· Executive Brief
Diabetic Eye Disease Awareness Month 2026
Use the observance to review how people find information, enter the right route, and remain connected through follow-through.
Leadership signal
Recognition becomes useful when leaders inspect the real route.
The strongest use of the observance is operational. Map where questions enter, how requests move, when ownership changes, and how unresolved work becomes visible. The goal is not to make a clinical claim. It is to make the organization’s response easier to understand and manage.
Start by distinguishing what is known from what the organization still needs to learn. A campaign plan, communication volume, or meeting count does not establish that a route is usable. Leaders need direct listening, clear ownership, locally defined measures, and a method for keeping exceptions visible.
Executive opening
Diabetic eye disease is a screening problem that lives outside the eye clinic
Most people with diabetes are seen in primary care, endocrinology or community health centers, not in an eye clinic. Their retinal exam therefore depends on a separate order, referral, appointment or imaging step that someone has to own. When that step fails, vision loss can progress without symptoms until it is hard to reverse. The failure is rarely dramatic; it is an order not placed, an image not read or a referral not scheduled.
The national picture shows how often that step fails. In a cohort of more than 1 million fee-for-service Medicare beneficiaries with diabetes, annual screening rates ranged between 51.7% in 2022 and 55.3% in 2016.[1] Put plainly, close to half of these older adults with a known diagnosis went unscreened in a given year. In rural areas, fewer than half of patients with diabetes complete the recommended screening.[2]
The gap also follows the map of disadvantage. Medicare beneficiaries in the most deprived neighborhoods had a 24 percent lower probability of annual screening than those in the least deprived neighborhoods.[1] That pattern makes diabetic eye screening an equity measure as much as a quality measure.
Health systems are not short of tools. Teleretinal imaging, autonomous artificial intelligence and mobile clinics can all bring screening to where patients already are. Yet AI-aided testing has been commercially available for five years and adoption is still relatively limited.[3] The executive question is whether the system around each tool captures the image, reads it, routes the result and confirms that follow-up care happened.
The sections below summarize current peer-reviewed evidence on burden, screening technology, referral completion, access models, equity, workforce and communication. They are meant to sit beside the executive guidance already on this page and give leaders a shared factual base for choosing where to act first.
Current evidence
What the research shows about screening gaps and screening tools
Low screening completion is not confined to one country, payer or care model. At five public hospitals in Addis Ababa, 223 of 405 adults with diabetes (55.1%) adhered to recommended retinal screening.[5] Urban residence was associated with higher adherence (adjusted odds ratio 2.783), while lack of formal education (0.242) and poor knowledge of diabetic retinopathy (0.281) were associated with lower adherence.[5] The setting differs from most U.S. systems, but geography, education and knowledge appear again and again in the studies that follow.
People who go unscreened often carry substantial disease. During a screening campaign at primary healthcare units in northeastern Brazil, diabetic retinopathy affected 50.0% of 95 examined patients and maculopathy affected 22.9%.[4] Most participants were economically inactive (81.1%) and reported low income (83.2%).[4] In two isolated rural U.S. primary care clinics, pop-up eye clinic exams recorded diabetic retinopathy in 21.2% of patients (11 of 52), compared with 7.7% (3 of 39) for teleophthalmology.[2] Cataract was recorded in 55.8% of pop-up clinic patients (29 of 52), against 7.7% (3 of 39) with teleophthalmology.[2] The authors described this as substantial unmet eye pathology among patients reached in primary care.[2]
Autonomous artificial intelligence has moved retinal testing into primary care, but performance depends on the setting. Among health systems that adopted these tools, reported nonmydriatic gradability ranged from 49% to 75%, sensitivity from 87% to 100% and specificity from 60% to 91%.[3] The same review found no studies comparing the cleared systems on the same patients.[3] In a small municipality in southern Brazil, a smartphone-based AI device showed sensitivity of 100% and specificity of 66.3%.[12] Its positive predictive value was 49.2%, and eight participants were excluded because of low image quality.[12]
Equipment choice matters as much as the algorithm. At a primary health center in Punjab, India, one algorithm’s sensitivity was 97.5% on one nonmydriatic camera and 81.7% on another.[13] Its specificity was 62.7% and 53.8% on the same two cameras.[13] The authors concluded that context-specific validation is needed to ensure safe integration into primary care.[13]
For executives, these findings set two expectations. Vendor accuracy figures are a starting point rather than a guarantee, and ungradable images and false positives create real work for someone downstream. A local validation and monitoring plan belongs in the business case before the first camera is installed.
Figure 1
Diabetic eye screening completion stays low across settings and models
Percent of eligible patients or scheduled appointments completing a screening step, as reported in each study. Values are copied from the cited studies [1], [2], [5], [7], [10]; none is estimated.
Data note. These values come from different studies, populations, years and definitions of completion, so this is an indirect comparison; the bars show how often a screening step is missed, not which model performs best.
Limitation. Each value reflects its own denominator: some count eligible patients, one counts scheduled appointments, and follow-up windows range from six months to one year.
