1. Program design
Who is the program for? What does each level of care actually include? Where should human time be spent, and where should the system carry the load?
Diabetes program design since 2005
I have helped shape how diabetes care is delivered between visits.
I design care programs that determine the right intervention for the right patient at the right time — then use real-world data to step care up, step it down, and improve the program itself.
A core design principle
Digital health should not mean digital-only.
A postcard can be an intervention. So can a home A1c kit, a CGM, a text message, a phone call, a family member, a clinician, or an automated workflow. The program should choose the channel most likely to work for this patient now.
Blue Cross Blue Shield of Texas
BellSouth / AT&T
Kaiser Permanente
Texas Medicaid / McKesson
USDA Telemedicine
Driscoll Children’s Hospital
NHS / Salford Royal
Philips
What I help with
Five design problems show up every time.
Most teams do not need more health-tech enthusiasm. They need a clean operating model.
2. Device and vendor selection
CGM, BGM, diagnostics, messaging tools, analytics, and platform vendors should be chosen for fit inside the care model — not because they look impressive in isolation.
3. Governance and compliance
Someone has to define the clinical guardrails, the data provenance, the privacy model, the validation expectations, and the vendor responsibilities.
4. Patient decision logic
The program needs a practical way to determine who should be stepped up, who can be stepped down, and which signals are strong enough to trigger a change.
5. Program optimization
Once the program is running, the data should tell you which features, vendors, interventions, and workflows are helping — and which ones are simply consuming time or money.
Lifecycle diagrams
This is the work in pictures.
A good diabetes program learns in two loops at once: one loop around the patient, and one loop around the program itself.
Patient care lifecycle
This loop answers the daily question: what does this patient need now?
Program learning lifecycle
This loop answers the management question: what should the program itself do differently?
Program stack
This is the architecture I think clients actually need to see and discuss.
Why listen to me
I have worked on this in the field, not just on slides.
These examples matter because they mirror the design questions ACCESS participants and device partners are facing now.
DiabetesHouseCall / BCBSTX
Remote diabetes program design, supporting evidence, actuarial review, and commercial reimbursement for the program as a whole.
BellSouth Active Disease Management
A fully remote program with connected glucose data, mail-in A1c, patient onboarding, copays, claims, medical necessity, physician participation, and care operations.
Kaiser Permanente research
Motivational interviewing and behavior-change principles translated into digital patient engagement and support workflows.
Texas DHHS / McKesson
Medicaid telemonitoring and disease-management design with explicit goals around A1c, self-management, satisfaction, ER use, and hospitalization.
HomeCheck-A1c
Remote outcome measurement and near-term reinforcement so the program could learn faster and patients could see progress sooner.
ACCESS + TEMPO
The need is finally obvious.
ACCESS creates room for outcome-oriented diabetes programs. TEMPO creates room for technology to prove its role inside those programs.
The work sits in the middle: program design, intervention logic, vendor choices, governance, and a data architecture that lets the model keep getting smarter.
If you are building one of these programs, the real question is not whether a device works.
The real question is whether the whole care system can choose the right intervention for the right patient, at the right time, and learn from the result.
Built, tested, operated
This work did not simply anticipate the market. Parts of it became the market.
Over two decades, I have worked across remote monitoring, disease management, behavior change, diagnostics, payer reimbursement, telehealth, real-world evidence, digital clinical development, and patient-facing care systems.
The common thread is not a device, an app, or a single delivery channel. It is the design of a care system that can observe, decide, intervene, measure, and learn.
Source evidence
This work changed care, not just software.
A short reading list for anyone trying to understand what was built, what was tested, and what the results actually showed. These are third-party publications, government records, and original public sources — not résumé claims.
01
Clinical results
The strongest place to start: controlled evidence that the system affected outcomes.
Read this first
Integrating an Automated Diabetes Management System Into the Family Management of Children With Type 1 Diabetes
Forty-eight children were randomized for 12 months. Families using the Diabetech ADMS finished with significantly lower A1c than controls, and more frequent use was associated with greater improvement. It is the clearest published answer to the question: did the system do anything clinically useful?
Why it matters now: the intervention was not just a meter. It combined automated data capture, trend reporting, real-time alerts, and family decision support — the same kind of program-level thinking ACCESS is now making economically relevant.
Read the full open-access article →
02
Behavior, engagement & social support
Evidence that the care model went beyond transmitting glucose values.
Diabetes Social Support Feasibility Pilot Study
Kaiser Permanente Georgia and Diabetech paired adults with type 2 diabetes with a supporter of their choosing, then used wireless glucose transmission and mobile feedback to make that support more timely and useful. Kevin McMahon is listed as a study author.
Why read it: it shows an early attempt to design the people around the patient into the intervention instead of treating adherence as a reminder problem.
Open the original journal issue →The Potential of Cellular Technology to Mediate Social Networks for Support of Chronic Disease Self-Management
Douglas Roblin documents the actual design choices behind the Kaiser/Diabetech intervention: wireless glucose transmission, supporter messaging, trigger logic, implementation tradeoffs, and lessons from participant use.
Why read it: this is less about a gadget and more about designing a chronic-care system around human behavior, self-management, and clinical escalation.
