About
Experienced Product Manager with a demonstrated history of working in the hospital &…
Activity
1K followers
-
Amjad Azizi posted thisMost lab automation conversations start in the wrong place. They start with what the technology can do. They should start with what the specific operation costs today, in labor, error-related rework, TAT delays, and staff hours spent on tasks that don't require their expertise. The pattern is familiar: impressive demos, carefully selected reference sites, and ROI projections built on best-case assumptions. Then implementation reality arrives. A peer-reviewed economic analysis of total lab automation at a tertiary care hospital found mean TAT improvement of 6.1%, and 13.3% at the 99th percentile, where clinical impact matters most. Payback period, counting only labor cost reduction, was 4.75 years. That's a real, defensible return. But it required rigorous baseline measurement, conservative modeling, and honest accounting of transition costs to arrive at that number. In 2026, with tighter capital budgets and increased scrutiny on every major IT investment, that discipline is no longer optional. Labs that can't articulate the true baseline cost of a process, with their own real data, are in a poor position to justify automating it. Know what you're automating. Know what it costs today. Build the case from your own evidence, not from someone else's projections. #HealthcareIT #Informatics #HealthcareLeadership #DataStrategy #ClinicalDecisionSupport
-
Amjad Azizi posted thisPatients are showing up to clinical appointments having already "consulted" an AI about their lab results. Sometimes those interpretations are wrong. A growing number of patients are using consumer AI tools to interpret their CBC, metabolic panels, and lipid panels — often before speaking with a clinician. The trend is picking up speed. The problem isn't patient curiosity. Patients wanting to understand their own data is reasonable. The problem is that these tools are largely unvalidated for clinical use. According to a Dark Daily analysis from April 2026, there is currently no standardized framework to benchmark accuracy at scale, and early evidence shows these tools can misinterpret biomarkers, miss critical context, and generate recommendations that sound authoritative but aren't. This creates a concrete downstream problem. Patients arrive at appointments pre-informed — or pre-misinformed — and the clinical conversation starts with correcting the AI's interpretation rather than discussing actual results. In my experience, the laboratory has always been the quietest voice in the patient care conversation. We generate the data. The interpretation happens elsewhere. AI-driven consumer tools are now expanding that gap in ways the field wasn't prepared for. Clinical labs have an opening here. Cleaner patient-facing reports, better plain-language result context, and genuine engagement with the developers building these interpretation tools — these are within reach. The profession should be shaping the standard, not reacting to it. The validation gap won't close itself. #HealthcareIT #Informatics #DigitalHealth #ClinicalDecisionSupport #DataStrategy
-
Amjad Azizi posted thisThe FDA's final guidance on clinical decision support software carries one clear message: AI that can't explain itself doesn't qualify as decision support. In March 2026, the FDA finalized its CDS software guidance. The core principle: for AI-enabled CDS to be exempt from medical device regulation, clinicians must be able to independently review the basis for any recommendation the tool makes. Not just that it's accurate — that a clinician can see why. That distinction matters enormously. I've worked in environments where an AI output became the answer — not a reference, not a starting point — the answer. When something goes wrong there, accountability becomes a conversation about who pressed approve, not who understood the underlying logic. That's a governance failure built into the deployment decision itself. The guidance also resolved a longstanding tension: developers no longer need to engineer artificial optionality into tools when clinical evidence points clearly to a single path. A single well-reasoned recommendation with transparent logic now satisfies the standard. That's the right call — forcing a list of choices where one is clinically obvious was never good science. For clinical IT leaders: any AI tool in your environment that cannot explain its reasoning should face a much higher bar of local validation before it touches live workflows. Transparency in clinical AI isn't a feature you add later. It's a prerequisite. (Source in first comment.)