Figure data (5 rows)
| Measure | Value | Population, setting and year as reported | Source |
|---|---|---|---|
| Screened within six months of a teleretinal bulk order | 17.5% | 5,665 eligible patients with type 2 diabetes in a U.S. health system, stepped-wedge randomized quality improvement trial | [10] |
| Screened after teleretinal imaging was introduced | 40.3% | Two clinics of an urban federally qualified health center in Madison, Wisconsin, 2022 to 2024; 14.4% at baseline | [7] |
| Annual screening, Medicare beneficiaries with diabetes | 51.7% | 20 percent sample of fee-for-service Medicare claims, 2022; 55.3% in 2016 | [1] |
| Adhered to retinal screening recommendations | 55.1% | 223 of 405 adults with diabetes at five public hospitals in Addis Ababa, Ethiopia, 2023 | [5] |
| Scheduled pop-up eye clinic appointments completed | 58.4% | 52 of 89 appointments over 18 months at two isolated rural primary care clinics in upstate New York | [2] |
Continuum of care
From a captured image to completed eye care
A retinal image has value only when an abnormal or ungradable result reaches an eye care professional and the patient completes that visit. That handoff is where the evidence is thinnest. A systematic review of 17 Canadian programs for communities facing care barriers found that programs frequently identified unmet need.[8] However, referral completion, repeat screening, treatment receipt and patient-reported outcomes were inconsistently reported.[8]
Better detection also changes the load on downstream services. When optical coherence tomography was added nationwide to an optometrist-based teleophthalmology pathway in 79 Danish stores, monthly teleophthalmology referrals rose from 703 to 2,359.[9] Referrals to public healthcare rose from 0.55% to 1.74%, with stable acute referrals but increases in subacute (+178%) and routine (+292%) referrals.[9] Because acute referrals stayed stable, the added volume landed in scheduled subacute and routine care.[9] The study covered many retinal conditions rather than diabetic disease alone, but the capacity lesson applies to any screening expansion.[9]
Practical barriers keep many patients from regular eye examinations. At a tertiary diabetes center in Jordan, 30% of 998 participants had never had an eye examination, and 14% of those previously screened cited transportation difficulties.[6] Patients experience screening and treatment as one journey, not as separate services. In northeastern Brazil, financial hardship and transportation challenges were among the major barriers to screening.[4] There, 78.1% of participants reported challenges scheduling ophthalmologic evaluations, and 76.3% said no ophthalmologist was available in their city through the public system.[4]
These barriers place follow-through squarely within the reach of operations, not only clinical decision-making. Scheduling, travel support, financial counseling and continuity with the same team are all levers a health system controls.
The financial case for screening often ignores this downstream step. A review of 18 health economic analyses of AI versus manual grading found that most did not consider improved follow-up after AI screening, potentially underestimating its economic value.[15] Only 39% of those studies followed standardized reporting guidelines.[15] Health systems using AI testing, by contrast, reported improved patient adherence to follow-up with ophthalmology.[3] Leaders building an investment case should insist on local follow-up completion data rather than accuracy alone.
Figure 3
Why the pathway breaks
People
- Image capture skills uneven and lost with staff turnover
- Limited clinician training and confidence in teleophthalmology
- Belief that screening is unnecessary without symptoms
Process
- Screening ordered visit by visit rather than for the whole population
- Teleophthalmology and pop-up clinics run as separate programs
- No owner for abnormal or ungradable results
Technology
- Low gradability of nonmydriatic images
- AI performance varies by camera platform
- False positives add downstream referrals
Environment
- Transportation needs and distance to eye clinics
- Housing and food insecurity
- Neighborhood socioeconomic deprivation
Policy
- Financial hardship and travel limit screening attendance
- Uninsured patients and limited coverage
- Investment cases that ignore follow-up completion
Measurement
- Referral completion and treatment receipt rarely reported
- Screening rates not stratified by deprivation or insurance
- No local validation of AI accuracy
Unranked qualitative cause-and-effect diagram for the outcome in the box: People with diabetes do not complete retinal screening or the eye care visit that an abnormal or ungradable result requires Causes are assembled from the cited literature [1], [2], [3], [4], [6], [7], [8], [10], [11], [13], [14], [15]; categories are not weighted and their order carries no meaning.
Data note. Qualitative; no frequencies are implied.
Limitation. Causes drawn from published studies may not all apply locally; validate them with frontline teams.