Read the original article →Using Mobile ICT to Enable Social Support in Chronic Care Management
This presentation exposes the practical lessons behind the pilot, including the uncomfortable ones: technology can improve engagement, but integration cost, delivery-system burden, and patient fit determine whether it actually belongs in the care model.
Why read it: the conclusion is surprisingly current — technology should be selected because it improves the care system, not because it exists.
Open the original AHRQ presentation →
03
Early connected-care architecture
Independent reporting and public records showing what Diabetech was actually building.
Help for Diabetics — Diabetech makes disease management simpler
A contemporary profile describes GlucoMON as immediate patient-to-caregiver connectivity, but more importantly documents the broader model: mobile data, an intelligent rules engine, and social networking as three parts of one diabetes-management system.
Why read it: it is third-party evidence from 2007 that the design intent was already disease management and decision support — not simply selling connected meters.
Read the original D CEO profile →GlucoMON — “In Touch When You Can’t Be”
David Mendosa documents the product while it was still new: automated wireless transmission, real-time family alerts, privacy controls, hospital studies, intensive-management protocols, and a nationwide study explicitly intended to support reimbursement.
Why read it: this is one of the best time-stamped records of how early the system moved from connectivity toward clinical intervention and payer evidence.
Read the original article →System and Method for Glucose Monitoring — US20050038680A1
The public record lays out a remote-care architecture spanning glucometers, insulin pumps, activity data, nutrition, remote telemetry, trend analysis, patient-management teams, and communications between the patient and care services.
Why read it: it provides a dated technical record of the architecture years before remote patient monitoring became a mainstream healthcare category.
Open the public patent record →Texas identifies GlucoMON as a wireless mobile patient diabetes-monitoring example
The state health plan describes GlucoMON as a wireless system for automatically transmitting glucose results and delivering configurable alerts by phone, fax, pager, or email — without requiring a computer or technical expertise from the patient.
Why read it: this is government-source recognition of the practical design principle that still matters today: reduce patient burden while moving useful data to the right people quickly.
Open the Texas State Health Plan →Kevin McMahon / Diabetech appears in the official witness record for diabetes self-care legislation
The Texas House Research Organization’s bill analysis lists Kevin McMahon of Diabetech among witnesses supporting legislation addressing individualized health plans and diabetes self-care in schools.
Why read it: it establishes a public policy footprint around practical diabetes self-management at the same time the remote-care work was being developed.
Open the official Texas House analysis →
04
Payer and disease-management context
The operating environment behind the Texas Medicaid / McKesson work.
Texas Medicaid Disease Management — claims-based identification, chronic-care management, and cost reduction
The public program record shows Texas contracting with McKesson Health Solutions to manage high-cost chronic conditions including diabetes, using claims to identify eligible patients and combining education, monitoring, medication support, RNs, and community health workers.
Why read it: it explains the disease-management program into which Diabetech later proposed its telemonitoring subcontract — the payer problem, the operating model, and the economic intent.
Read the government program summary →CMS records McKesson Health Solutions as the participating plan for the Texas Medicaid Enhanced Care Program
CMS documents diabetes as one of the included chronic conditions and describes data analysis, quality improvement, program evaluation, and program modification as explicit uses of collected data.
Why read it: this is the closest public analogue to the program-optimization thesis: collect data not just to report activity, but to evaluate and change the care program.
Open the CMS national summary →
05
Where the market is now
Why this old operating experience suddenly matters again.
Philips Virtual Care Management — configurable chronic-care programs, connected devices, coaching, and actionable data
Philips describes a virtual-care model built around configurable programs rather than standalone devices: connected monitoring, coaching, analytics, and timely intervention for patients, providers, and payers.
Why read it: it shows where connected-device value ultimately migrated — from hardware toward integrated program operations and intervention.
Read the Philips announcement →ACCESS — recurring Outcome-Aligned Payments for technology-supported chronic care
CMS is explicitly testing a model in which organizations receive recurring payments for managing chronic conditions and earning full payment is tied to measurable outcomes rather than simply performing a specific activity or supplying a device.
Why read it: this is the commercial opening. The payer is buying a result-producing care program, which means someone has to design the program, intervention logic, technology stack, governance, and analytics underneath it.
Read the official CMS ACCESS model →TEMPO — real-world evidence for digital-health devices used inside ACCESS care models
FDA’s TEMPO pilot is designed around devices operating in real care settings, with manufacturers expected to collect, monitor, analyze, and report real-world data relevant to ACCESS outcomes.
Why read it: device strategy and program strategy are converging. The technology has to fit the care model, create measurable value, and produce evidence that survives regulatory scrutiny.
Read the official FDA TEMPO pilot →
Historical primary artifacts not published here
The BellSouth / AT&T Active Disease Management pilot poster and the Diabetech response to the Texas Medicaid / McKesson telemonitoring RFP are in the source library. They are not linked publicly here because they contain historical proprietary / program material. Selected access can be discussed directly.
Contact
If this is what you are building, we should talk.
I understand the opportunity in front of you. I have been here before — starting in 2005. Let me know if some of that experience would be useful.