Signal versus Noise
August 14, 2011
At DocSpot, our mission is to connect people with the right health care by helping them navigate publicly available information. We believe the first step of that mission is to help connect people with an appropriate medical provider, and we look forward to helping people navigate other aspects of their care as the opportunities arise. We are just at the start of that mission, so we hope you will come back often to see how things are developing.
An underlying philosophy of our work is that right care means different things to different people. We also recognize that doctors are multidimensional people. So, instead of trying to determine which doctors are "better" than others, we offer a variety of filter options that individuals can apply to more quickly discover providers that fit their needs.
August 14, 2011
In a prior blog post, I referenced the balance between privacy and transparency in our quest to empower patients to make better decisions. The discussion around patient reviews seemed sufficiently complicated that I decided to address that separately. So, picking back up on the topic, the question is whether we would like to allow providers to hide certain reviews.
For us, there's actually not much of an issue if we knew that all reviews were true (and objective). If that were the case, the answer would essentially be the same answer to the question of whether or not to display something like disciplinary actions: yes, we can understand why providers might want that hidden, but no, we wouldn't want to hide that. The question becomes more complicated in light of allegations that have surfaced that people write false information in their reviews (whether that be one provider writing a false review for a competitor or whether that be an unhappy patient making up facts to strengthen his case).
It'd be interesting to read a study on the estimated number of patently false reviews (although, I'm sure that such a study would be expensive and difficult to pull together). Overall, patients very much look to see what other patient say -- for a while, that was the most requested feature. The sense that we got in talking with people is that patient reviews needed to be taken with a grain of salt, much as reviews for other products and services. While it might be easy to fake one or two reviews, it seems unlikely that someone would go through the trouble of creating ten different reviews for the same provider. Impossible? Nope. Improbable? We think so. And that brings us to our take on whether or not to show patient reviews: a few reviews is not a lot of signal, but the aggregation of many reviews that voice the same feedback over a long period of time is likely to mean something. Interestingly, someone just wrote in with the same opinion.
And for those who are curious, we try to offer something extra when people leave a review on our site. After typing in their feedback, we give users the opportunity to upload some documentation (e.g. an explanation of benefits form or a receipt) and request a verification. If everything checks out, we'll mark that review as a "Documented Encounter" -- it's a little like an online retailer saying that a review is from a "Verified Buyer" since they know the online identity of the individual and as well as his purchase history. Have thoughts on this topic? Let us know.
August 06, 2011
Amidst all of the news regarding the debt ceiling debate, there was one line in particular that caught my attention regarding potential health care reform. In his editorial, Paul Ryan wrote "There is a better way [to curb health-care spending]
July 31, 2011
Product development never seems to go quite as fast as we would like, and we've ended up with an impressive stack of front-end web development tasks still to be done. Thus, that's the next area that we would like to hire for. Know anyone who might be interested? Here's the scoop.
In case you're wondering, we value people's enthusiasm and their ability to learn far more than we value prior experience. For this role, we're happy to train the right person. Are you someone who yearns to make a difference in the world of health care? Let us know!
July 22, 2011
Have you ever stared at a word for so long that it starts to look funny to you? I'm pretty sure this has happened to me with words as diverse and, well, embarrassingly normal, as 'roof', 'color', 'chimney', and 'fridge'. (Fortunately, they look fine to me now, so don't worry too much.)
Unfortunately, it seems that the opposite happens when you spend too much time looking at a website. It doesn't start to look funny; in fact, the things that may have looked odd at first start to seem completely normal, and you'll probably even forget that they ever looked strange to you at all. This is why we value user feedback - hopefully none of our users spends quite as much time on our site as we do (if you do, you should probably go out and play - or get back to your own job), so your perspective is different from ours, and we need to hear it. We want to know what's intuitive, what's abysmally counter-intuitive, what you want to see that we don't currently have, what you do see but honestly would rather not.
One thing we've been learning from the user feedback we've gotten so far is that the sheer length of the lists of medical schools, professional interests, insurance plans, etc. in our filters can be, to make a large understatement, a bit overwhelming. I, with my irreparably skewed perspective, may be familiar enough with the site that I have an idea of what's in the list, how it's organized, when it will stop, and how I can efficiently use it to hone in on what I'm looking for - but it doesn't make as much sense to someone seeing it for the first time. Some people told us the lists are so long they wouldn't look through them; others see them and then assume that just scanning through the results of their initial search is the most efficient way to find an appropriate provider. If you want to find a provider who accepts Blue Cross insurance, attended Stanford for medical school, and treats a lot of patients with high blood pressure, you're something close to infinitely better off if you use the filters than if you look through the list of search results until you find the right doctor - but this isn't necessarily obvious, and we want to fix that.
The plus side of having unmanageably long lists of options is that we really do have a huge breadth and depth of information on doctors and other health care providers, and we've spent a lot of time standardizing and organizing it to make it digestible. It's just a matter of making it easy to use at the same time. This could involve using drop-down menus instead of lists, adding search boxes for people to type in the medical school or insurance company they're looking for, organizing the lists differently (for example, separating medical schools into tiers based on their rankings), or something else entirely. As we consider these and other potential changes, we'd love to hear your thoughts at our discussion boards. Just don't stay there so long that 'forum' starts to look funny - I hear that can happen.
July 16, 2011
We've known for a while that our user interface needs a redesign. The search results page sports a powerful filter interface. You can say, for example, that you want you to see doctors that accept any Aetna insurance except for Aetna Select. You can say you want to see any doctor in a certain area who has graduated from either University of Texas Southwestern or University of Virginia medical schools, all in the same query. As you update, the counts next to each option also update, indicating how many providers would still match if you were to filter on that option. The problem?
The interface ends up with a lot of checkboxes. And long lists. We closed some of the filter trees so that users wouldn't be overwhelmed with the number of choices, but that created the problem of some users not knowing about the options. There are other issues as well. With so many filter options, which attributes are most important? Some people know what's important to them (whether that be residency, years of practice, or patient reviews). Some others would like more guidance in sorting out which doctors to further research. Another issue that has come up is people wondering how are the results sorted.
Lots of questions to tackle. And this time, we'll have to keep in mind that we ourselves are not the target users. If you have ideas on how the user interface should be re-designed, please post it to our discussion board or contact us. It'd be great to hear from you. Thanks.