Figure data (6 rows)
| Category | Causes drawn from the cited literature |
|---|---|
| People | Image capture skills uneven and lost with staff turnover; Limited clinician training and confidence in teleophthalmology; Belief that screening is unnecessary without symptoms |
| Process | Screening ordered visit by visit rather than for the whole population; Teleophthalmology and pop-up clinics run as separate programs; No owner for abnormal or ungradable results |
| Technology | Low gradability of nonmydriatic images; AI performance varies by camera platform; False positives add downstream referrals |
| Environment | Transportation needs and distance to eye clinics; Housing and food insecurity; Neighborhood socioeconomic deprivation |
| Policy | Financial hardship and travel limit screening attendance; Uninsured patients and limited coverage; Investment cases that ignore follow-up completion |
| Measurement | Referral completion and treatment receipt rarely reported; Screening rates not stratified by deprivation or insurance; No local validation of AI accuracy |
Access
Bringing retinal imaging to where patients already receive care
Placing a camera in primary care can change screening rates quickly when the program is designed around the clinic. In an urban federally qualified health center in Madison, Wisconsin, the baseline population of 1,094 patients was 24.0% Black or African American and 44.8% Hispanic or Latino.[7] Thirty percent were uninsured.[7] After teleretinal imaging was introduced at two clinics, diabetic eye screening increased from 14.4% to 40.3%.[7]
Ordering can also move from the individual visit to the population. In a stepped-wedge randomized quality improvement trial, a centralized provider placed teleretinal screening bulk orders for 5,665 eligible patients with type 2 diabetes.[10] Each order included a retinal photo order and an electronic patient education message.[10] Six months after the first bulk order, the overall screening completion rate was 17.5%, and receiving a bulk order increased the odds of completing screening (odds ratio 1.78).[10] The approach removes the need for each clinician to remember the order during a busy visit. It also gives a central team a defined cohort to track from order to completed image.
Program performance improves when leaders manage it as an operating system rather than a device rollout. In the second year of an AI-based screening program in Northern California primary care and endocrinology clinics, gradability rose from 62.3% to 71.2%.[14] Patient encounters increased by 21.9% after targeted training protocols and camera utilization reports were introduced.[14] The authors described the result as gradability comparable to nonmydriatic rates observed in clinical trials.[14] The lesson for leaders is that utilization and image quality each need a named owner and a regular report.
Rural programs face a different mix of constraints. One study drew on records from 885 adults with diabetes at two isolated rural primary care clinics in upstate New York.[2] Of these, 586 had an exam with an eye doctor, 91 had an on-site primary care-based exam and 208 had no recent exam.[2] Over 18 months, 58.4% of scheduled pop-up clinic appointments were completed, and 92.3% of teleophthalmology images were gradable.[2] Staff proposed making same-day teleophthalmology a default part of annual diabetes visits, with targeted pop-up referrals for abnormal findings.[2]
The common thread is that access improved when imaging was built into an existing visit and someone owned the order, the image quality and the result. Separate programs running in parallel created inefficiencies in both programs, while integrated workflows were the direction staff themselves recommended.[2]
Executive guidance · System route
Map the path from first contact to verified follow-through.
Use this route as a discussion framework. Replace each stage with the organization’s approved process, qualified roles, current service capacity, and applicable requirements. Do not infer a clinical protocol from the observance title.
The route at a glance
- Listen
- Clarify entry
- Design for visual access
- Protect transitions
- Verify and learn
Reliability rule: Every stage needs a named owner, an observable status, an approved escalation path, and a defined signal that shows whether responsibility moved or remained unresolved.
Listen
Gather questions and friction points from people using the route, frontline teams, and community-facing staff. Separate observed barriers from assumptions.
Clarify entry
Publish a plain-language first step and identify who receives requests through each approved channel.
Design for visual access
Review whether information, digital channels, physical wayfinding, and transition instructions are usable with locally appropriate access support.
Protect transitions
Use acknowledgment and status visibility when responsibility moves between services, locations, or organizations.
Verify and learn
Review unresolved work, repeat contacts, and avoidable handoffs. Use the findings to improve the route rather than blame individuals.
Figure 2
The system route as one flow
The five stages of the system route described in this brief, drawn as one flow with a learning loop back to the start. The stages are the page’s original executive guidance, not a clinical protocol.
Data note. Shows sequence and handoffs only; it carries no counts or rates.
Limitation. Local processes may add, merge or reorder stages under organizational policy.
Figure data (5 rows)
| Stage | What the stage requires |
|---|---|
| Listen | Gather questions and friction points from people using the route, frontline teams, and community-facing staff. Separate observed barriers from assumptions. |
| Clarify entry | Publish a plain-language first step and identify who receives requests through each approved channel. |
| Design for visual access | Review whether information, digital channels, physical wayfinding, and transition instructions are usable with locally appropriate access support. |
| Protect transitions | Use acknowledgment and status visibility when responsibility moves between services, locations, or organizations. |
| Verify and learn | Review unresolved work, repeat contacts, and avoidable handoffs. Use the findings to improve the route rather than blame individuals. |
Equity and trust
Who is missed, and which programs reach them
Neighborhood disadvantage tracks closely with missed screening. In the national Medicare cohort, screening rates fell consistently as the Area Deprivation Index rose.[1] Beneficiaries in the most deprived group (index 86 to 100) had a relative risk of 0.76 for annual screening.[1] The comparison group was beneficiaries in the least deprived group (index 1 to 15).[1] That gradient gives leaders a ready way to stratify their own screening data.
On-site programs can reach people that off-site referrals miss. In the rural primary care study, patients who received an on-site exam were the most socially disadvantaged group.[2] Only 46.2% reported no transportation needs, 30.8% reported no housing instability and 39.6% reported no food insecurity.[2] For these patients, a referral to a distant eye clinic may simply never be completed.
