Harry Goldberg (@HarryCGoldberg) header
Harry Goldberg (@HarryCGoldberg) profile picture

Product @neonhealth | Harry's HealthTech Hang

San Francisco, CA
Joined August 2016
HIMSS just told CMS what a lot of us building this have been thinking out loud: the 2027 deadline is running ahead of the runway. That's my read too, as someone who has to turn these timelines into something access teams can actually run. The new rule (CMS-0062-P) wants drug prior auth to go electronic and run on new FHIR-based APIs by October 2027. Genuinely good goal! HIMSS just flagged three things that aren't ready yet: the APIs haven't been tested at scale, there's no CMS payer directory to look a payer up in the ePA flow, and payers are each on a different version of the implementation guide. If you're on an access team staring at this same timeline, come find me. I'd genuinely love to compare notes on what "ready" really takes.
1
1
70
Here's a scene anyone in patient services will recognize. Your manager emails: "Your agents sound dismissive on Brand X this quarter. We need a solution by end of week." What we've seen play out next is always the same. Two QA analysts, three days, audio pulled by hand, a hedged response, and a lot of stress. We built the Neon Quality Engine to shrink that to the same business day. Every call on Brand X is already scored, on protocol and on sentiment, every score linked to its transcript. You filter by brand, time, and agent in a few clicks. You open the calls that make the case. You send back an evidence-linked answer from 100% of calls, the day your manager asked. That's rapid protocol monitoring, not a monthly post-mortem. Be right. Be nice. On every call. Blog post here πŸ‘‰οΈ neonhealth.com/blog/ai-quali… Built for hubs, specialty pharmacies, and manufacturer patient services teams. If this is your world, DM me. πŸ™ #PatientServices #SpecialtyPharmacy #HubServices #HealthcareAI
1
43
Detecting "bad words" in patient support calls is NOT quality assurance. It's sentiment scoring with a QA sticker on it. You've probably heard of these vendors. They listen to PSP/PAP calls, flags negative-sentiment patterns, and sends the hub a review queue (read: expensive wild goose chases). Picture a patient saying: "Pretty crappy day. My flight was delayed two hours because of the wind." That call gets flagged. The manufacturer pings the hub. The QA team pulls the audio looking for a complaint that isn't relevant. A word-list can't tell you whether the agent asked the right questions or pushed back when they should have. A word-list can’t statistically measure adherence to a standard operating procedure. At Neon Health, we have both sentiment AND adherence. Happy to compare notes on what good QA actually looks like. #PatientSupport #HubServices #SpecialtyPharmacy #QualityAssurance
51
Pharma channel consolidation has blasted past anything our integration calendar was sized for. We build patient access workflows assuming a partner stack that's stable enough to hand a benefit verification off to a prior auth, off to a financial assistance enrollment, and the partner on the other end of that handoff turns out to be owned by someone different than our roadmap said six months ago. That's why Drug Channels' 2026 vertical-integration map is worth reading carefully. Adam Fein's annual update shows the channel running in two directions at once: 2025 unwinds alongside fresh integration moves, in the same window. Centene outsourced PBM operations to Express Scripts in 2024 and divested Magellan Rx, PANTHERx Rare Pharmacy, and Magellan Specialty Health. In 2025, Cigna Healthcare sold Evernorth Care Group's 18 primary care clinics to HonorHealth. And in the same window, Evernorth invested $3.5 billion for an undisclosed stake in Shields Health Solutions, and UnitedHealth Group acquired CPS Solutions for hospital specialty pharmacy services. The divestiture column and the integration column are running on the same calendar. The architectural read is that the channel was rebuilt last year while a lot of patient services teams were actively integrating against the version of it that existed in 2023. The partner stack you stood up benefit-verification logic against in 2024 is partly owned by different parents in 2026, and the integration surface moved with the ownership. In life sciences, the system you integrate with isn't the system you signed up to integrate with, and at Neon | AI-powered patient access, we have some fun solutions to deal with this ever-changing ecosystem.
2
1
46
The prescription gets written. Then the patient hits the wall. Insurance hurdles, financial assistance enrollment, prior auth back-and-forth, payer appeals, the right form to the right fax line, the call to the patient explaining what's actually happening. The gap between Rx written and therapy started is where adherence is won or lost. 4 in 10 patients fall off entirely! Closing that gap is the actual job: the work optimizes for the patient who would otherwise never get the first dose, while cost reduction shows up downstream in fewer FTE-hours and faster cycles. That distinction matters when you're scoping automation for patient services. Teams getting it right build toward access first, even when the slide deck speaks in cost-out terms. The patients meet the medicine sooner, and everything else follows.
