Patients who fail to take medicines as prescribed are costly to health systems and risk seeing their own conditions worsen. HealthNet Homecare, a home care services company, is trying to make it easier to predict which patients are most likely to fall in to this category.


The company is developing Adhere Predict, a solution based on predictive algorithms, in collaboration with machine learning experts at University College London and Brunel University. 

Poor medication adherence is due to a multiplicity of clinical and non-clinical factors, including disease severity, length of illness, access to therapy and education levels. The interplay between those factors can vary from one patient to another and even change with individual patients. 

Medication non-adherence costs close to £1 billion annually and affects between 30% and 50% of patients, primarily those with chronic diseases, according to Ejike Nwokoro, head of patient insights and data strategy at HealthNet Homecare. Non-adherence can involve either patients who drop their medications completely or those who are not taking them as directed. 

“It’s a huge issue and not good for the patient,” he told Digital Health News. “If patients aren’t taking their medicine, their quality of life nosedives.” 

The company provides patient support programs that are primarily designed to help patients with administration of tightly controlled therapies, he says. Patients are referred by the NHS, but the service is paid for by pharmaceutical companies that wrap services around their products. 

Established in 2015, HealthNet Homecare serves more than 160,000 patients in the UK and does a million medication deliveries every few months, involving a high volume of patient interractions. The data from those interractions will form the basis of Adhere Predict, Nwokoro says. 

“It’s gold dust, because it gives you the opportunity to intervene early,” he adds. “The time you have the greatest chance to alter treatment adherence is at the beginning of the journey, before the patient deteriorates.” 

Using just the first four to five months of data the company provided, Adhere Predict has been able to predict with just under 90% accuracy which patients are likely to see their medication adherence drop, Nwokoro said. A Beta version of the platform is due in late summer, and Adhere Predict is expected to officially launch around November. 

Adhere Predict is just the latest innovation from HealthNet Homecare. In August 2023, the company launched a phone app to help patients manage their medication deliveries. 

Source link