Teleretinal imaging can narrow some gaps, but not automatically. In the Madison community health center program, teleophthalmology increased screening equity by sex and insurance status.[7] The authors noted that clinic-level strategies for patient outreach, staff engagement and resource allocation may allow further gains in screening equity.[7] Equity, in other words, depends on how the program is run as much as on the technology.
Programs built for specific communities show what low-barrier design looks like. In the Canadian review, seven studies focused on Indigenous or First Nations communities and five on people experiencing homelessness or marginal housing.[8] Teleophthalmology, teleretinal and mobile diabetes screening models were the most common program type, at 8 of 17 studies (47.1%).[8] The evidence supported co-location, culturally safe partnerships, low-barrier referral pathways and navigation, though certainty remained limited.[8]
Cost and travel recur as structural barriers across very different settings. Researchers in northeastern Brazil pointed to telemedicine-based screening and subsidized transportation as structural responses.[4] U.S. leaders can apply the same test locally by asking whether the next step after a positive image is affordable and reachable for the patients least likely to complete it. Education and geography shape screening outside the United States as well. In the Ethiopian study, urban residents were more likely to adhere to screening, and people without formal education were less likely to adhere.[5] Leaders should read findings like these as a prompt to stratify, not to stereotype. Breaking screening and referral completion out by neighborhood, insurance, language and rurality shows where a route is failing. The usual remedy is a change in the route rather than in the patient.
Human infrastructure
The people and skills that keep screening programs running
Screening technology succeeds or fails through the staff who operate it. Health systems that adopted autonomous AI described proper site selection, alignment with primary care workflows, streamlined patient engagement and referrals, and ongoing staff training as elements of success.[3] Those adopters also reported improved quality measures, health equity, productivity and patient adherence to ophthalmology follow-up.[3]
Frontline teams name the friction points clearly. In the rural primary care study, staff identified training and workflow complexity as barriers for teleophthalmology.[2] For pop-up clinics, the barriers were underutilization and scheduling challenges.[2] Implementation champions across roles then proposed the integrated workflow themselves, which shows frontline staff acting as design partners rather than recipients of change.[2] In Northern California, targeted training protocols and camera utilization reports were among the interventions that improved screening efficiency.[14] Image capture is a skill that must be taught, monitored and refreshed as staff turn over.
Clinician readiness is uneven. In a survey of 489 general practitioners in Colombia, undergraduate telemedicine training was uncommon (9.0%) and prior teleophthalmology experience was rare (1.8%).[11] Even so, 57.3% reported positive perceptions of teleophthalmology.[11] Greater undergraduate ophthalmology training was associated with high ophthalmologic confidence, including adjusted odds ratios of 8.82 for 51 to 100 hours and 14.71 for more than 100 hours.[11] Respondents also raised concerns about privacy, workload, job security and implementation readiness.[11]
Specialist reading and triage capacity is the other half of the workforce equation. The Danish pathway combined standardized optometrist image acquisition, consultant ophthalmologist review and structured teleophthalmology triage.[9] Its authors stressed that the results reflected that integrated model rather than the imaging device alone.[9] Expanding capture without planning reading time and appointment slots simply moves the queue. Navigation belongs in the same plan. The Canadian review found support for low-barrier referral pathways and for spectacle or treatment navigation in programs serving communities facing barriers.[8]
For executives, the workforce plan for diabetic eye screening should name who captures images, who checks quality, who reads results, who calls patients and who holds open referrals. Each of those roles needs training, protected time and a clear escalation path. Leaders should also expect turnover and build refresher training and quality checks into the operating budget rather than treating them as launch costs. A program that depends on one trained person at each site is fragile by design.
Executive guidance · Operating dashboard
Measure whether the operating route works as intended.
Do not publish a number until the organization defines its numerator, denominator where relevant, owner, source system, exclusions, review cadence, and limits. The candidates below are management prompts, not benchmarks or reported results.
Four operating signals
Each signal pairs an executive question with one candidate local measure, set out below.
Executive guidance · Signal
Entry-point clarity
Can people and staff identify the approved first step without searching across conflicting instructions?
- Candidate local measure
- Locally defined contact, routing, and repeat-contact signals by channel and service.
Executive guidance · Signal
Ownership reliability
Does a specific team accept each request and make its status visible?
- Candidate local measure
- Requests without an owner, aged work, transfers, and escalations using locally approved definitions.
Executive guidance · Signal
Visual usability
Do communication and navigation channels work for people with differing visual-access needs?
- Candidate local measure
- Locally defined format requests, wayfinding barriers, inaccessible content, repeat contacts, and corrective-action status.
Executive guidance · Signal
Improvement closure
Does the organization correct recurring friction and report what changed?
- Candidate local measure
- Selected barriers, assigned actions, completion status, and post-change review.
Figure 4
The operating system behind the dashboard
The four operating signals from this brief’s dashboard, each tied to the candidate owner named in the executive scorecard, around one accountable route. Drawn from the page’s original executive guidance.
Data note. Structural diagram; no measured values.
Limitation. Owners are candidates to be confirmed locally.
