Hacking Health Care : My Journey finding the Elusive PCP
The Setup
I work for a respectable Silicon Valley company and have the luxury of having a decent health care plan. But with all good things comes some unexpected problems.
Like many of you, last November, I choose a health plan and then forgot about it. Unfortunately, life had other plans for me.
Recently, I hurt my self and wanted to see a doctor (sports/general, frankly anything) and then I realized that I had not even completed the first (and somewhat crucial ) step of choosing a health plan: Choosing a Primary Care Physician!
So I need a PCP, really how hard can it be?Apparently very hard. Here is where my story begins
The kind folks at Connected Care California (my health plan) were indeed trying to help me. I found all the doctors in the network by a simple search. Example: https://www.connectedcarehealth-ipf.com/ca/
The Tragedy
The real challenge came right after that: the 10-mile radius search netted a “mere” 143 results!
In a world where we are used to rating and reviews for practically all things, here I was, faced with 143 choices!
A glaring example of review overload taken from Amazon pages.
The kind folks at Connected Care California, exported a pdf with 8 pages of doctor names (that too in font 9!!). I could have taken a dart and randomly picked one, or I could do something about this.
The Battle
I am telling you a story, so (you guessed it right) I did something about this.
My problems were the following:
- How are these doctors? What do their patients think of these Docs?
- How to sort them by specialties?
- Where are they located? San Jose is a big city, could I filter and then map these doctors?
Weapons: Github/Python/Vim/
For the first two questions, Yelp came to the rescue.
I noticed that Yelp already had reviews about most of these Docs. All I needed was to somehow feed Yelp.com all these names and locations and then get star rating and number of ratings for each of them. (Just like you, I rather go to a 4-star doc with 50+ reviews versus a 5-star Doc with 5 reviews)
I started off from a script from scrapehero/yelp_search.py on Github
And made some tweaks to it. You can see my version here.
Some pre-processing of the text file was necessary before feeding it to Python which I did it with my handy dandy Vim text processor: removing extra data (example exact address, phone, gender, network, language etc ), making each entry as one line, sorting the data alphabetically, replacing white space with “+” and comma by “2%” etc.
Here is what the input to python script looked like
Yelp is great because the queries are very easy to pass
Within a minute I had the Name/Address/ Star Rating/ Rating Count / Category / and Yelp Url for each of the doctors.
Bingo, I am in business!! One last thing, plotting all this data.
Light at the End of the Tunnel
Google Labs created a wonderful tool, fusion table, which for some reason never got out of beta. But thankfully it is still available for use.
Its beauty is that I could mark addresses and filter results.
I filtered the list by doctors with +3.5-star rating and having a review count more than ‘4’. Fusion table also allows for filtering by specialization. Further more, I color coded the dots by the review rating.
Again a great way to visually identify how far a Doctor is and what his ratings were (red dot < 4* and blue dot > 4*).
½ a day spent, and I was finally able to make an informed choice !!
TLDR; found independent reviews of my in-network doctors via a script on Yelp and Google Fusion tables.
