“There's no machine learning model that can solve everything. In the end, it's the human who has to decide, but PFS gets pretty close.”
–EVP of Revenue Cycle, Northeast Health System
PFS Group operates a <#>-person customer service center out of our headquarters in Houston, TX, in addition to three other locations across the continental U.S. Our clients are large health systems collectively representing $3B in outstanding A/R.
As we continue to grow with our success in customer service-fueled revenue enhancement and conversion, outbound call volume rose as a result of increasing demand for our services. In our early days, while we were still growing into our telephony infrastructure, our outbound dialing strategy involved regularly calling every one of our patients. In the process, we misallocated resources by contacting patients who either had no intention or ability to pay their bill. Coupled with growing demand outpacing our ability to add manpower or resources, the cost to run outbound campaigns steadily rose while the profitability of each outbound call remained uncertain.
We started by analyzing our data archives to gain insights into patients’ financial circumstances and their responsiveness to our outbound campaigns. Leveraging the cutting-edge in machine learning techniques and high-speed data processing hardware, we were able to prioritize our dialing efforts to focus on willing-payer patients and qualify indigent patients for charity, while minimizing resources spent on unreceptive, non-paying patients. By correctly identifying paying patients vs. unreceptive patients, we were able to maintain revenue, reduce current operating costs and obviate near-term expenses to scale infrastructure.