-
Amjad Azizi posted thisAnyone can prompt an AI tool. Not everyone can answer for what it decides. There's a distinction I've been thinking about more as AI moves deeper into clinical operations: the difference between being a user of AI and being accountable for it. Being a user means you interact with the output. You prompt it, you evaluate the suggestion, you move on. That's fine for low-stakes tasks. But healthcare isn't primarily a low-stakes environment. Being accountable means you can answer the harder questions. What population was this model trained on? How was it validated for your specific workflow? Who monitors for drift? What's the fallback when performance degrades in ways the tool doesn't announce? In my experience, the most dangerous moment in any AI deployment isn't the launch. It's six months later, when the initial enthusiasm has settled, the IT team is onto the next initiative, and the tool is quietly drifting from its original accuracy baseline. Healthcare IT leaders who understand this dynamic are building governance infrastructure alongside capability, monitoring schedules, defined ownership, regular revalidation checkpoints. Not as bureaucracy. As professional responsibility. The FDA has signaled this direction in its updated digital health framework: post-market monitoring is moving to the center of how AI tools are regulated, not treated as an afterthought. The standard I'd apply: if an AI system in your environment made a consequential error today, could you walk your CMO through exactly who was responsible for its ongoing performance? If not, that's the gap to close first. #HealthcareIT #HealthcareLeadership #Informatics #DataStrategy
-
Amjad Azizi posted thisEvery lab I've encountered calls itself automated. Many still have staff functioning as human middleware, toggling between disconnected systems and manually re-entering data. There's a meaningful difference between having automation and having automated workflows. The analytical phase is well-covered in most modern labs. Instruments interface, auto-verification runs, barcodes track specimens end to end. The gaps live in the connective tissue: add-on requests arriving by phone or fax, billing reconciled manually, result review still consuming the majority of staff time even in otherwise advanced facilities. I've seen labs where the LIS doesn't fully integrate with billing, where middleware between instruments and the LIS creates its own data silo, and where downstream clinical systems receive a flat file at best. Each gap becomes a manual workaround. Each workaround becomes an unofficial FTE. With staffing pressures unlikely to ease, those workarounds are not a sustainable strategy. Labs facing rising test volumes without proportional staffing growth need to close integration gaps, not add more people to bridge them! The question worth asking isn't whether your lab has automation. It's where that automation stops, and who is covering the handoff. Real automation is a workflow problem. The hardware just makes it visible. #HealthcareIT #Informatics #DataStrategy #HealthcareLeadership
-
Amjad Azizi posted thisIf you've spent time in a clinical lab, you already understand the most important concept in AI governance. In laboratory medicine, we validate before we deploy. We define analytical performance thresholds. We establish reference intervals in the target population, not a reference population from a published paper. We run proficiency testing continuously, not just at go-live. When something drifts, we investigate. We document everything. That discipline is exactly what AI deployment in health systems requires. And most healthcare IT teams don't have it, not because they're careless, but because they're borrowing frameworks from software engineering rather than from the clinical sciences. A software QA mindset asks: does it work in our testing environment? A lab validation mindset asks: does it perform consistently across my patient population, my workflows, and my operators, and do I have evidence to prove it? I've seen well-resourced health systems deploy AI tools across multiple clinical domains without a defined performance monitoring plan. No baseline. No population-stratified analysis. No defined threshold for clinical review. The tool goes live, and the assumption is that ongoing performance mirrors validation performance. That assumption fails in the lab. It fails in AI, too. Healthcare IT leaders don't need to reinvent governance frameworks. One already exists in the clinical sciences. It's been refined over decades by lab professionals, codified by CLSI and ADLM, and field-tested across thousands of high-stakes diagnostic decisions. The methodology is there. The question is whether we choose to use it. #HealthcareIT #Informatics #HealthcareLeadership #DataStrategy #ClinicalDecisionSupport
-
Amjad Azizi posted thisYour AI vendor can update their algorithm without seeking new FDA approval. That is not a problem — unless your health system has not built the processes to detect when performance changes. As part of its January 2026 guidance updates, the FDA formalized the Predetermined Change Control Plan, a framework that allows AI developers to make planned algorithm updates without resubmitting for approval, provided those changes were documented in advance and remain within pre-defined performance boundaries. The logic is sound. Requiring re-approval for every improvement to a machine learning model would stifle responsible iteration. The PCCP approach allows AI tools to evolve in a controlled, documented way. But it creates a specific and underappreciated responsibility for health systems: you can no longer rely on FDA clearance events to signal when something has changed. Real-world performance monitoring is now an internal function, not an external checkpoint. In practice, that means healthcare IT teams need performance tracking embedded in their AI governance infrastructure — not just at deployment, but continuously. Segment performance across patient populations. Track for drift over time. Define thresholds that trigger clinical review. Establish clear accountability for when outputs shift. The teams I have seen navigate this well treat AI governance the way they treat laboratory quality management: documented, continuous, and clinician-informed. That parallel is not accidental. The stakes are the same. The FDA's adaptive oversight model is the right direction for a field that moves as fast as clinical AI. But it works only if health systems build the internal infrastructure to complement it. That gap is wider than most leaders currently acknowledge. #HealthcareIT #DigitalHealth #Informatics #ClinicalDecisionSupport #HealthcareLeadership