Figure data (4 rows)
| Operating signal | Candidate owner |
|---|---|
| Entry-point clarity | Executive sponsor and operational lead |
| Ownership reliability | Operational owner and participating teams |
| Visual usability | Sending and receiving owners |
| Improvement closure | Improvement owner and executive sponsor |
Executive guidance · Handoff workflow
Keep work owned when it crosses a boundary.
A handoff is reliable only when the sending party knows who is expected to receive the work, the receiving party acknowledges responsibility, and any unresolved exception remains visible. Local policy and qualified judgment control the actual process.
How responsibility moves
Lane A
From message to accepted ownership
- Confirm the approved message and intended audience.
- Publish the actual entry point and receiving role.
- Acknowledge receipt using the locally approved method.
- Escalate when ownership or response remains unclear.
Lane B
From accepted ownership to verified follow-through
- Record the next accountable owner and visible status.
- Record any locally approved communication or access support needed for the next step without making assumptions.
- Keep unresolved work visible across boundaries.
- Confirm closure or document the continuing exception.
Executive guidance · Executive scorecard
Ask questions that connect intent, ownership, and results.
The scorecard keeps review focused on decisions. Leaders should set definitions locally, examine variation without overclaiming cause, and pair quantitative signals with the experience of people who use and operate the route.
Who reviews what
- Executive sponsor and operational lead
- Operational owner and participating teams
- Sending and receiving owners
- Improvement owner and executive sponsor
Candidate owners named in the scorecard, confirmed locally.
| Operating signal | Executive question | Candidate owner | Suggested review point |
|---|---|---|---|
| Entry-point clarity | Can people and staff identify the approved first step without searching across conflicting instructions? | Executive sponsor and operational lead | At campaign launch and close |
| Ownership reliability | Does a specific team accept each request and make its status visible? | Operational owner and participating teams | At a locally defined cadence |
| Visual usability | Do communication and navigation channels work for people with differing visual-access needs? | Sending and receiving owners | At a locally defined cadence |
| Improvement closure | Does the organization correct recurring friction and report what changed? | Improvement owner and executive sponsor | Until action is verified |
Figure 5
Control points for a reliable diabetic eye screening pathway
| Control point | Risk if missing | Owner | Evidence of control |
|---|---|---|---|
| Population-level screening orders for every eligible patient | Patients who miss visits or whose clinicians forget the order are never screened | Ambulatory quality lead with population health team | Share of eligible patients with an active order, reviewed monthly |
| Local validation of AI or teleretinal performance on installed cameras | Accuracy differs from vendor figures and false positives overload referrals | Chief medical information officer with ophthalmology lead | Documented validation sample and periodic accuracy review by site |
| Image quality monitoring and staff training | Ungradable images stall the pathway and patients are lost | Primary care practice managers | Gradability rate by site and camera, with training records |
| Owned result routing for abnormal and ungradable images | Positive findings sit unread or unreferred | Teleophthalmology program manager | Time from image to read and from read to referral, with open items listed |
| Referral completion tracking and navigation | Patients with disease never reach an eye care visit | Care navigation lead with eye care service line | Share of referred patients with a completed visit, stratified by insurance and deprivation |
| Specialist reading and appointment capacity planning | Better detection creates queues for subacute and routine care | Service line administrator for ophthalmology | Referral volume against available reading time and appointment slots |
| Financial and travel support for ongoing treatment | Patients miss screening or treatment because of cost, travel or low awareness | Patient access and financial counseling leaders | Use of assistance and reported barriers among patients who miss treatment visits |
Control points derived from the cited evidence [1], [2], [3], [6], [7], [9], [10], [13], [14] and this brief’s executive guidance.
Data note. Each row is a control to verify, not a measured result.
Limitation. Owners and evidence of control must follow local policy and applicable requirements.
Public communication
Messages that move patients to the next step
Awareness campaigns work best when they close specific knowledge gaps. In the Ethiopian hospital study, poor knowledge of diabetic retinopathy was associated with lower screening adherence.[5] The authors urged clinicians to take a more proactive role in education, with emphasis on rural residents and people without formal education.[5] In northeastern Brazil, limited awareness of the importance of screening was among the major barriers identified.[4]
Trust in familiar staff is an asset to build on. Rural patients described on-site teleophthalmology and pop-up clinics as a convenient one-stop shop and valued their trust in primary care staff.[2] Messaging that comes from a patient’s own care team, and offers screening during a visit already scheduled, can draw on that trust. Health systems that adopted AI testing likewise listed streamlined patient engagement and referrals among the elements of successful implementation.[3]
Leaders should also check whether staff assumptions match patient priorities. Asking patients directly, rather than inferring priorities from missed appointments, gives outreach teams a sounder starting point. In northeastern Brazil, 33.3% of participants reported difficulties attending medical appointments for diabetes management.[4] The researchers there recommended targeted educational initiatives alongside structural interventions.[4]
Digital outreach can be paired directly with the order. The bulk order trial attached an electronic patient education message to each retinal photo order and examined the effect of opening that message.[10] Pairing a message with an action the patient can take is more useful than a general reminder that eye health matters. The authors concluded that population-level interventions should be implemented to improve diabetes preventive care.[10] Messages should also speak to practical worries such as cost, travel and scheduling, which patients in northeastern Brazil identified as barriers.[4]
Public communication should also be honest about what screening tools can and cannot do. Because AI performance varies by camera and setting, claims about accuracy should reflect local validation rather than vendor figures.[13] This month is a chance to tell patients exactly where to get screened, what happens after an abnormal result and who will call them. Every public message should lead to a route that the organization has tested and can staff.