39
I think that the specialty pharmacy our hub partners are working with five years from now will be a completely different operator than the one they're working with today. That's what we keep hearing from pharmacy ops leads, and Adam Fein's Drug Channels piece spells it out πŸ‘€ Drug Channels Institute, an HMP Global Company counted almost 2,000 accredited specialty pharmacies in the U.S. at the end of 2025, and independents are down to 29% of those locations from 59% in 2015. Hospitals and health systems went from 15% of the count in 2017 to 28% in 2025, with another 16% owned by other healthcare providers. So, the top 3 specialty pharmacies hold roughly two-thirds of specialty drug Rx revenues, and all three are vertically integrated with a PBM. The accredited landscape looks more diverse than ever on the surface, but the economics are moving in the opposite direction! Adam Fein and the DCI team are doing the kind of work most operators don't have time to do themselves. Worth a follow if you're tracking specialty pharmacy, they have a great sense of things πŸ™
1
96
When I walk into a manufacturer's patient services planning session, the strategy consulting firm's already there, and that's healthy. I’m saying this as a former consultant and current VP of Product! They've already mapped the workflow, the benchmarks, and the recommendation by the time we sit down, so the build sits inside their roadmap, not next to it. They scoped what to build and why, and we focus on executing what they've delivered. The two roles aren't competing for the same dollar. The cleanest engagements I've sat in this year had one of the strategy consulting firms I respected. The manufacturer's patient services lead already knew exactly what role they'd handed to whom going into the kickoff. That's the healthy version! If you're scoping a patient services build with a strategy firm, drop me a note.
38
Cleveland Clinic, Sanford Health, Providence, CommonWell. The Centers for Medicare & Medicaid Services Electronic Prior Authorization Acceleration recently added 29 named partners to a list that until recently was mostly health plans (Aetna, Cigna, Humana, UnitedHealthcare among the nine health plans that signed the pledge last June). What changed last week is that the implementation cast finally includes the providers, the EHR vendors, and the network operators who have to actually plug FHIR endpoints into intake workflows. What we keep hearing from hub services and specialty pharmacy operators is that the January 2027 deadline isn't the real planning question anymore. The real question is which of these named partners will have a working API endpoint in a sandbox you can integrate against six months from now. That's the gap between a pledge and a workflow. The teams that look ready for 2027 aren't the ones still tracking the rule; they're the ones lining up integration partners by name, mapping which provider system feeds which payer endpoint, and writing the fallback paths for the EHRs and networks that won't be ready when the deadline lands. If you're running hub intake or PA operations and already triangulating between these named partners for your integration roadmap, would love to compare notes. Shoot me a text, always learning from how teams are sizing this in practice.
1
72
Every specialty pharmacy AI roadmap starts with voice agents right now. Last week a COO walked me through theirs, and we landed on a different first step. It's been working! Phase one: we run our QA AI on every human call. The team gets visibility into what good calls look like, the scoring engine gets calibrated against real client work instructions, and there's no risk because the humans are still doing the work. In parallel, we layer in workflow automation on prior auth. The time-to-therapy ROI shows up fast there, and every automated PA is one less swivel-chair hour for the team. Voice agents come last. By the time we layer them in, the QA scoring is already running, the work instructions are documented and refreshed as the team learns, and there's pattern data on what 'good' means for that brand. What we've found is each phase pays for the next: the QA scoring funds the workflow automation, and the workflow data trains the voice agents. None of it requires a leap of faith! 🀝
31
Two GLP-1 competitors. One self-pay platform. GoodRx is now offering both Novo's Wegovy HD ($399/month) and Lilly's oral GLP-1 (Foundayo) on the same self-pay platform. It’s the aggregator pattern landing in real time. The pricing tells the velocity story too: $399/month, $798 for two months, $1,197 for three. FDA approved the 7.2-mg dose on March 19, and GoodRx has it on shelf less than a month later. What's striking is how fast the parallel rail is building out. Manufacturer-direct (LillyDirect, Novo's own subscription) and platform-aggregator (GoodRx) are both live now, competing for the same patient. How long before self-pay platforms like this start competing with insurance for routine specialty access, not just GLP-1s?
138
What a week at Asembia's AXS26 Summit 🫨 Dozens of meetings. 4 days.The Neon | AI-powered patient access team was out in force. Some of the smartest patient services leaders in the industry, all in one Wynn Resorts tower. A few patterns I kept hearing from manufacturers, hubs, and SPs: βœ… Hub differentiation is collapsing βœ… AI must past the Turing test βœ… "Sentiment β‰  accuracy" finally landing on the manufacturer side Huge thanks to everyone who carved out time. To the folks I missed -- DM me to catch up! πŸ™ to Stedman Hood, Kyle Woodard , and Chris Dimos for an absurd 4-day push. #AXS26 #PatientServices #SpecialtyPharmacy #HealthTech
36
Hospitals quietly became the fastest-growing specialty pharmacy operators. 106 locations in 2017, 553 today. That's 28% of accredited specialty pharmacies. Adam Fein's new Drug Channels rankings put a number on the shift that operators have been feeling for a few years. What's easy to miss is that the rest of the accredited layer barely moved. ~1,900 locations grew only 3% YoY, while health systems did 5x in 8 years against that flat backdrop. That's one segment doubling down on the patient-access stack while the rest of the market sits flat! Most of the specialty-dispensing conversations I've had in the last six months start with a hospital ops lead, not a pharma manufacturer. Would love to hear from folks closer to this than me πŸ™ What's powering the build-out on your end? Margin capture, patient-access strategy, both, or did I miss something?
