Highlights
In this retrospective study of 1121 eyes from 600 patients followed at a tertiary care center, delayed follow-up (DFU) was common, occurring in 53.4% of visits.
Across the full cohort, several DFU metrics, including percentage of visits with delay, mean days delayed, and maximum delay, were not significantly associated with visual field progression after multivariable adjustment.
Progression was more frequent in eyes with worse baseline disease: 32.8% for baseline mean deviation (MD) ≤ -6 dB versus 19.8% for baseline MD > -6 dB.
In a matched case-control sensitivity analysis, mean DFU >60 days was associated with nearly doubled odds of progression in patients with baseline MD < -6 dB, suggesting that prolonged delays may still be clinically relevant in more advanced glaucoma.
Background
Glaucoma is a chronic optic neuropathy and one of the leading causes of irreversible blindness worldwide. Because structural optic nerve damage and visual field loss typically accumulate slowly and often without symptoms until late stages, long-term surveillance is central to management. Follow-up intervals are intended to balance competing priorities: detecting progression early enough to modify treatment, minimizing patient burden, and using limited clinic capacity efficiently.
In practice, however, actual follow-up timing often diverges from planned scheduling. Patients may miss appointments because of transportation barriers, cost, comorbidity, caregiving obligations, or limited health system access. Clinics may also reschedule or defer visits because of capacity constraints. Despite the importance of surveillance in glaucoma care, there remains limited empirical evidence defining how much delay is clinically meaningful across disease severity strata. As a result, recommended follow-up intervals often rely on expert consensus, perceived risk, and local practice patterns more than high-quality comparative data.
The study by Kolli and colleagues addresses this evidence gap by examining whether delayed follow-up is associated with worsening on Humphrey Visual Field (HVF) testing. The question is clinically important because it touches both individual patient care and service design. If modest delays do not materially worsen outcomes for most low-risk patients, follow-up schedules could potentially be individualized more efficiently. Conversely, if delays are especially harmful in higher-risk groups, that would support targeted efforts to protect timely access for these patients.
Study Design and Methods
Design and setting
This was a retrospective cohort study with a matched case-control sensitivity analysis conducted at a tertiary care center. Investigators used electronic health record data linked with HVF results from 2014 through 2023.
Population
The primary analysis included glaucoma patients and glaucoma suspects with at least 5 HVF tests during the study period. A total of 1121 eyes from 600 patients met inclusion criteria. Mean follow-up was 7.3 years, mean baseline age was 73.7 years, and 50.2% of participants were female.
The authors also performed subgroup analyses by baseline disease severity, stratified by MD greater than -6 dB versus MD less than or equal to -6 dB. This threshold is clinically familiar and separates milder from more established functional loss.
Exposure: delayed follow-up
Delayed follow-up was defined relative to the interval recommended by the treating provider. The study quantified delay in three ways: the percentage of visits with delayed follow-up, the mean number of delayed days per patient, and the maximum delay experienced.
This approach is attractive because it captures not only whether delay occurred, but also its frequency and magnitude. At the same time, it depends on the quality and consistency of provider-documented intended follow-up intervals, which may vary by clinician and by contextual factors not fully measurable in the record.
Outcome
The primary endpoint was visual field progression, defined as an MD linear regression slope of at most -0.5 dB per year across all available HVF tests. This is a clinically interpretable threshold commonly used to identify meaningful functional decline over time, though any slope-based definition has known limitations related to test variability, regression assumptions, and the number and timing of examinations.
Statistical approach
Associations between DFU metrics and visual field progression were examined with covariate-adjusted multivariable logistic regression. Subgroup analyses were performed by baseline MD. A sensitivity analysis included patients with at least 2 HVF tests and used matched case-control methods, matching on age, follow-up duration, number of HVF tests, and baseline MD. This secondary analysis was designed to test the robustness of the main findings under a different analytic framework.
Key Results
Progression rates
Visual field progression was not rare. It occurred in 19.8% of eyes with baseline MD > -6 dB and in 32.8% of eyes with baseline MD ≤ -6 dB. This gradient is biologically and clinically plausible: eyes entering follow-up with greater functional loss may have either more aggressive disease, less reserve, or both.
