Is using patient intervention data to predict adherence a new idea?
William: The concept and history definitely goes back a long time. Most organizations who have used this data are looking at response rates and trying to better segment the interventions they perform. But up until recently, there hasn’t been a lot of true AI (artificial intelligence) use of this intervention data.
What challenges can patient intervention data overcome?
Clifford: Organizations use intervention data to reach a number of objectives. Some organizations try to get patients to initiate therapy, some try to get them to be adherent generally – so that they maximize fills. Some are trying to help patients be adherent in a way that achieves a particular quality metric.
What are the different kinds of patient intervention data, and how are they captured?
Clifford: There’s structured intervention data, which is all data except free text. And then there’s unstructured intervention data, which is free text, like notes from an intervention call. There are also call recordings, which can then be turned into transcripts. Another source of data from call recordings is contextual analysis.
How is AllazoHealth’s use of intervention data different?
William: A lot of organizations first commit to a specific intervention, and then try to maximize patient engagement with that intervention. For example, when planning phone calls an organization might consider “Well, should we change time of day? Do we change days of the week? Do we use a different language at the beginning? Do we introduce the topic differently? Do we have an automatic dialer, versus a live dialer?” Those are all factors that can affect engagement rates. But it’s focused on just that one intervention.
Instead, AllazoHealth looks at intervention data holistically. We predict which combination of channels, content and timing will have the most impact on an individual patient, optimized for your clinical outcomes.
Clifford: I think it’s important to note AllazoHealth combines intervention data with other data to provide the most accurate view of patient behavior. It can be combined with enrollment, eligibility, and historical claims data. It can be combined with copay card data. It can be combined with social determinants of health, and consumer behavior data.
In addition, we have access to data from interventions that have been delivered to millions of patients to complement the data from a specific program’s population.
How broad does a client’s set of intervention data need to be for AllazoHealth to deliver effective outcomes?
Clifford: Sometimes we have prospects who feel that “I don’t have emails and text, I only have live calls. So I need to get these other things in place to make use of AllazoHealth.” But that’s not true. Even with just one intervention channel, there’s are a lot of options, such as different timings of interventions, targeting which patients you do more with and who you do less, and different content. Of course, it’s better to have more diverse channels. But we’ve delivered effective optimization with a limited data set.
AllazoHealth utilizes artificial intelligence to make a positive impact on patient healthcare outcomes. We help pharmaceutical companies, payers and pharmacies improve patient adherence, gaps in care, and return on program investment.
Our AI engine learns from a comprehensive data set of over 14 million lives, including payer, provider and retail claims, social determinants of health, and patient interventions. We can access broader and deeper data on individual patients than any individual organization.