28
DiMe just pulled together Welldoc, Wheel, PHIL, Inc. in one room to write the rulebook for direct-to-patient drug models. Companies that have been building parallel rails for five years, now in the same working group. The market has been sprinting (LillyDirect in early 2024, TrumpRx earlier this year, J&J last month) while the operating standards lagged. Anand Iyer at Welldoc put it cleanly: the goal is to "tame the Wild West before it becomes the Wild West." What we keep seeing across patient access is that the hardest part isn't the prescribe step. It's everything around it: oversight, transparency, the handoff between the platform that prescribes and the team that supports the patient on therapy.
40
Back from Becker's Healthcare Spring CPO Summit in Chicago. πŸ™Œ One theme kept surfacing across the exhibit hall and Mark Cuban's keynote: pharmacy teams are covering more volume, more programs, and more compliance, without proportional investment in the people or tech that actually helps. Huge thanks to my booth bro, Stedman Hood, for two days of great conversations with chief pharmacy officers from our nation's top hospitals. We held down the Big Red Fort! What we're building at Neon | AI-powered patient access: AI agents that plug into specialty pharmacy workflows, not around them (read: off hold, out of data entry, onto the clinical work that actually requires a pharmacist). If you're a specialty pharmacy or patient services leader thinking about this, shoot me a DM. β˜• #BeckersPharmacySummit #SpecialtyPharmacy #PatientServices #PharmacyOps
42
I am, apparently, going to see Charlie Sheen speak at a specialty pharmacy summit next week. Still processing that sentence. (read: this is not where I expected Charlie Sheen to turn up in 2026.) The charitable read: Resilience. Recovery. Meeting people at their most vulnerable. Rebuilding. That is the patient access job. If the fireside actually gets to the "meet people where they are" truth, there's something real in it. Benefit of the doubt, cautiously extended. @stedmanblake and I are at Asembia's AXS26 Summit all week. Come find our booth #1838 or suite -- coffees and contails on us. Text, DM, or comment if you'll be there! πŸ€— #Asembia26 #AXS26 #SpecialtyPharmacy #PatientServices #MarketAccess
1
2
49
76% of biopharma manufacturers are outsourcing at least some hub functions BUT the hub model is shifting, and the numbers back up what we've been hearing everywhere. Cardinal Health's latest Sonexus pulse study found that confidence in bringing capabilities in-house jumped 17 points in just two years. Nearly half have looked at vendor consolidation recently. What we've been seeing on the ground now has hard data behind it: the monolithic single-vendor hub is giving way to something more modular. Best-of-breed tools for specific functions, clear SLAs, shared KPIs across partners. 𝗧𝗡𝗲 π—”π—œ 𝗽𝗢𝗲𝗰𝗲 97% of respondents are in some stage of AI implementation. 65% plan to adopt within the next 12 months! We're talking about tools that cut SOP and reference data lookups from 15-20 minutes down to seconds. The organizations getting real traction are the ones asking the hard question early: where does our core competency end and our partner's begin? That answer is different for every program, and it's changing fast. This is exactly the kind of conversation we're fired up to have at Asembia! If you're going to be there, come find us. We'd love to dig into what this shift looks like up close πŸš€ πŸ‘‡ #PatientAccess #HubServices #LifeSciences #Asembia
38
We built an AI QA system and the first thing we learned: sentiment tells you almost nothing about whether the call was actually handled correctly. Think about it this way. A patient services call has maybe forty steps in its standard operating procedure. The agent needs to verify insurance, confirm the right medication, walk through side effects in the right order, follow up on prior authorization status. Sentiment analysis will tell you the agent was polite. It won't tell you they skipped step 22. So we built something that monitors adherence to the actual SOP, step by step, at scale. In highly regulated, technically complex calls in life sciences, that's the name of the game.
26
Nobody's writing love letters to payment infrastructure. But try running patient access without it. Virtue VC put together a solid look at the B2B payment layer in healthcare. The companies building here don't get the headlines, but they're solving real plumbing problems that affect whether patients actually get served. The stat that stuck out to me was that pre-pay integrity went from nearly zero a decade ago to 40% today, heading past 60%. The whole layer is being replaced while it's still running. Worth your time ⏰ #HealthTech #PatientAccess
1
109
72% decreased risk of discontinuation. That's not marginal. πŸ‘€ AbbVie's patient support program studied over 10,000 patients on adalimumab. Nurse-led care coaching calls. The result: patients were also 48% more likely to stay adherent. And it's not one study. Patients enrolled in a Crohn's disease support program were 17% more likely to start therapy and achieve remission. An integrated oral chemo program hit a medication possession ratio of 0.92. The evidence keeps stacking up. When you actually support patients through their therapy, they stay on it. πŸ‘πŸ‘πŸ‘
50
Lots of enthusiasm for clinical AI. Fewer receipts. The ARISE State of Clinical AI survey shows the gap between pilot excitement and production deployment is still wide. Familiar pattern for anyone who's been in life sciences tech for a minute. It's the most grounded look at where AI actually helps in practice and where the evidence still has gaps. The teams closing that gap are worth watching. #ClinicalAI #HealthTech
52