Canadian Government Executive - Volume 26 - Issue 03

8 / Canadian Government Executive // May/June 2020 TECHNOLOGY By Nicholas James Rumble, Betty Ann M. Turpin, Ph.D., Jack Bryant Artificial Intelligence and Data Analytics Better Health Outcomes by Applying Table 1. Target Disease classes Asthma Atrial Fibrillation Cancer CHD (Chronic Heart Disease) CKD (Chronic Kidney Disease) CDPP (Cardiovascular Disease Primary Prevention) COPD (Chronic Obstructive Pulmonary Disease) Diabetes Dementia Depression Epilepsy Hypertension Heart Failure Hyper/Hypothyroidism Mental Health Obesity Stroke T he 21st century has brought to the forefront the power of data analytics (DA) and artificial in- telligence (AI) as a means of im- proving health care around the globe.[1] As governments embark on data analytics and AI solutions, the question must be asked if the solution is able to deliver positive out- comes. This paper presents a solution that can be applied within a healthcare system and draws upon advanced Big Data analyt- ics and Artificial Intelligence to influence patient prescription (drug) adherence [2] and leads to four key outcomes: [3,4] • Increased drug adherence • Improved patient outcomes • Actionable data that supports disease management and research • Improved cost efficiency through re - duced waste. Background Medicines represent the first line of defence against illness and are one of our most cost-efficient health interventions. They improve our quality of life and, most criti- cally, keep us alive. Despite a worldwide ac- knowledgment of the seriousness of the is- sue and efforts around the globe to improve medication adherence, there has been no significant change for the last 40+ years. The effectiveness of medicines, however, is predicated upon patients adhering to their prescribed medication(s) regimen. Non-ad- herence affects the patient’s health status, leading to unmanaged illness and a higher rate of avoidable, premature death.[5] Below is a list of well-known chronic dis- eases that require medications to support a person’s health, well-being, quality of life, and life itself (Table 1). Hence, the impact of medication adherence on patient out- comes can not be under-estimated. The economic and human costs are stag- gering as a result of drug non-adherence: • More than 50 per cent of visits to a doc - tor, 40 percent of long-term care admis- sions and more than 50 per cent of hospi- tal readmissions; [6] • Is one of the key drivers to rising health care costs in Canada; [7] • Cost of treating a patient with low adher - ence is twice that of a patient with high adherence; [9] • It is estimated that as many as 40 to 70 per cent of Canadian patients are not taking their medications as directed by their phy- sicians; the more severe the medical condi- tion the less likely patients adhere. [18] • Medication non-adherence is the source of an estimated $700 billion in “otherwise avoidable healthcare expenditure annu- ally”; [11] • Medication non-adherence is the 4th leading cause of death; [12] Chronic disease and long-term illness ac- count for 90 per cent of total healthcare expenditure;[13] over 50 per cent of total drug expenditure; [15] and escalating costs threaten drug program(s) sustainability. [14] Benefits and implications Below are key benefits that can be gained by adopting an AI-based healthcare solu- tion: [18] 1. Improved drug adherence - It has been shown that increases in adherence of between 30-37 per cent are very achievable.[19] 2. Reduced healthcare costs - A 10 per cent improvement in medication ad- herence reduces healthcare costs by up to 29 per cent.[20] 3. Faster time to treatment - With to- day’s huge patient caseloads, treating patients sooner saves both lives and healthcare costs. 4. Big data analytics tools expedite the process by factoring in unique circum- stances, such as adherence, lifestyle choices and demographics, along with the patient’s symptoms to help provid- ers make more accurate diagnoses and to formulate the best treatment regi- men in real-time. 5. Reduced hospitalizations and readmis- sions – One of the best ways to curb healthcare costs is to keep patients from entering the hospital system in the first place. 6. Risk stratification – A fully anonymized data helps track and identify the sick- est and most at-risk, and often the cost- liest patients, in a proactive way. 7. Improved medication therapy man- agement –Big Data analytics helps cli- nicians and clinical pharmacists better co-manage drug therapies, all in real time – leading to better patient out- comes.