Serif Health’s Post

I have a lot of thoughts about all of the news this week. Folks have done a great job explaining it (hat tip to Brendan Keeler for his excellent explanation). So my thoughts will be brief. 1) Misusing purposes of use or setting up shady practices which just barely meet the definitions if you squint real hard at them is NEVER ok. We owe so much more to patients than that. Be better. 2) It is especially reprehensible if data is being used to allow law firms to search for class action lawsuits. My private healthcare data should never be used to allow ambulance chasing lawyers to find cases. If I want to file a case that's for me to decide, accessing my personal information to decide that for me, NEVER ok. Again, be better. 3) Many of us have been pushing hard (some of us since 2017 and the first draft of TEFCA) to expand the purposes of use within HINs so that folks have legal and appropriate avenues to access data bringing transparency to data sharing. This gives patients more choices not less and ensures an even playing field. The slow walking being done on payment and ops and broadcast/identity based individual access (and let's be fair their are financial motivations by some of the orgs to slow this down) is not acceptable. We have to move to expanded permitted purposes that take into account who the requestors are and how they will be using data and ensure both of those are clear to data holders. 4) Trust but audit/verify is key to trust at a national scale. If you are an EHR vendor you must have systems in place that can identify abnormalities and patterns in data requests. HIEs have had this technology in place for years. HIEs that have very little funding, have systems in place that can identify odd requests and immediately stop them. EHR vendors have no excuse not to have equivalent technology in place. Bottom line: Be better. Don't break an entire system that is vitally important to the health of patients just to pad your pockets. Maybe take a second to think, would I want my health data or my family member's health data used in this way. This is not just data. This is incredibly personal information about people; information we want them to share with their doctors freely, which they will not do if they can't trust it won't be kept private. So be better.

At Serif Health, the beginning of the month is all about ingesting the latest updated machine readable files from each payer. This month, one of our lead engineers Derek Pedersen, is sharing just a few of the intense challenges associated with the scope and complexity of this process and how we continue to automate and improve our monthly ingestion pipeline! Read on if you want to learn a little bit more of the day-to-day challenges of working with the price transparency data!

In emergency care, every second matters. Delayed decisions can mean the difference between life and loss. 𝗥𝗲𝗮𝗹-𝘁𝗶𝗺𝗲 𝗱𝗮𝘁𝗮 𝗮𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀 transforms how healthcare providers respond, act, and save lives. But how can hospitals and emergency responders make split-second decisions confidently using real-time data analytics? Here’s how: 👉 𝗘𝗮𝗿𝗹𝘆 𝗪𝗮𝗿𝗻𝗶𝗻𝗴 𝗦𝘆𝘀𝘁𝗲𝗺𝘀: Real-time analytics processes live patient data to detect critical changes like heart irregularities or oxygen dips. Instant alerts allow healthcare providers to act before it’s too late. 👉 𝗔𝗺𝗯𝘂𝗹𝗮𝗻𝗰𝗲-𝘁𝗼-𝗛𝗼𝘀𝗽𝗶𝘁𝗮𝗹 𝗖𝗼𝗼𝗿𝗱𝗶𝗻𝗮𝘁𝗶𝗼𝗻: Data from ambulances, including vitals and diagnostics, can be shared live with ER teams. This ensures hospitals are prepared for the patient’s arrival, reducing treatment delays. 👉 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗶𝗻𝗴 𝗘𝗺𝗲𝗿𝗴𝗲𝗻𝗰𝘆 𝗥𝗼𝗼𝗺 𝗙𝗹𝗼𝘄: Emergency rooms often face overcrowding. Data analytics helps forecast patient influx, allocate resources, and reduce bottlenecks in emergency departments. The result? Shorter wait times and better patient outcomes. 👉 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗻𝗴 𝗖𝗿𝗶𝘁𝗶𝗰𝗮𝗹 𝗘𝘃𝗲𝗻𝘁𝘀: Real-time analysis of patient data can identify early warning signs of complications like sepsis, heart failure, or stroke. Early alerts allow for proactive intervention when it matters most. 💡 𝗪𝗵𝘆 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀: Emergency care doesn’t wait, and neither can data. Real-time analytics gives healthcare professionals the tools to act swiftly, efficiently, and accurately when lives are on the line. Learn more at dataengite.com #HealthcareInnovation #RealTimeAnalytics #DataEngineering #EmergencyCare #AIinHealthcare #DataDrivenCare #Dataengite #Dataanalytics

