Recap. Research in digital health (eHealth, mHealth, big data in healthcare) is a hot area with considerable room for growth. But working in this field has some potential pitfalls that early-career researchers (ECRs) need to consider before jumping in. Last time, I described some of the expenses associated with this work and some ways that ECRs can plan for these. I also noted that this work is time-consuming, which is today's topic.
Time is a commodity in academia, and the stress of effectively managing one's time is especially high for ECRs - we're just getting the hang of what works and what doesn't, after all. Of course, any research project* can be time- and effort-consuming, and perhaps more so for ECRs. We're less likely to have large collaborative networks, and in our networks, we have less clout than more established researchers. We're also just establishing our labs and identifying responsible RAs. So projects aren't always completed on our intended schedules, and we end up doing much of the work ourselves. This stalls our progress and creates undue worry about achieving tenure.
In digital health, it's possible that you're actually unable to do all of the work yourself. For example, if I wanted to design a web program to deliver a new intervention, I would need a computer programmer's assistance. Same goes for collecting data in real time via smartphone app. As with reducing costs and/or accessing technology, maximizing the benefits of collaborations is essential for time management. If you're collaborating with non-academics, there is a good chance that your ideas about "complete" and "timely" work will not align with theirs. This could lead to a great deal of time and effort spent chasing down the products you need. See here for some signs that your collaborations are toxic.
Two other considerations about time management in digital health research. It's intuitive to expect that technology will decrease the time and effort that you need to spend managing data collection. Indeed, something like switching from in-person to internet survey administration frees up many hours that would have been devoted to participant supervision. But this exchange is not equal across all devices, platforms, and methods.
Take the aforementioned real-time data collection as an example. Once the app is set up, you just let it run and the data come rolling in, right? Anyone who has ever worked with this type of ecological momentary assessment procedure is laughing right now. Bugs in your app, device problems, and most of all, participant error will take more time to address than you can imagine. And managing the resulting data is a job I would never, ever want.
Similarly, take a web-based behavior change program. Once the platform is set up, even if it's been plot-tested, monitoring use, responding to problems, and managing any participant interaction takes more time than you expect. Students are fantastic and helpful in some respects, but they have myriad other responsibilities and will never treat the project with the same care you will. One more reason to pursue funding as early and often as possible: if you have professional staff, you don't have to bear this burden yourself. (It is my goal in life to get a grant that allows me to hire a professional research coordinator.)
Finally, keeping up with advancements in the field presents unique challenges in digital health. As noted, the field is exploding, with new papers and journals appearing every week. Moreover, technology evolves, especially in commercial industry, at an alarming rate. By the time you hear about a device or platform, design a study, learn how to use the technology, hire/train others, and collect the data, your technology is out of date. Then you have to write papers, wrangle any collaborators, deal with peer review and rejection.... by the time you actually publish your findings, the current technology is much more sophisticated than what you used. If behavioral science is your field, then the technology matters less than the way it's used. But if not, have a plan for how you'll deal with criticisms about living in the digital past.
What other challenges do you see for digital health ECRs, or ECRs more broadly? Post your comments below.
*I'm currently reading Joel Cooper's Cognitive Dissonance: Fifty Years of Classic Theory, which describes many of the key studies in this social psychology area. It's reconnected me with my incredible respect for social psychology research. You want time-consuming? They use rigorous experimental methods to test complex theories, using multiple moderators, confederates, and elaborate cover stories. Impressive.
We're back! A late Happy New Year, with wishes that 2015 has started off well. If you're an academic, it's likely that your Spring semester has started. Scranton has the luxury of a long intersession, so we still have 10 days to go. (If I'm being honest, I'll acknowledge that I've been back to work every day in January. Who can resist the call of flexible time to write papers?) You can rub it in when I'm still teaching in May.
I have had a productive intersession so far, and both papers and a fellowship application have given me plenty of opportunity to consider where I want my work to go next. Most signs point to digital tools for health behavior change, though as an early-career researcher (ECR), I hesitate to throw all of my energy in this direction. For a few reasons.
Full disclosure. Digital Health is not my area of specialization, but it is one of my interests, and my work moves more and more in this direction. In fact, this theme has come to unify a few of the disparate threads of my research. This post is a reflection on the opportunities and challenges for an ECR with interest in digital health who does not already have ample funding.
