I was fortunate to make it down to SAS’s Health & Life Sciences Conference in Cary, NC last week. One of the keynote speakers, Clay Christensen explained his thinking on how to improve care and decrease costs in the healthcare industry. Although not directly mentioned, he also touched on different aspects of Michael Porter’s thinking on healthcare and “competition at the disease level.” It struck me that Clay’s thinking on disruption as applied to the healthcare industry also occurs at the analytics level and in a very pointed way, at the level of analytics architecture inside a healthcare insurance company.
Clay’s point was that the healthcare industry needs to decentralize itself from the hospital-centric, speciality-centric model that drives most healthcare costs today. If you think about hospitals, they are like factories. If a factory is designed to manufacture any item, then within the factory, there will be dedicated stations that perform one function extremely well but the factory is not optimized to do one thing well. Hence, it has a higher cost structure to create any specific item, but it is capable of creating any specific item you may want. Factories that are designed to create just one item can be optimized to create that one item. This applies to hospitals. Hospitals are generally designed today to solve any healthcare issue that walks in the door. Porter says that if hospitals are designed to solve one specific disease, say hip replacement, then they can become much more efficient over time at solving that one issue over time. Hospitals can solve any problem but they do so at higher cost. Dedicated facilities, say hip and knee surgery centers, can innovate and decrease costs over time and this is where competition needs to occur.
That’s nice and we want that. But even if the U.S. healthcare implemented just this idea well, it would not solve the healthcare cost and quality issues we face today. Clay also seemed to imply that because travel costs (airfare or gas, for example) are so much less expensive compared to the healthcare costs than it was years ago, that the concept of having dedicated centers and even fewer hospitals works because we can travel to them easier. The original geographic distribution of hospitals was driven in part by the fact that many people could not travel long distances for healthcare due to costs. Today this is a different story.
So where does this leave us? Clay also discussed the evolution of the computer industry. It’s this analogy that is the key to cost and quality puzzle. In the early days, the computer industry was highly centralized. Giant mainframes dominated and the costs of mainframes and solving really big, hairy problems was large. Then came mini-computers. These were department targeted machines that solves smaller problems but allow the workers in the departments to avoid the wait for the mainframe “queue” to solve their problem. When a mini-computers could solve a department-level problem, it did so at lower costs then the mainframes. Then came the desktop PC. These desktops allowed people to create documents and spreadsheets. Normal workers could solve problems that the mainframe and mini-computer could solve, but desktops could solve the problems cheaper. This push outwards to allowing people anywhere in the organization to solve their problems themselves without depending on others was key to improving productivity and lowering costs for solving problems. This is also the reason why when someone innovates analytics using a spreadsheet then asks IT to automate the solution, the cost goes from a spreadsheet-like cost to a huge cost that seems disproportionate to the original solution costs.
To decrease costs and improve quality, healthcare must follow the same trajectory. Healthcare delivery has to be decentralized from hospitals and even specialists’ offices. Healthcare needs to be delivered in clinics and in less costly venues. It’s not that all healthcare issues can be solved at a clinic, but a clinic needs to solve those healthcare problems that can be solved in a clinic environment. You are not going to have knee surgery at a clinic but you are going to go there for a physical, or some tests or a cough during the cold season. In this approach of decentralization of problem solving, hospitals do not go away but the venue that SOME healthcare is delivered in changes to a lower cost model. Essentially, by decentralizing the delivery of care, but still having care based on standards and policies directed by physicians, it is possible to lower cost and improve outcomes.
At Ajilitee we think the same way about analytics. We recognize that innovation needs to occur with people who know the clinical world and who know how to use data mining to find patterns that decrease costs and improve outcomes. In other words, decentralize and drive analytics so that an analyst is more in control of the data analysis process using tools that they want to use. Analysts like to use easy-to-use tools like Excel, Access, Tableau, QlikView, or SAS. Ajilitee is not afraid to say that data warehouse groups should embrace and support Excel and Access usage. Enabling these tools to work in a safe, enterprise way is the key to pushing down analytics. Anything we can do to make it easier for people to use data is the key to driving analytics and data driven decisions into the organization. Let’s not take away tools; let’s make it easier and safer to use them. We can do this while still being observant of HIPAA and other regulatory needs.
To make it easier to perform these analytics using these tools we have to make it easier to access the data. Sometimes this means just collecting datasets and putting them into one location then letting the analyst figure out how to use the data–part of their job is to figure out how to use the data correctly. Sometimes it means that we need to build integrated datasets in one location to allow analysts to decrease the amount of effort to work with data from different areas of the organization.
To make it easier to use these “desktop” analytical tools (or even if they are server based) companies need to build a place where data is collected in various stages of integration. Perhaps all the data is wonderfully integrated, Claims, Providers, Members, and life is good. Perhaps not. Perhaps some of the data is integrated but perhaps not all of it. At the very least, let’s just put it all in one place that analysts can go to to find the data easier. Lets centralize data for access but decentralize analytics. Let’s get serious about IT being helpful and move away from the need for IT to solve the entire data integration problem. This is the collaboration we need between the business and IT, this is why some Payers sometimes put Informatics and IT into the same group to solve collaboration and coordination problems. This is why centralizing data access (but not necessarily solving the entire data integration problem all at once or even making it the end goal) justifies cloud based analytical repositories that can scale up or down as analysts need it. This is why for some companies, a data warehouse appliance is a great answer to get data into one place as fast as possible then collaborate between business and IT to improve it over time. Let’s let analysts play in the cloud or appliance sandbox–it’s their database too!
From my perspective, Clay’s keynote supports the Ajilitee point of view. We see the analytics industry moving over time. Ajilitee can help you outsource your analytics and have a place that is inexpensive and can scale up or down as your needs change over time. Ajilitee can identify the optimal relationship model between IT and the business. Ajilitee can provide SaaS software, like our Managed Analytics offerings, that solve real problems. Ajilitee has helped Payers justify and obtain data warehouse appliances, like Netezza and Teradata, to help IT become more helpful in create successful and innovative analytics. That’s why I liked Clay’s talk and that’s why I think Ajilitee can help you power up your analytics.