We’ll meet a couple of times on the phone or in person to make sure we are in full agreement as to what the problem is and what the deliverables are. In other words, we’ll be 100% in synch.
Sometimes, the problem may not be clear. No worries. It is part of our job description to unfold and lay bare messy problems. We have been in this business for 25+ years, which allows us to provide perspective and enrich the conversation with similar situations we can compare to.
Several clients of ours regretted they had not worked with us earlier. They wanted to make sure the problem was very clear first before reaching out.  They did not realize we could help articulate and frame their problems.
We start off by drawing up an inventory of data sources that are available at the client. Oftentimes, the client is delighted to learn that they own several data sources that they did not know they had. Some data sources require a TPA (we know which ones they are) and that’s no problem at all.
We then expand that list to include data sources that are freely available in the public domain (Sunshine Act, Referrals, Part B/D, NPPES, etc.).
Here’s the kicker. We bring our Panorama database to the table. This is a truly awesome database that draws upon 65+ publicly available data sources that we diligently stitched together to describe the key dynamics of Patient, Payer, and Provider Behavior at the zip level and higher.
The above may not be sufficient. The problem at hand may require additional data sources. No problem. We’ll draw upon our deep knowledge of the data landscape to recommend which ones to acquire (open/closed claims, EHR, GPO, CDM, etc.). When needed, we’ll run a head-to-head comparison to make the determine which data source to go with.
Our premise is that you should be able to run the model by yourself without us being around should you decide to.
For starters, our models are completely transparent, the antithesis of black boxes.  You will know exactly how the model works, what the assumptions are, what the inputs and outputs are, and why we chose this model in the first place. You’ll be able to take a deep dive in the innards of the model and see for yourself the cogs and wheels in action.
We’ll provide user-friendly and engaging explanations that allow  you to explain the model to any stakeholder.
There are three parts to this:
The first part is about the findings. We’ll highlight the key nuggets this study brought us and what they mean to us.
The second part is the action plan. It’s a clear set of recommendations given our goal and what has been uncovered.
The third part is the story. It is the proverbial one-minute executive summary that ties everything together. It provides perspective on the problem and clarity as to how to move forward.

Q & A

In a nutshell, we provide clarity and actionable insights from analyses. This involves a whole range of activities. From data strategy to data set building. From targeting and segmentation to resource allocation. Worth mentioning are our house specialties which include influence mapping, hospital retail spillover, managed care access, and the like.

Three things. First, you’ll get a good idea of how the industry approaches the problem of interest and how you stack up against the competition. Second, stimulate your team with new ideas, data sets, algorithms, paradigms, and ways of looking at the problem. Third, pick out and customize the best solutions out there and make them work for you.

Several ways. We are always talking with clients and data vendors and as a result are aware of the latest developments.  We actively participate in conferences, symposia, and workshops where we meet people from across the industry, people that are eager to exchange ideas. We are also avid readers and publish on a regular basis, which puts us in touch with passionate and knowledgeable people who want feedback on their views.

Three sources. First from the client and this includes data assets the client purchases from data vendors. Second, free data sources that are in the public domain and they include Open Payments (Sunshine Act), Referrals, Medicare Part B/D, Physician Compare, NPPES, etc.  Third, our proprietary database, Panorama, which merges data from 65+ sources and describes the behavioral dynamics of the Patient, Payer, and Provider at the zip level and higher.

We do. Over time, we have developed a slew of data matching tools that we deploy each time we need to combine data assets. The latest addition to the suite is MergeOO and MergeOC. These tools merge PHI-free patient-level data assets whenever the two databases to be merged encrypt their patients and physicians using different encryption engines. We have also developed several data bridges over time. One that we use frequently is the ship-to bill-to bridge which we call upon when we need sales at the account level and are given shipment sales data.

Machine Learning is the way of the future there is no doubt about that. What we know for a fact is that there are several use cases where ML is the way to go.  Take patient discontinuation. We found that Bayesian optimization leaves all the other approaches in the dust. Take response to promotion of physicians or accounts. We found that Boosted Trees, SVM’s, Random Forests and SVD’s outperform the more traditional approaches. Take matching of physician names, account names, payer and plan names. We found that an ensemble approach involving a CNN (convolutional neural net), Boosted Trees, and SVM’s takes the cake.

Of course, it depends on the project. Most projects are between 4-16 weeks. Flat rate and always very reasonable.

Yes, we do. This is a great way for us to acquire new clients as it gives us the opportunity to stand out.

Yes, we welcome them. It’s a good way for the prospective client to size us up and understand the breadth and depth of our skill set. We love it because it almost always means we’ll be working together shortly after.

Really easy. We send you an SOW (Statement of Work) that describes the objective, deliverables, methodology, fees, and timeline for your review. You sign off on the work stream and send us a PO (Purchase Order). We make sure we have a TPA in place for those data assets that require a TPA and start gathering the data. Then it’s kick-off time and the project begins.