Executive guidance · 90-day plan
Move from listening to a verified improvement in 90 days.
The time boxes below are a management framework, not a clinical timeline. Adjust them to risk, scope, governance, resources, stakeholder participation, and applicable review requirements.
Ninety days, three moves
- Days 1–30See the current route
- Days 31–60Build the reliable path
- Days 61–90Make the change visible
Days 1–30 · Listen and map
See the current route
- Name an executive sponsor and an operational owner.
- Listen to people using the route and representative frontline teams.
- Map how Diabetic Eye Disease Awareness Month communications connect to real entry points, owners, transitions, and escalation.
- Select one recurring barrier that the organization can responsibly address.
Days 31–60 · Design and test
Build the reliable path
- Define the trigger, receiving owner, required information, acknowledgment, escalation, and closure signal.
- Validate the design with every sending and receiving team involved.
- Audit one high-use communication or navigation route for visual usability and correct a verified barrier.
- Test the route in more than one representative setting or operating period where practical.
Days 61–90 · Implement and verify
Make the change visible
- Launch the redesigned route with a small set of locally defined measures.
- Review unresolved work and recurring friction using a learning approach.
- Correct implementation barriers and record who owns each remaining dependency.
- Report what changed, what evidence was reviewed, and what still requires live oversight.
Figure 6
The 90-day plan on one timeline
The 90-day plan from this brief drawn as a timeline: each action sits in the 30-day window the plan assigns it. The full wording of each action is in the figure data and in the plan above.
Data note. Time boxes are a management framework, not a clinical timeline.
Limitation. Durations are uniform 30-day windows; real tasks may span windows.
Figure data (12 rows)
| Window | Action |
|---|---|
| Days 1–30 · Listen and map | Name an executive sponsor and an operational owner. |
| Days 1–30 · Listen and map | Listen to people using the route and representative frontline teams. |
| Days 1–30 · Listen and map | Map how Diabetic Eye Disease Awareness Month communications connect to real entry points, owners, transitions, and escalation. |
| Days 1–30 · Listen and map | Select one recurring barrier that the organization can responsibly address. |
| Days 31–60 · Design and test | Define the trigger, receiving owner, required information, acknowledgment, escalation, and closure signal. |
| Days 31–60 · Design and test | Validate the design with every sending and receiving team involved. |
| Days 31–60 · Design and test | Audit one high-use communication or navigation route for visual usability and correct a verified barrier. |
| Days 31–60 · Design and test | Test the route in more than one representative setting or operating period where practical. |
| Days 61–90 · Implement and verify | Launch the redesigned route with a small set of locally defined measures. |
| Days 61–90 · Implement and verify | Review unresolved work and recurring friction using a learning approach. |
| Days 61–90 · Implement and verify | Correct implementation barriers and record who owns each remaining dependency. |
| Days 61–90 · Implement and verify | Report what changed, what evidence was reviewed, and what still requires live oversight. |
Executive guidance · Leadership close
Carry one accountable improvement beyond the campaign window.
Recognition matters. It becomes durable when leaders also listen, clarify the route, support the people operating it, close an identified gap, and report the result without claiming more than the evidence supports.
Leadership commitment
Listen to the people closest to the route, assign one barrier, and verify the change.
Executive action kit · Month observance
Turn Diabetic Eye Disease Awareness Month into accountable action.
Use the observance to review how people find information, enter the right route, and remain connected through follow-through.
Leadership focus
Use this observance to examine navigation, longitudinal ownership, and specialty-care coordination. Select one verified barrier, assign an accountable owner, and carry the improvement beyond the campaign window.
Workforce lens
Ask whether staffing, role clarity, training, tools, workload, and escalation support the people operating the accessible information, evaluation navigation, accommodation, and follow-through pathway.
Patient and community lens
Make the approved first step clear. Test whether a person can move from information to an acknowledged next action without navigating conflicting instructions.
Equity and access lens
Review visual accessibility, language, transportation, digital design, disability accommodation, and access to appropriate services.
Five-minute briefing
Inspect the operating sequence
Listen with people using the pathway, caregivers, eye-care services, primary care, rehabilitation, and access teams. Find one inaccessible information, unclear entry, or a referral that is not completed, then test the locally approved route from entry through acknowledged follow-through.
- Listen
- Clarify entry
- Design for visual access
- Protect transitions
- Verify and learn
Leadership actions for this week
- Name an executive sponsor and frontline operating owner.
- Ask people using and operating the pathway where ownership becomes unclear.
- Test one representative route from first question to acknowledged next step.
- Select one barrier that can be corrected without overstating the evidence.
- Set a review date and define how completion will be verified.
