By Rob Korom, Chief Medical Officer, Penda Health

A mother brings her two-year-old daughter to a local hospital with several days of diarrhea.  Although diarrhea is a common condition, it can also be deadly. In 2011, diarrhea was the leading cause of death in children under five in Kenya. 

Unfortunately, all too often, children do not receive appropriate treatment for diarrhea. 

While many clinicians are quick to prescribe antibiotics for childhood diarrhea, the most effective, evidence-based treatments are actually Oral Rehydration Solution (ORS) and zinc sulphate.  A CHAI study in 2013 showed that despite guidelines clearly recommending ORS and zinc as the mainstays of therapy for diarrhea, “less than 40% of children in Kenya used ORS and less than 1% used zinc.”

The difference between what guidelines recommend as evidence-based treatment and what is actually delivered to patients in the real world is referred to as the “implementation gap.”  There are various reasons for the discrepancy, including insufficient clinician training, busy and chaotic working environments, and inability to easily access guidelines in real time while the patient is in front of you.  

Penda Health is closing the implementation gap with clinical decision support.

Penda Health is a global leader in closing the implementation gap in primary care.  We do this by heavily leveraging technology throughout the patient’s visit to any of our 20 outpatient medical centres around Nairobi.  The core of our approach is to use Clinical Decision Support at the point of care in as many visits as possible.  This means ensuring standard guidelines are linked automatically to diagnoses within our Electronic Medical Record system.  We have jointly built this point of care decision support tool with our fantastic EMR colleagues Easy Clinic in India.

To give a sense of how this directly improves patient care, imagine you are a healthcare provider diagnosing diarrhea with moderate dehydration at Penda Health.  Here is what will pop up on your screen as soon as you enter that diagnosis:

Importantly, we don’t allow healthcare providers to move on in the visit without acting on these suggestions one way or another.  Clinicians still have the option to reject suggestions in case the unique circumstances of the case require a different approach.

The implementation of the Clinical Decision Support guidelines is only half the battle of closing the implementation gap and achieving high-quality care for our patients.  The other half is continuous, automated measurement of our performance in each of these areas and responding to areas of concern.

Making data actionable is the other half of closing the gap.

Thanks to our outstanding data analytics team at Penda, we’ve built detailed, real-time, interactive dashboards using Microsoft Power BI to constantly monitor our performance against our own internal guidelines.  Because the acceptance or rejection of the various guidelines is in structured data fields, we’re able to easily view the provision of quality care across our branches, over time, and even down to the level of individual clinicians.  This has unlocked an ability to immediately discover areas that require additional clinician training or further assessment.

Gaps in providing recommended care are not always due to clinicians themselves, but can be related to very low-tech structural challenges.  In one example, we used our automated dashboards to assess how often we recorded the patient’s oxygen saturation when they presented with possible Covid-19 symptoms.  Our dashboards showed that one of our medical centres was performing much worse on that metric than others, which prompted me to visit and ask questions.  

We quickly discovered that because one of the pulse oximeters in the clinic had disappeared, the team had decided to keep it under lock and key in the pharmacy area.  That had created a significant barrier to clinicians using the pulse oximeter to measure and record the oxygen saturation of our patients, and they only took the time to do it when the patient appeared very short of breath.  We agreed with the team that we needed to make the pulse oximeter more accessible to clinicians, even though there was a risk the device would be lost again.

We would never have been able to even identify that problem if we didn’t have access to automated, real-time, comparative data across our network.  It is clear how much potential there is in tech-enabled quality measurement in primary care.

As of early April, we have clinical decision support recommendations for more than 100 common conditions – or about 60% of visits – and real-time visibility into our performance on all of them.  Our goal is to expand our automated quality measurement to include more than 90% of all visits at Penda (that’s over 900 per day across the network!) by June of 2021.  

When we think again of the child who comes to a medical centre with diarrhea, it should be clear that Penda’s Clinical Decision Support and automated quality measurement make it vastly more likely that she will get safe, effective, and evidence-based care.