and diagnoses based on the other clinical data that is available.  At first, the goal would be to make the workflows more efficient.  As we go on, we hope that AI predictive models could actually improve the quality of care across the nearly 30,000 monthly outpatient visits we have at our network each month.  We would be thrilled to partner with the McGovern Foundation to bring AI to primary care in Kenya!