Delayed follow-up was common
Any delayed follow-up occurred in 53.4% of all visits. This is an important descriptive finding on its own. It suggests that deviation from the prescribed schedule is not an exception in routine tertiary glaucoma care but a frequent occurrence. From a health services perspective, this prevalence underscores the need to distinguish benign scheduling variability from delays that truly place vision at risk.
Primary multivariable analysis
In the adjusted analyses, none of the three main DFU metrics was associated with visual field progression in the overall cohort. Specifically, neither the percentage of visits with delay, nor the mean number of days delayed, nor the maximum delay predicted progression after accounting for covariates.
The absence of a broad association is the study’s central finding. For clinicians, the most immediate implication is that follow-up intervals longer than originally prescribed may not invariably translate into measurable visual field worsening for the average glaucoma patient or suspect in this tertiary-care population. This result challenges a common assumption that any delay is necessarily harmful.
Subgroup analyses
The null findings remained in subgroup analyses stratified by baseline MD. This suggests that, within the primary cohort analysis, the overall lack of association was not fully explained by baseline disease severity. Still, subgroup analyses can be underpowered, particularly when progression is defined by a relatively stringent functional endpoint.
Other factors associated with progression
Several variables were significantly associated with higher odds of visual field progression: older age, larger cup-to-disc ratio, shorter follow-up duration, higher number of HVF tests, and unmet social needs.
Some of these associations deserve careful interpretation. Older age and larger cup-to-disc ratio are clinically credible risk markers. The association with higher numbers of HVF tests may reflect surveillance intensity or confounding by indication, where patients perceived to be at higher risk are tested more often. Shorter follow-up duration being associated with progression may seem counterintuitive, but it can arise when eyes with more rapid decline accumulate sufficient signal to meet slope criteria earlier. The link with unmet social needs is particularly notable because it points beyond biology and toward structural determinants of vision outcomes.
Matched case-control sensitivity analysis
The most clinically provocative findings emerged in the matched analysis among patients with baseline MD 60 days was associated with significantly higher odds of visual field progression compared with a mean DFU of ≤30 days, with an adjusted odds ratio of 1.95 and a 95% confidence interval of 1.03 to 3.68. A maximum DFU of >365 days showed a borderline association compared with ≤120 days, with an adjusted odds ratio of 1.77 and a 95% confidence interval of 0.97 to 3.23.
These results do not overturn the main negative finding, but they refine it. They suggest that while modest or occasional delays may not materially affect most patients, longer average delays could still matter in eyes with more advanced baseline functional loss. That pattern is consistent with current clinical intuition: patients with less reserve have less room for undetected deterioration before meaningful disability occurs.
Clinical Interpretation
The study’s overall message is not that follow-up timing is unimportant. Rather, it suggests that the relationship between delayed visits and measurable functional progression may be more nuanced than many clinicians assume. Several interpretations are plausible.
First, current prescribed intervals may be conservative for many low-risk or stable patients, particularly in specialty settings where clinicians already individualize monitoring based on perceived risk. In that scenario, a modest overrun beyond the recommended date may not be sufficient to change outcomes.
Second, treatment changes and disease trajectory may be driven more by intraocular pressure control, medication adherence, baseline structural damage, and intrinsic disease aggressiveness than by the exact spacing of visits, at least within the range of delays observed here. If a patient’s pressure is well controlled and disease is stable, a somewhat later visit may have limited consequence.
Third, the study raises the possibility that the harm of delay is concentrated in specific subgroups rather than distributed evenly across all patients. The sensitivity analysis supports this possibility for patients with baseline MD worse than -6 dB. In practice, this would favor a risk-stratified scheduling model rather than a uniformly rigid timetable.
Strengths
This study has several notable strengths. The sample size was substantial, with more than 1100 eyes and a mean follow-up exceeding 7 years. The use of repeated HVF measurements over time provides a clinically meaningful functional endpoint. The exposure definition reflected provider-intended care intervals, which is more relevant than using arbitrary calendar thresholds alone. The authors also examined multiple delay metrics and performed a matched case-control sensitivity analysis to test robustness.
Equally important, the study reflects real-world practice rather than idealized trial conditions. That makes the findings especially relevant to health systems trying to design follow-up pathways under capacity constraints.