-
Amjad Azizi posted thisThe leading professional organization in laboratory medicine just told Congress it needs to update a 30-year-old law to cover AI. They're not wrong. In February, ADLM released a position statement calling on Congress and federal agencies to modernize the Clinical Laboratory Improvement Amendments to explicitly include AI systems within their regulatory scope. The concern at the center of this ask deserves more attention than it has received: AI models trained on historical laboratory data often underrepresent racial, ethnic, and lower-income populations. When those models are deployed in lab settings, they can systematically underestimate risk in the patients who already face the greatest health disparities. ADLM's recommendations are specific: bring AI into the CLIA framework, partner with professional societies to develop federal validation and verification standards for lab AI tools, and standardize data reporting across health systems to enable independent algorithm verification. These are not radical demands. They are minimum conditions for responsible deployment. In my experience, the hardest part of implementing AI in clinical laboratory environments is not the technology. It is the absence of shared standards for what validation looks like, how to monitor for performance drift, and who is accountable when an AI tool performs differently across patient populations. ADLM named that gap clearly. The question is whether federal agencies will act before deployment moves faster than governance can keep up. Regulatory frameworks designed before machine learning existed cannot adequately govern AI-enabled diagnostics. Healthcare IT leaders should be paying attention, and contributing to this conversation wherever they can. #HealthcareIT #Informatics #DigitalHealth #HealthcareLeadership #DataStrategy
-
Amjad Azizi posted thisThe evidence on clinical AI is maturing. That means we finally have enough data to stop asking "does AI work in healthcare?" and start asking "where does it work, and why?" A report released in early 2026 by the ARISE network, drawing on contributions from researchers at Stanford, Harvard, and affiliated health systems, reviewed a year's worth of influential clinical AI research with exactly that question in mind. The findings are worth reading carefully. AI delivers the clearest, most consistent results in prediction tasks: flagging deterioration risk in hospitalized patients, generating risk scores beyond what a clinician can manually calculate, identifying early warning signals in complex data streams. These use cases share a common trait, they extend human capacity rather than replace human judgment. The same report found AI performs well as an optional second opinion in radiology and primary care. The key word is "optional." The models that showed the strongest results were integrated so that clinicians reviewed AI output rather than deferred to it. Where the evidence is weaker: patient-facing AI is expanding rapidly, but most research on it measures engagement, not health outcomes. That is a meaningful gap. Any healthcare IT leader investing deeply in that category without a defined outcomes framework is building on weak ground. In laboratory informatics, the pattern holds. Predictive tools, risk stratification, reflex logic, critical value flagging, fit the profile of where AI earns its return. Autonomous, unreviewed decision-making does not. Pick your use cases based on evidence. Not on what is generating buzz. #HealthcareIT #Informatics #DigitalHealth #ClinicalDecisionSupport #DataStrategy
Amjad Azizi commented on a post
1w
Congratulations! 🎉
-
Amjad Azizi reacted on thisAmjad Azizi reacted on thisAfter 15 incredible years at #Abbott, filled with invaluable learnings, unforgettable experiences, and memories I will always cherish, I am excited to begin a new chapter in my professional journey. I am grateful to have joined PubliCare GmbH as Managing Director. I am inspired by #PubliCare’s mission and the opportunity to make a meaningful impact in HomeCare—helping improve patients’ lives every day. Here’s to new beginnings!
-
Amjad Azizi reacted on thisAmjad Azizi reacted on thisI am truly honored and thrilled to receive the 2026 Association for Diagnostics and Lab Medicine (ADLM) Data Science & Informatics Division Achievement Award. I am deeply grateful to the ADLM community and DSI board committee for this recognition, and to the many colleagues, collaborators, mentors, and friends who have supported and inspired me over the years. This award reflects not only my own work, but also the collaborative spirit and shared commitment of our community to advancing data science, informatics, and innovation in laboratory medicine. I hope to see many of you at the DSI Research Hour during ADLM. I look forward to exchanging ideas, sparking new collaborations, and exploring together how data science and AI can continue to shape the future of laboratory medicine. Thank you again for this meaningful honor and for all of your support!
-
Amjad Azizi reacted on thisAmjad Azizi reacted on thisYesterday I had the opportunity to visit two customer site labs at Northwestern Memorial Hospital and Ann & Robert H. Lurie Children’s Hospital alongside fellow Sysmex interns! Seeing these labs in action, using the instruments that Sysmex produces, highlighted the real-world impact of the work that is done each day. It was incredibly insightful to connect what we’re learning with how it directly supports better outcomes for patients and more convenance for the medical workers. I am so grateful for the lab teams for taking the time to share their expertise and give us a closer look at their everyday operations. It was so great to see the behind the scenes of healthcare.