Digital health is a term that encompasses the use of technology in health promotion and healthcare, including mobile health (mHealth, or mobile applications), eHealth (electronic health, or web platforms/email), and wearable technology (Fitbits and the like). As such, it brings together computer programmers and software designers, engineers, medical specialists, big data analysts, entrepreneurs looking to design "the next big thing." And - arguably the most important - behavioral scientists. I'm biased, of course, but hear me out.
For example, I can hardly count the number of recent scientific papers and news articles that point to the simultaneous potential of wearable technology and/or mHealth apps and their failure to promote lasting health behavior change. Nearly all of these articles explicitly name behavioral science as missing or underappreciated in these domains. Indeed, what interests me in this area is identifying the missing link; my recent and forthcoming work has shown preliminary support for improvements to our use of technology-connected online social networks to facilitate and sustain behavior change.
Each of these publications, including mine, involves a call to arms for larger, more rigorous, more cost-effective tests of improvements to digital health interventions. The time is now to strike in a hot area, which few ECRs ever get to to. Perfect! Except.... damn, this work is expensive and time-consuming. Today, I'll focus on expense; part two will cover the time commitment.
Expense. With wearables, it's the device ($90-$150) for each participant; with new apps or web platforms, it's their development (which I would have to pay someone to do); for either one, the convention is to offer some sort of monetary compensation for participants' time. I did get away with offering only the device and treatment in a recent study; I had 100% retention of 12 participants over four weeks. But this sample size and time frame aren't that impressive.
An obvious source of funding for these needs is a new faculty member's startup package. If you're a new faculty member at a research-oriented institution, and your institution understands the resources required for digital health, and you did your homework before negotiating, skip this section. But let's examine whether this is a common situation. Startup packages seem to be the privilege of only those hired to the tenure track. (A good friend of mine just accepted a non-tenure track, teaching/research faculty hybrid at an R1, and got only enough to cover her statistical software needs.) If you're at a non-R1, you might have to play hardball to get an administrator who is not familiar with digital health to pony up more startup money than s/he believes you deserve.
On, and you may have heard that, in the US and elsewhere, the tenure track is disappearing. So if you're hired into a short-term teaching faculty position, you'll need to write a grant proposal. Are you eligible for internal grants? Pray that the answer is yes; at many institutions, your status excludes you from both internal funds and government grants. Unless you can get a longer-term commitment based on your grant award. Meaning you have to get the grant.
The same goes for many postdoctoral fellowships in the US. If you're in a formal training program that provides individual funding, great! Again, skip. But a good number of fellowships do not come with project funding, and same goes for internal and external grants. For example, at my postdoc institution, available funds went to graduate students and faculty, not postdocs. (I have heard that this is changing, which is great for the new class.) In health psychology, many organizations that offer small grants for which students are eligible, not postdocs.
What are your remaining options? Again, without a longer-term commitment from your institution, you won't be eligible for many internal or government grants. The NIH K-series* is one exception (see Part 2). Aside from the K, I took advantage of every possible opportunity for funding on postdoc; most required that I be listed as a co-investigator, as I was not eligible to be the PI, though the project would have been mine. I also applied for ECR grants and fellowships through professional organizations. No luck for me, but I do recommend this route. It can't hurt to try, and your vita will show that you're knowledgeable about the process.
Your other option is collaboration. On postdoc, I was fortunate to be able to collect some pilot data by adding components to existing or new studies that were initiated by my mentors. And, as noted, I ran a pilot with no funding, using devices leftover from a previous study. This worked for me, but it's not the same as having control over study design and fund allocation. For behavioral scientists, sometimes partnering with those in basic science or computer programming will make your applications or final products look stronger to those evaluating them.
Whatever you do, think beyond your first study. If funds are limited, your devices or program will need to last you a while. Can you design a second study, perhaps even in a topic area that isn't quite digital health, so that will benefit from the money you spent on project one? For example, after my upcoming pilot intervention study, my devices will be used as assessment tools in a larger longitudinal project. It was a huge relief to realize that the devices could work for a distinct purpose and interest.
The bottom line. If you're a graduate student or postdoc with interests in digital health, think strategically. Use whatever funding you have or startup you get to purchase materials that will serve multiple purposes. Same goes for any early-career faculty member thinking of getting into the area. Apply for grants as early as possible, and use the money to build toward bigger and better projects. Collaborate. And try to keep up with the newest developments; unfortunately, your ideas and technology will be obsolete in a few months. More on this next time.