Candidate measures
Define every numerator, denominator where relevant, owner, data source, exclusions, cadence, and limitation locally. These are management prompts, not external benchmarks.
- Locally defined contact, routing, and repeat-contact signals by channel and service.
- Requests without an owner, aged work, transfers, and escalations using locally approved definitions.
- Locally defined format requests, wayfinding barriers, inaccessible content, repeat contacts, and corrective-action status.
Department readiness checklist
- The public and staff entry points match the actual approved process.
- A specific role accepts each request, referral, or escalation.
- Handoffs include acknowledgment and a visible unresolved state.
- Language, disability, digital, transportation, and trust barriers receive explicit review.
- Communications do not introduce unsupported themes, statistics, or clinical advice.
- A named leader will review what changed after the observance.
Intended audiences
- Executive and Operational Leaders
- Patient Experience and Access Leaders
Staff communication template
During Diabetic Eye Disease Awareness Month, our organization will connect awareness with a practical review of the accessible information, evaluation navigation, accommodation, and follow-through pathway. Use approved information, identify the correct entry point, confirm ownership when work moves, protect privacy, and escalate unresolved barriers through local channels.
Community communication template
Diabetic Eye Disease Awareness Month is an opportunity to share trustworthy information and make the next step easier to find. Use our approved channels for information and support. If a request changes hands, our goal is to keep ownership and follow-through visible.
Measurement worksheet
- Signal
- What observable condition will show whether the route works?
- Definition
- What is included, excluded, and counted?
- Owner
- Who reviews the signal and acts on exceptions?
- Cadence
- When will leaders review it?
- Equity check
- Which differences require protected, locally appropriate review?
- Closure
- What evidence will confirm the improvement was completed?
- Editorial review
- Completed
- Source records
- 2 listed under Authoritative resources
- Review standard
- Sourced facts remain separated from original executive guidance
- Clinical boundary
- No new statistical or clinical claim was added by this action kit
Sources and content boundary
Authoritative resources
Use note: All operating models, measures, scorecard questions, and action steps are original executive guidance. This article supports leadership planning and education. It does not replace clinical judgment, emergency procedures, legal advice, professional standards, organizational policy, or applicable federal, state, and local requirements.
Peer-reviewed references
References
Fifteen peer-reviewed sources, newest first. Each was individually verified through authenticated academic research databases. Bracketed numbers in the text link here; the original executive guidance on this page carries no citations because it is guidance, not a sourced claim. Sources last checked in those databases on .
- 1.Borkar DS, Ford CB, Song A, Lusk JB, Bosquet A, Muir KW, et al. Neighborhood Socioeconomic Deprivation and Adherence to Diabetic Retinopathy Screening Guidelines. American Journal of Ophthalmology. 2026;292:275-283. https://doi.org/10.1016/j.ajo.2026.07.054
- 2.Arya NG, Koltchine V, Gearinger TC, Buehler S, Yousefi-Nooraie R, Ramchandran RS. Implementing Diabetic Eye Care in Rural Primary Care Through Teleophthalmology and Pop-Up Clinics: Mixed Methods Study. JMIR Diabetes. 2026;11:e106290. https://doi.org/10.2196/106290
- 3.Teng CW, Patel SD, Barkmeier AJ, Liu TA, Myung D, Henderer J, et al. Autonomous Artificial Intelligence in Diabetic Retinopathy Testing—Lessons Learned on Successful Health System Adoption. Ophthalmology Science. 2026;6(1):100935. https://doi.org/10.1016/j.xops.2025.100935
- 4.dos Santos ML, Florêncio LBDS, de Almeida LRB, Zwecker LJ, Alves M, de Carvalho AB, et al. Barriers to diabetic retinopathy screening in public healthcare in northeastern Brazil. Arquivos Brasileiros de Oftalmologia. 2026;89(3):01-14. https://doi.org/10.5935/0004-2749.2025-0248
- 5.Dinsa F, Aga F, Gela D, Wldesilase A. Adherence to retinal screening recommendations and its associated factors among patients with diabetes in Ethiopian public hospitals: a cross-sectional study. BMJ Open. 2026;16(4):e100026. https://doi.org/10.1136/bmjopen-2025-100026
- 6.Albakri YJ, Aldabbagh FA, Sabbagh HM, Khashman MK, Farahid O, Ali RM, et al. Barriers to Regular Eye Examination in Individuals with Diabetes at a Tertiary Diabetes Centre in Jordan: A Cross-Sectional Study. International Journal of Environmental Research and Public Health. 2026;23(2). https://doi.org/10.3390/ijerph23020147
- 7.Khoussine J, Peterson K, Fowlkes C, Maldonado C, Hoang J, Emmet T, et al. Teleretinal Imaging Enhances Equity in Diabetic Eye Screening at Urban Community Health Clinics. Telemedicine and e-Health. 2026:15305627261477819. https://doi.org/10.1177/15305627261477819
