Digital algorithm ePOCT+ successfully reduces antibiotic prescriptions in children, combating antimicrobial resistance

A new study recently published in the journal Nature Medicine explores the potential application of a digital algorithm named ePOCT+ to help healthcare providers decide how and when to prescribe antibiotics for treating sick children.  

Study: A digital health algorithm to guide antibiotic prescription in pediatric outpatient care: a cluster randomized controlled trial. Image Credit: Ground Picture / Shutterstock.com

Background

The World Health Organization (WHO) has declared antimicrobial resistance (AMR) one of the biggest threats to global health, food security, and development today. AMR caused nearly 1.3 million deaths from bacterial infection in 2019, most heavily affecting sub-Saharan Africa. This is equivalent to the sum of malarial and human immunodeficiency virus (HIV) deaths worldwide.

Unnecessary and excessive antibiotic prescription is among the major causes of the emergence of AMR. For example, in Tanzania, over 50% of sick children were prescribed antibiotics during medical consultations, primarily as outpatients, and most of these prescriptions were considered unnecessary on audit.

Electronic clinical decision support algorithms (CDSAs) are being developed to guide health professionals on the clinical management of patients, including important clinical signs and symptoms, indicated tests, and required diagnoses and treatments. Thus, CDSAs could reduce antibiotic prescription rates in children; however, these effects have not been observed in studies resembling actual practice conditions. In addition, such studies were often flawed.

The algorithm in the current study is designed to be used in combination with a series of clinical and laboratory tests under supervision to train health professionals in prescribing acute illness among children below the age of 15 years. The efficacy of this approach was tested by comparing it in a randomized controlled trial conducted in Tanzanian primary care centers.

The study extended over 11 months and involved over 23,500 consultations carried out at facilities using ePOCT+ compared to over 20,700 at usual care facilities. The investigators explored the impact of the algorithm on antibiotic prescription rates and the clinical outcomes on day seven in children under 15 years of age.

What did the study show?

The routine incorporation of the ePOCT+ algorithm led to a significant reduction in the number of antibiotic prescriptions at the primary care level at less than 25% of consultations as compared to nearly 75% at usual care facilities.

The relative risk reduction was 65% for the ePOCT+ facilities. This was not accompanied by poorer outcomes at day seven in ePOCT+ facilities, either in terms of mortality or hospitalization due to worsening of the patient’s condition.

Secondary non-referred hospitalizations, when the child required admission after the first visit without being referred by the healthcare practitioner, indicating worsening of the clinical condition, were not significantly different between the two groups.

Neither group showed significant differences in the need for additional medications after the index visit. The mean antibiotic prescription rate appeared to decline over time in ePOCT+ centers but not in usual care centers. However, about 25% of patients in the ePOCT+ arm did not receive management based on the use of this tool.

More severe diagnoses were made at ePOCT+ centers at 3.6% compared to 2.6% for usual care facilities at the index visit. The risk of referral for hospitalization was twice as high for cases attending ePOCT+ centers at an absolute risk of 1.2% compared to 1% for children at the usual care centers.

The difference in antibiotic prescription rates was most evident among children with respiratory complaints, with an absolute reduction of over 60% in this group. Among children below the age of five years, the absolute reduction was about 50%. The least difference, 25%, was in the group below two months of age, which also showed the greatest drop in clinical failure rates at day seven, by about 40%.

Overall, every diagnostic or symptomatic group showed reduced antibiotic prescription rates by 25 percentage points or more. The difference in absolute terms was low among those with malaria at about 20%.

What are the implications?

Using ePOCT+ could help address the urgent problem of antimicrobial resistance by safely reducing antibiotic prescribing.”

CDSA, coupled with laboratory and clinical guidance, reduced antibiotic prescription rates in children by almost three times compared to usual care practices. This was accompanied by comparable rates of clinical recovery, however.

The marked reduction in clinical failure among infants below the age of two months is important, as this subgroup accounts for more than half the deaths in children below five years of age.

While these findings differ from those of other studies conducted in other African countries, mainly in the size of the effect observed, this may be explained by the broader clinical approach used in the ePOCT+ algorithm. The variation in the algorithm has been demonstrated in earlier research to impact the effects of its use on antibiotic prescription rates.

Clinical uptake of tools like ePOCT+ may require its integration with existing digital tools such as electronic medical records (EMR) to minimize the increase in time needed per visit to enter the data necessary for these tools. Mentoring and feedback on-site also boost uptake rates.