Joining NPI to Overture with Placekey, https://lnkd.in/ed-suKgK This was a fun one that provided some insight into both the NPI and Overture datasets. In joining these, I found the record in Overture for a Rite Aid a few blocks from my apt with GERS ID of 08f2a10720ad9386038cefa3fb288f81 and NPI id of 1255749818. In NPI, the record is listed as 'ACTIVE' and in Overture it comes from a 'msft' (Microsoft) source with confidence of '0.77' - pretty high on 0-1. I had never seen it before, so I did some digging and could not find much on it until ... 2007 came calling and I found that MapQuest was the only mapping aggregator to note its existence and having it permanently closed. So, really the takeaway is that reality models are challenging. The Colab notebook is linked in the blog post if you want to run it yourself. Like any generalized matching problem, we are constantly iterating so if you have any thoughts or feedback always open to it! Stats on join: Enriching the NPI data with Overture lead to: - 173,646 centroids on Entity Code One records - 527,652 centroids on Entity Code Two records - 137,279 websites attributed to Entity Code One records - 476,671 websites attributed to Entity Code Two records - 63,300 socials attributed to Entity Code One records - 413,882 socials attributed to Entity Code Two records

Hospital Price Transparency: Get Resources to Help You Comply Starting July 1, 2024, hospitals must conform to a CMS template layout and data specifications for making public their standard charge information in a comprehensive machine-readable file (MRF). Starting January 1, 2025, you’re also required to encode additional data elements. We have resources to help you meet these new requirements: Data Dictionary GitHub Repository: https://lnkd.in/dgTKTW4Y Required CMS template layouts Data dictionary Examples of how to encode standard charges in the MRF Q&A discussion board Tools webpage: https://lnkd.in/dqEVFfip Online validator Command-line interface validator TXT file generator MRF naming convention tool Resources webpage: https://lnkd.in/d7TbkxpZ Final rules FAQs Guides Webinar materials Hospital Price Transparency regulations require each hospital operating in the U.S. to publish a comprehensive MRF with the standard charges for all items and services they provide. More Information: Register for the October 21 webinar on meeting the upcoming January 2025 requirements Email questions to PriceTransparencyHospitalCharges@cms.hhs.gov

Time is running out for payers to comply with the #CMS Interoperability and Prior Authorization Final Rule. Waiting until 2025 or 2026 may mean you’re not ready to exchange real data in production by the January 1, 2027 deadline. Availity’s latest white paper reveals the critical steps you should take now to ensure compliance. Learn from our inaugural Payer-to-Payer Data Exchange cohort and start your journey today. https://okt.to/F1tUMk

💡 Ready to tackle the challenge of low practitioner response rates? 💪 Our latest article has you covered with three effective strategies to get those responses rolling in. Say goodbye to missed deadlines and hello to a more efficient process. Check it out now and level up your data accuracy game! #QuestForSuccess

No matter where you are on your Enterprise Clinical Data Management (eCDM™) journey, we have the guidance for you. Visit the new Q-Centrix website, select your biggest eCDM obstacle, and we'll provide you with information about our process, experts, technology, or how to request a consultation. All at the stroke of your finger: https://bit.ly/3CPTHLq #ClinicalDataManagement #EnterpriseData #DataAnalytics

📈 Struggling to get practitioner responses for data accuracy? 😫 We've been there! But fear not, we've got three proven strategies to boost those response rates. Say goodbye to missed deadlines and hello to efficiency! Check it out now. #QuestForSuccess

What Is Your Historic Data Worth? We often measure value with tangible assets or expected cash flows. There is often a measure of value right in your hands that you forget to pay attention. For healthcare providers, your visit history has significant worth. A few years ago, I championed a project where we did a detailed analysis of our patient visits. Our system was solid in that each appointment that was booked finished with a status. Those status labels were COMPLETED, RESCHEDULED, CANCELLED or NO SHOW. There may have been another status or two, but you get the idea. We required a status of an appointment within two hours of the scheduled appointment time and if none was entered, the appointment hit an exception report. We also cross-referenced any COMPLETED appointments to see that charges occurred on that date of service and for any charges on a date of service, it had to have a corresponding status of COMPLETED. My objective in this project was to data mine the likelihood of NO SHOW cases and see if repeatable and predictable patterns occurred. By analyzing several years of data across 120 locations, we began to see clusters of NO SHOW appointments for certain times. By taking advantage of the likelihood of the NO SHOW we found that a large portion of our cases occurred in certain time of day slots. With several years of data in our hands, we were able to identify opportunities to schedule an additional patient based upon the likelihood of a NO SHOW. We set the system's initial threshold to be a high likelihood and dropped the threshold over time to find a "sweet spot". The outcome? We had a 5% lift in revenues. For a location that did 75 visits per day, we found an additional 3 to 4 visit opportunities per day. Yes, we did occasionally have everyone show up but it was never a patient satisfaction issue. Instead of an empty spot, we had a patient. Yes, we tracked those instances where we had overbooked and the statistics were self-correcting as time went on. The system became sophisticated enough to point to times on the calendar where a NO SHOW would occur and you could view the likelihood going back 12 months, 24 months or all data combined. The first step in this project was to have reliable data in our history. It does no good to have appointments without a reliable status. Your data has value.