Check out the photo to the right. The patient did not do well. First, they stopped taking medicine on time, then they stopped taking medicine. The patient fell, and then died. The patient was seen by his oncologist well into the period of the chart showing his steep decline, without benefit of this data. Would would have happened if this data was available at the time of the visit?
I am in Nashville, TN, at the Vanderbilt Center for Better Health, the über cool innovation center at Vanderbilt University (everyone has one these days), where Project HealthDesign teams do much of their work. Five teams are examining the integration of patient generated data, which they refer to as “observations of daily living” or ODL’s into clinical care and health decision making. This includes more than just medication adherence data – anything from moods to peak flow to ability to make coffee in the morning (through the use of clever sensors).
I am always grateful when an innovative project like this has a well stocked web site so I don’t to explain how amazing it is here –feel free to explore – instead I am posting my impressions from the review of the 5 teams’ work.
And here they are –
- ODL are a form of the patient voice, and a codified version of it that can be analyzed – it appeals to the science and the art of medicine
- Their integration into care could change the conversation from what doctors think they need to know from patients to what is actually helpful to know about their patients. For example, a patient that fits the physician profile of being non-adherent but in fact has an excellent adherence record may not need to be counseled about medications relative to other health issues at a visit.
- ODL could serve as an early warning system as in the example above, or a recognition system for the times when patients are able to demonstrate achievement of their health goals. Both could be motivating for care provider and patient (and family) at critical times.
- ODL, ironically, could promote the health system’s understanding of social determinants of health – when asthma controller medicines are shown not to be used, we saw evidence of a search for the “causes of the causes” of poor adherence, such as housing conditions that created reactive airways
- This work will help health care understand which people and in what states are successful tracking ODL (of various types, active, passive, hybrid). I was cautioned, though, that the teams have not arrived at demographic segmentation of which people are most likely to capture and use ODL for their conditions.
- I think a lot of people believe that understanding the patient experience around an illness or wellness and the community conditions around those may be more important than understanding therapeutic treatment options. This work brings us closer to making the gathering of that data possible.
And…I’m not sure how algorithmic it has to be, or how much computing needs to be done. I can imagine the image at the top of this post being presented to patients’ primary care physician and them knowing immediately that something is wrong. The human brain coupled with a strong patient-physician relationship is an amazing piece of computational hardware.
Because this work is still in progress and will eventually be submitted for publication, I am not able to share any preliminary results here. LIke my other favorite Robert Wood Johsnon Foundation Projects (repeat after me: @myopennotes , @AF4Q ), stay tuned to the literature, it’s worth the wait.
More photos from the day below, including the crazy hats ….