I have been leading the QI data analytic team at Cincinnati Children’s for more than four years. This team has developed expertise and systems around managing data and deploying analytic methods for patient-centered improvement opportunities in areas such as safety, wait times, and disease outcomes.
The QI data analytic team has never tackled cost in a systematic way. There wasn’t the same QI urgency around reducing the cost of patient care, for many complex reasons, including the belief that since third parties pay the service-driven fees, cost didn’t affect the patient in any direct way. It did, of course, affect the organization, but the Finance Department handled that — revenues, expenses, costing systems, budgets, all the data with dollar signs on them.
Because Finance handled dollar signs, for 4-plus years, my interactions with the department had been minimal. It took me more than a year to learn that finance had a data warehouse (we did not) in which they stored the results of functions that mapped source data into organizational concepts that we used as well. A year later, we learned that IS had collaborated with finance to develop a new way to calculate and report daily census, leveraging data from the EHR. My team was still using the old way — leveraging a daily report from the scheduling office, based on telephone calls to nursing unit directors about bed occupancy. We needed to work with finance to obtain the new census results.
Two years later, we reached out to them again. We had acquired a Pediatric Medical Complexity Algorithm from Seattle Children’s Hospital, which ran against claims data. We hadn’t used claims data since the EHR went live and didn’t need to until we acquired this algorithm. We asked finance to provide claims data to run the algorithm, and then we reverted to encounter data anyway. They produced nearly identical results; plus, encounter data are current while claims data are always lagged.
So my department and finance lived in our own silos. We seldom, if ever, visited the others’ offices. Staff seldom attended meetings together. I did not know who in finance reported to whom, or why. I did not know what job titles their staff held. They had their questions, we had ours. They had their reports and ways of publishing them, and so did we. They had their data (revenue data from PeopleSoft, and claims from the EHR), and we had ours.
Of course we share customers, for whom our silos are a burden. If the customer wants a balanced scorecard, with some measures from finance, some from QI, and others from patient services, IS, or biomed, the customer has to behave like a hunter-gatherer. If the customer “wants data from Epic” (a common refrain), the customer has to guess whom to ask. (The really savvy ones ask everyone and wait to see who delivers first).
Our silos are the product of historic emergent forces. For many years, IS focused on implementing and optimizing the EHR. The goal was to improve the experience of performing routine transactions: Scheduling an appointment. Registering patients. Checking and recording the patient’s vital signs. Reconciling the patient’s medication lists. Placing a procedure order, making a referral, and sending a prescription to the patient’s pharmacy. Compiling and sending a correct bill.
Initially, leveraging the data generated from these transactions was an afterthought, if it was a thought at all. Thus, there was no central entity charged to engineer a comprehensive data management and analytics solution that would meet many diverse needs, and there certainly wasn’t a budget allocated to do so. But demand for analytics was always present, and it grew more intense over time. The customers could not and would not wait for a well-constructed centralized data delivery system to be devised. So every group that needed data and analytics improvised. And in this way, different groups, all with their own equally compelling needs, each developed methods to acquire the data they required, hired staff, and created reporting systems and analytic methodologies — and did so independently. The siloes emerged.
Payment reform theoretically could, and perhaps should, change all this. Fee for service is yielding to accountability for both cost and quality for defined populations, which can force an organization to concern itself with both, simultaneously. In theory, the people who deal with data and analyses for cost and quality should work together; they should not be in separate silos. In practice, however, payment reform has not broken down the siloes between finance and QI. And in response to payment reform, another silo is emerging: for data and analytic services dedicated to support the Health Network, an ACO lookalike.
Yet in early May, data and analytic staff from finance and QI came together, along with staff from IS and Patient Services. For the first time, staff from these silos sat in the same room, listening intently to the same presentation, paying attention to others’ questions, and striving to learn. There have been several gatherings since then, and there will be more. For the first time ever, I made my way to the finance offices and spent several hours in small group meetings learning about their concerns and sharing ours.
Problems with the infrastructure to support analytics had brought us together. Like bridges across the US, cracks had started appearing everywhere, almost at the same time:
- Concerns about the ETLs from source systems.
- Readier access to source data, already transformed to be analytically digestible.
- More sophisticated visualization and reporting tools.
- The general provision of analytic public goods.
At that first meeting in early May, we listened to a vendor describe its offerings. There have been, and will be, more meetings since then. When the invitations go out, they are accepted almost immediately. The meeting rooms have been packed. The tone is respectful. The faces reveal the dawning awareness of the common needs for which a common solution would be ideal, along with the different needs, for which reconciliation requires governance.
The silo-crumbling role of technology is to use its possibilities as a convening mechanism. It is bringing together, in a calm, non-threatening social space, those who are keen on the possibilities. It is still a work in progress. I am sure the silo phenomenon is widespread, so I’m very curious to learn how others have been dealing with it.