-
Amjad Azizi reacted on thisAmjad Azizi reacted on this🎙️ NEW SUPERSTAT EPISODE ALERT 🎙️ Leadership. Change. Accreditation. The Future of Laboratory Medicine. This week on SUPERSTAT, we're joined by a very special guest, Nancy Stratton, CEO of COLA, for Part 1 of an important conversation about where our profession is headed and the leaders helping shape that future. Together, we explore: 🔬 The evolving landscape of laboratory medicine 📈 Leadership during times of change 🏛️ The role of accreditation in quality and patient care 🚀 What the future may hold for laboratories, professionals, and healthcare Whether you're a laboratory leader, bench professional, educator, student, or simply passionate about the future of diagnostics, this episode offers valuable insights from one of the profession's most respected voices. The future doesn't just happen. It is built by the decisions we make today. 🎧 Listen now on Spotify, YouTube, and wherever you get your podcasts. What do you think is the biggest challenge—or opportunity—facing laboratory medicine over the next decade? #SUPERSTATPodcast #LaboratoryMedicine #COLA #NancyStratton #Leadership #HealthcareLeadership #ClinicalLaboratory #LabLife #MedLab #LaboratoryProfessional #Diagnostics #QualityMatters #Accreditation #FutureOfHealthcare #HealthcareInnovation #MedicalLaboratoryScience #LabLeaders #Pathology #PatientCare #ScienceLeadership
-
Amjad Azizi reacted on thisAmjad Azizi reacted on thisThe best skill in sales isn’t objection handling. It’s not discovery either. It’s knowing when to press. Watch Messi play and you’ll notice something weird, he barely moves for long stretches…but he’s not coasting. He’s reading the entire game. And when the conditions shift, a defender steps too far up, a midfielder loses shape, he turns on full intensity. He recognizes that moment has disproportionate value. Your pipeline works the same way. Many reps spread their effort evenly across deals. Same energy, same cadence, same follow-up rhythm regardless of what’s actually happening. But the best reps I’ve coached do something different. They’re constantly reading the signals. And when one hits, they press with intensity. …Your champion just told you leadership moved up the evaluation timeline, a competitor got brought into the conversation, the CFO started scrutinizing the ROI data unprompted. Those aren’t just updates, they are windows of opportunity. And they don’t stay open long. The reps who consistently overperform aren’t always working harder across the board. They recognize when a deal has entered a moment where their effort compounds, and they match their intensity to it. Messi doesn’t treat minute 12 like minute 78 with the defense stretched thin. You shouldn’t treat every deal in your pipeline the same either. #alltimeleadingscorer #salesenablement
-
Amjad Azizi liked thisAmjad Azizi liked thisI mapped 16 ML algorithms across 8 decision factors so you never have to guess again. Most people treat algorithm choice as a knowledge problem. It's a constraint problem. The cheatsheet below isn't a menu you order the best item from. It's an elimination filter. Before accuracy ever enters the conversation, three questions quietly kill most of your options: → How much labeled data do you actually have? Small data eliminates neural networks, Transformers, and Gradient Boosting before you start. That's why Naive Bayes and KNN still ship in 2026. They work when you have almost nothing to train on. → Do you have to explain the decision? A loan denial, a fraud flag, a medical triage. The moment a human has to defend the output, you lose Random Forest, Boosting, and deep nets, not because they're worse, but because "the model said so" is not an answer a regulator accepts. Logistic Regression and Decision Trees survive here for one reason: you can read them. → What's your latency and compute budget? Real-time scoring eliminates KNN (no training phase means it does all the work at prediction time) and Hierarchical Clustering (memory-intensive by design). Constraints decide this, not capability. Only after those three filters do the trade-offs on the sheet start to matter: ☑ Linear and Logistic Regression assume a straight-line world. Fast and readable, useless the moment your data bends. ☑ Decision Tree is the most interpretable and the most unstable. Random Forest trades that interpretability away to buy stability. ☑ Gradient Boosting gives you state-of-the-art accuracy and bills you in tuning hours. ☑ Transformers win on long context and lose on everything cheap. Here's the uncomfortable part. The most powerful model is almost never the right one. Capability is the easiest axis to optimize, so it's the one beginners fixate on. Senior engineers optimize the opposite direction: the simplest model that clears the bar, because simple models are cheaper to run, easier to debug, and survive contact with messy production data. That's why the column everyone skips, "When Not to Use," is the one that actually separates the two. Accuracy is the last thing you tune. Constraints are the first thing you check. Learn how to train AI to write exactly like you using the voiceprint here: https://lnkd.in/d3VE4dzm What's the last time you shipped a simpler model than the one you wanted to? ♻️ Repost to help someone stop over-engineering their model. #AI #GenAI #MachineLearning #DataScience #ML
