In a hospital setting, automated dispensing cabinets are used to increase workflow efficiencies and patient safety, Unni says. Unit dose medicines are prepared and stored in these cabinets, which a nurse can access after a pharmacist verifies the prescription order entered by a physician.
Pharmacies are using the large amount of data they generate in a variety of ways. The current pharmacy management systems used through pharmacy benefit managers have information about each patient’s prescriptions and prescription pick-up habits, Unni says. Several health insurance companies use this data to flag patients who haven’t picked up their medicine in recent months—an indication of medication non-adherence. This provides an opportunity for the pharmacist to either call or speak with the patient about the necessity of taking their medicine as prescribed.
Additionally, some pharmacies and health insurance companies are using big data to predict the risk of medication non-adherence, Unni says. This medication adherence prediction model is based on several variables, such as patient demographics, the number of medications a patient takes, out-of-pocket costs for medications, and past refill habits. When a patient is identified as high-risk for medication non-adherence, a pharmacist provides them with additional counseling to reduce the possibility of non-adherence. They focus on educating the patient about the importance of adherence, the need for the medicine, what to expect for treatment outcomes and side effects, cost concerns, and so forth.
Read More: 5 Pipeline Drugs to Watch
In clinical pharmacies, in collaboration with physicians, big data is now used to develop algorithms to predict treatment outcomes or medication errors. Based on past outcomes and medication errors with various patient and condition characteristics such as age, gender, duration of disease, and severity of disease, risk models are developed to determine treatment outcomes and medication errors, Unni says.
Medisafe’s AI technologies are based off of machine learning analyzing more than 2 billion managed doses. The company leverages this data to identify persona types based off of behaviors such as reasons for non-adherent behavior, demographics, digital profiles, and therapeutic needs. “Once patients match persona types, then the power of AI guides patients through their personalized journey in their medication management needs,” Shor says. For example, periodic surveys or interventions check in with patients to assess their progress. If needed, persona-based management designs will change to meet patient’s needs.
Pharmacists are also using AI to monitor and prevent opioid abuse. “Incorporating AI can remove some of the burden from healthcare providers and improve patient safety,” says Rachael Fisher, PharmD, senior clinical implementation analyst, Wolters Kluwer, Health, which provides evidence-based health information and technology. Nationwide databases and prescription history can be used alongside AI to identify patients at risk for opioid misuse or patients who are doctor shopping, which could be an indicator that they are obtaining multiple controlled substance prescriptions illegally. Also, AI can identify physicians who overprescribe opioids, so education can be recommended.
Many state-specific laws have been created in response to the epidemic. “Pharmacists can use AI to assess compliance of local laws in addition to national recommendations from organizations such as CMS and CDC,” Fisher says. “The environment is ever changing; AI can help providers be more vigilant, effective, and improve overall opioid monitoring and safety.”
Continue reading on page 3...