Limitations
As with any retrospective observational study, residual confounding remains a major concern. Patients with delayed follow-up may differ in important ways from those seen on time, including adherence, frailty, access barriers, treatment intensity, and disease stability. Some confounding may bias toward the null if clinicians intentionally schedule lower-risk patients less urgently or tolerate modest delays in those thought to be stable.
The analysis was performed at a tertiary center, which may limit generalizability to community practice, safety-net settings, or populations with different access barriers. The requirement for at least 5 HVF tests in the main cohort may also select for patients more engaged in care.
The outcome relied on MD slope, which is practical but imperfect. Visual field testing is noisy, progression may be nonlinear, and global MD can miss localized changes. Structural progression on optical coherence tomography was not part of the primary outcome presented in the abstract, so some clinically relevant worsening may have gone undetected.
The exposure definition depends on documented recommended follow-up intervals, which may not always capture informal clinician judgment or evolving plans. Finally, because the study evaluated associations rather than randomized scheduling strategies, it cannot establish that longer intervals are safe; it can only suggest that observed delays were not consistently linked to worse functional trajectories in this dataset.
How This Fits With Current Glaucoma Care
Current glaucoma guidelines generally recommend individualized follow-up based on severity, rate of progression, intraocular pressure, treatment complexity, and risk of vision-related disability. That principle is echoed by the present findings. The study supports the idea that follow-up intensity should be tailored rather than standardized.
For stable glaucoma suspects or patients with mild, well-controlled disease, occasional delays beyond a recommended appointment date may not necessarily imply substantial harm. For patients with worse baseline visual field loss, however, the sensitivity analysis argues for greater caution. In these individuals, prolonged average delays may increase the chance that progression goes undetected until after additional irreversible loss has occurred.
The association between unmet social needs and progression is especially relevant for clinicians and health systems. It suggests that preserving vision may depend not only on disease metrics but also on whether patients can reliably access care. Risk-based scheduling should therefore incorporate social as well as biological vulnerability.
Practice Implications
Several practical lessons emerge. First, clinics may be able to reserve the most time-sensitive slots for patients with more advanced disease, recent progression, uncontrolled pressure, or substantial social barriers. Second, lower-risk patients might be candidates for more flexible scheduling without major compromise in visual field outcomes, though decisions should remain individualized. Third, tracking unmet social needs may help identify patients at greater risk for progression independent of traditional ophthalmic markers.
For researchers, the next step is prospective validation. Pragmatic studies comparing risk-stratified follow-up intervals, ideally incorporating both structural and functional outcomes, would be especially useful. It would also be valuable to distinguish patient-driven no-shows from system-driven rescheduling and to examine whether teleophthalmology, home monitoring, or technician-led testing can offset the effects of delayed physician review.
Conclusion
Kolli and colleagues provide timely real-world evidence that delayed glaucoma follow-up, although common, was not broadly associated with visual field progression in a large tertiary-care cohort. The data challenge the assumption that every departure from the prescribed schedule carries measurable functional harm. At the same time, the matched case-control analysis suggests a more cautious interpretation for eyes with baseline MD worse than -6 dB, where longer average delays may be associated with higher odds of progression.
The study therefore supports a risk-stratified approach to glaucoma surveillance. For many patients, modest delays may be tolerable. For those with more advanced disease or social vulnerability, timely follow-up remains a priority. The central clinical message is not to relax monitoring indiscriminately, but to target intensity where it is most likely to preserve vision.
Funding and Trial Registration
No funding information or ClinicalTrials.gov registration number was provided in the abstract. As this was a retrospective observational study, trial registration may not have been applicable.
References
1. Kolli A, Chen E, Xu X, Lu J, Takla P, Liang J, Huh M, Zhao Y, Lee J, Nguyen K, Liang E, Yu Y, McLeod SD, Stein JD, Gedde SJ, Aung T, Li F, Ramulu P, Ying GS, Han Y. Visual Field Progression in Glaucoma Patients with Delayed Follow-Up. American Journal of Ophthalmology. 2026-05-13. PMID: 42134617.
2. American Academy of Ophthalmology Preferred Practice Pattern. Primary Open-Angle Glaucoma. San Francisco, CA: American Academy of Ophthalmology. Most recent online update available through AAO PPP resources.
3. European Glaucoma Society Terminology and Guidelines for Glaucoma, 5th Edition. British Journal of Ophthalmology. 2021;105(Suppl 1):1-169.
