In this final installation, we will present the last three semi-finalists for the iMedicalApps Research Award at Medicine 2.0. Here we share the experiences of researchers aiming to improve screenings for cardiovascular disease in far flung regions of India, understand the underlying theory associated with diet apps, and leverage the plethora of information in consumer drug reviews to improve therapeutics.
SMARTHealth India: Development and Evaluation of an Electronic Clinical Decision Support System for Cardiovascular Diseases in India
Summary of Abstract:
In this pilot project, researchers targeted the growing rate of cardiovascular disease in underserved communities in India. Using an Android tablet app, they developed an algorithm based on a pre-existing dataset of patients for non-physician healthcare workers to use in community-based screening. The algorithm was validated against physician assessments as well as another software platform. The app was then deployed in eleven rural villages for a four week period, with data captured into an open source health record.
A total of 227 patients were screened by 11 non-physician health workers in 4 weeks, with 57% identified to be at high risk and referred for further care. Of note, however, only 30% of those referred actually made it to the appointment. Subsequent interviews with participants also revealed several important insights. There are now plans for a much larger cluster randomized trial involving 16,000 patients.
This project is an outstanding illustration of the potential that mobile health technology affords while simultaneously highlighting the fact that it is not a panacea for all that ails world health. The platform used here is low cost, mobile, and can be used without extensive training. However, it exists within a greater social context where access to care and medications are limited. Perhaps modifications where the health workers deliver education about behavior changes are needed. Another possibility is that if the lab-on-a-chip concept reaches clinical practice, doctors may be able to run a panel that involves lipids, creatinine, LFT’s, and A1c – and then start metformin and pravastatin. The potential is there and this study is clearly an important step in realizing it.
Summary of Abstract:
In this cross-sectional analysis of apps available for weight loss, the investigators sought to assess the state of the market when it comes to weight loss. Whereas many “content” studies are purely based on the app information page, researchers here actually downloaded and tested individual apps to understand how they applied (or didn’t) existing ideas around health behavior theory. They analyzed 58 apps for five levels of user interaction. These levels included: 1) general information or guidelines; 2) assessment; 3) feedback; 4) general assistance; and 5) individually tailored assistance. Investigators found that most apps provided just general information/assistance and did not follow any traditionally accepted behavior theory. More importantly, they delved into the specifics of the types of strategies used and how they fit with health behavior theory.
There is a large and growing literature around health behavior theory, an area of critical importance as we face an epidemic of disease largely related to behaviors. These investigators delve into the available apps in great detail, to a far greater extent than other studies which simply use the information pages for the apps. In doing so, they provide not only a better understanding of the existing market but highlight opportunities for improvement. We look forward to hearing the details of their assessments of these apps and the proposals for how apps can apply behavior theory to better achieve the goal of promoting healthier lifestyles.
Summary of Abstract:
Not only are patients reviewing physicians and hospitals, they are also reviewing prescription drugs. In this study, investigators attempted to use these consumer-generated reviews to garner insights into drug efficacy. They analyzed the reviews of 96 drugs on webmd.com that treat a total of 76 conditions. While it’s unclear how free text reviews were translated into standardized efficacy assessments for comparisons, they found that they were able to identify statistically significant differences for drugs that treat 11 of the 76 conditions. Of the 11, six had head to head comparisons available for the drugs in question; of these six, four of these comparisons were in agreement with the assessments based on consumer reviews.
In some ways, this study is reminiscent of the challenge of counting the number of jelly beans in a jar; while an individual guess is likely to be inaccurate, the collective mean of many will get reasonably close to the mark. While there are innumerable factors related to how a particular patient will react to a particular drug, the collective experiences being captured on a number of web-based forums has potential to yield valuable insight. As it’s unclear from the abstract how many reviews were used or how they were quantified, it is tough to draw conclusions about the validity of the strategy used and its results. Nonetheless, it is a very interesting idea and one we look forward to hearing more about.