-
Amjad Azizi liked this💜Amjad Azizi liked this🎓 A milestone worth celebrating! In partnership with Eastern Florida State College, Health First launched a two-year program in 2024 to develop future Medical Laboratory Technicians through hands-on experience working as laboratory assistants. We’re proud to celebrate the graduation of our first scholarship cohort! 🎉 In just a few weeks, these graduates will join our laboratory team as full-time Medical Laboratory Technicians, helping strengthen the future of healthcare on Florida's Space Coast. Investing in our workforce today means better care for our community tomorrow. Congratulations to our graduates—we can't wait to welcome you to the team! Interested in learning more about Medical Lab Technician opportunities and our Come Grow with Us program? Visit hf.org/grow Prefer to connect directly? Email us at ComeGrowWithUs@hf.org
-
Amjad Azizi liked thisAmjad Azizi liked this🔬 The future of laboratory medicine isn't reporting. It's forecasting. 🔮 Clinical Lab 2.0 is re-engineering diagnostics from reactive confirmation to proactive prediction. #AI models trained on single-site datasets fail in the global landscape of #laboratorymedicine. 📉 📝 Our new manuscript introduces the CL2.0 Lab-Initiated Care Model: a governance-backed framework for a global, de-identified laboratory data repository that enables truly generalizable AI. 🌎 👀 Inside: a nine-step business canvas 📋, real-world vignettes from cancer cluster detection to OGM validation 📊, and a privacy-first architecture built for cross-border compliance 👷♂️. ✍️ Authors include Ulysses G. J. Balis, MD, FCAP, FASCP, Fellow AIMBE from the University of Michigan, Amjad Azizi of Sysmex America, Inc., and Dr. Myra Wilkerson of Geisinger. Dr. Balis and Amjad are the Co-Chairs of the CL2.0 Pathology Informatics-AI Committee, and Dr. Wilkerson serves as the Committee's PSFF Board Liaison. 👏 🔮 The future of #diagnosticstewardship starts with shared data. 🔗 Read the full manuscript: https://lnkd.in/gTAta-mH 📻 Stay tuned for our next publication, which will be the second from our Education Committee, written in conjunction with #DCLS professionals and advocates about their specific role in the clinical care continuum. 📧 To join our database and receive our once-weekly emails directly in your inbox, fill out our contact form here: https://lnkd.in/ehDWjBXD or email us at: info@cl2lab.org. #healthequity #wholepersoncare #crucialconversation #reactiveconfirmationtoproactiveprediction
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
Licenses & Certifications
Projects
-
POCT connectivity projects at NGHA-KSA, SKMC-Abu Dhabi, TAWAM Hospital-AlAIN, HMC-QATAR
-
Introduction of the POCT concept, installations of the various POCT devices and connecting via middleware to the main host.
Other creators -
Other similar profiles
Explore more posts
-
Havcom Technologies
Validated. Compliant. Trusted. Why It Matters for Safer Healthcare In healthcare, every decision relies on accuracy from diagnostics to data interpretation to the systems that support patient care. That’s why validated and compliant systems aren’t just technical requirements. They are patient safety requirements. At Havcom Technologies, we guide organisations in strengthening the backbone of their healthcare operations through: - Risk-based validation approaches - Robust compliance frameworks - Quality-driven processes that support safe and reliable outcomes Because in healthcare: - A validated diagnostic system means fewer errors. - A compliant data environment means stronger privacy and integrity. - A reliable operational ecosystem means safer experiences for patients and providers. Effective healthcare doesn’t start at the bedside it starts behind the scenes, with systems that are trusted, transparent, and built on accountability. Safe care is powered by compliant care. And compliant care starts with doing validation the right way. #HAVCOM #HealthcareQuality #ValidatedSystems #ComplianceMatters #PatientSafety #HealthcareConsulting #DiagnosticsQuality #MedicalDataIntegrity
1 Comment
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content