- 8.Kaleem S, Tuitt-Barnes D, Rahman U, Kaleem S, Balian C, Chandrakumar M, et al. Vision programs for Canadian communities facing care barriers: a systematic review. Canadian Journal of Ophthalmology. 2026. https://doi.org/10.1016/j.jcjo.2026.08.003
- 9.Muttuvelu DV, Cehofski LJ, Faber C, Falk MK, Giannaccare G, Grauslund J, et al. Referral patterns after nationwide implementation of OCT-enabled teleophthalmology in optometrist stores in Denmark. European Journal of Ophthalmology. 2026:11206721261488973. https://doi.org/10.1177/11206721261488973
- 10.Takimoto SW, Freeby M, Han M, Vangala S, Fujimoto J, Dermenchyan A, et al. Population-Level Teleretinal Screening Bulk Orders to Improve Diabetic Retinopathy Screening Rates: A Randomized Quality Improvement Trial. Telemedicine and e-Health. 2026;32(9):903-911. https://doi.org/10.1177/15305627261450353
- 11.Loaiza-Guevara V, Martinez-Rivera JP, Restrepo N, Quintero-Patino L, Pinto-Arias A, Martinez LM, et al. Teleophthalmology adoption and perceived barriers among Colombian general practitioners: a cross-sectional study. Frontiers in Digital Health. 2026;8:1814108. https://doi.org/10.3389/fdgth.2026.1814108
- 12.Alves DO, Lavinsky D, Strzalkowski MEB, Moreira LB. Accuracy of portable artificial intelligence-supported fundus camera in the screening of diabetic retinopathy in primary care. Archives of Endocrinology and Metabolism. 2026;70(4):01-11. https://doi.org/10.20945/2359-4292-2026-0051
- 13.Chauhan A, Rana G, Verma P, Yadav M, Kumar L, Kaur G, et al. Diagnostic performance of multiple artificial intelligence (AI) algorithms for diabetic retinopathy screening in primary care: Evidence from real-world settings in India. Indian Journal of Ophthalmology. 2026. https://doi.org/10.4103/ijo.ijo_2774_25
- 14.Chen KM, Zhao CS, Knapp A, Dow E, Phadke A, Tan M, et al. Targeted Interventions Lead to Quality Improvement in Year 2 of an Artificial Intelligence-Based Diabetic Retinopathy Detection Program in Northern California. Retina. 2025;45(8):1469-1480. https://doi.org/10.1097/iae.0000000000004499
- 15.Leigh J, Drinkwater J, Turner A, Schroeder E. Health Economic Considerations for the Implementation of Artificial Intelligence‐Enabled Diabetic Retinopathy Screening: A Review. Clinical & Experimental Ophthalmology. 2025;54(1):144-161. https://doi.org/10.1111/ceo.70016
Claim-to-source matrix
What each source supports
Every sourced claim in the new sections maps to the references that support it, the section where it is used, and the limitation a reader should keep in mind. Original executive guidance carries no sources because it is guidance, not a finding.
| Claim group | Sources | Years | Used in | Limitation |
|---|---|---|---|---|
| National screening adherence and the neighborhood deprivation gradient | [1] | 2026 | Executive opening | Fee-for-service Medicare claims, 20 percent sample, 2016 to 2022; screening identified from billing codes; excludes Medicare Advantage and younger adults. |
| Screening adherence, disease found and barriers in international settings | [4] [5] | 2026 | Current evidence | Single-city cross-sectional studies in Ethiopia and Brazil; small or campaign-based samples; associations are not causal. |
| AI screening accuracy and gradability vary by system, camera and site | [3] [12] [13] | 2026 | Current evidence | Small diagnostic studies and a narrative review of adopters; no head-to-head comparison of cleared systems on the same patients. |
| Referral completion and downstream capacity after screening | [8] [9] | 2026 | Continuum of care | Canadian review reports inconsistent downstream outcomes; Danish study covers many retinal conditions, not diabetic disease alone, and is observational. |
| Patient-reported and practical barriers to follow-through | [4] [6] | 2026 | Continuum of care | Jordan data are self-reported barriers from one tertiary diabetes center; Brazilian data come from a small single-campaign cross-sectional sample. |
| Economic evaluation of AI screening omits follow-up | [15] | 2025 | Continuum of care | Review of modeled analyses with wide methodological variation. |
| Primary care teleretinal imaging, bulk orders and program improvement raise screening | [2] [7] [10] [14] | 2026, 2025 | Access | One randomized quality improvement trial; others are single-system observational studies with before-and-after designs. |
| On-site and community programs reach socially disadvantaged patients | [1] [2] [7] [8] | 2026 | Equity and trust | Observational designs; small subgroups in the rural study; certainty limited in the Canadian review. |
| Workforce training, readiness and implementation elements | [2] [3] [11] [14] | 2026, 2025 | Human infrastructure | Clinician survey from Colombia and qualitative or descriptive implementation reports; self-reported confidence. |
| Knowledge gaps, trust and communication levers | [2] [4] [5] [10] | 2026 | Public communication | Cross-sectional and qualitative evidence; the bulk order trial tested an order plus message, not messaging alone. |
About the author
Greg Wahlstrom
MBAHCM
President and Chief Executive Officer, The Healthcare Executive
View Greg’s profile©2014 – 2026 The Healthcare Executive. All rights reserved.