When Does Size Matter? Promises, Pitfalls, and Appropriate Interpretation of "Big" Medical Records Data

Application of big data in ophthalmology

The advents of information technologies have led to the creation of ever-larger datasets. Also known as big data, these large datasets are characterized by its volume, variety, velocity, veracity, and value. More importantly, big data has the potential to expand traditional research capabilities, inform clinical practice based on real-world data, and improve the health system and service delivery. This review first identified the different sources of big data in ophthalmology, including electronic medical records, data registries, research consortia, administrative databases, and biobanks. Then, we provided an in-depth look at how big data analytics have been applied in ophthalmology for disease surveillance, and evaluation on disease associations, detection, management, and prognostication. Finally, we discussed the challenges involved in big data analytics, such as data suitability and quality, data security, and analytical methodologies.

Characterization of poor visual outcomes of diabetic macular edema: the Fight Retinal Blindness! Project

PURPOSE

To investigate the incidence, characteristics and baseline predictors of poor visual outcomes in eyes with diabetic macular edema (DME) receiving intravitreal therapy in routine clinical practice.

DESIGN

Observational study.

PARTICIPANTS

Treatment-naive eyes starting intravitreal therapy for DME between 2014 and 2018 tracked in the Fight Retinal Blindness! registry. We examined two groups with poor visual outcomes: 1) Those with sustained vision loss of >10 letters from baseline without recovery of visual acuity (VA) or 2) Those with VA<55 letters at 2 years. Respective controls were eyes that did not experience poor visual outcomes.

METHODS

Kaplan-Meier curves analyzed proportion of eyes that experienced poor outcomes. Cox proportional hazards models evaluated potential baseline predictors of poor outcomes.

MAIN OUTCOME MEASURES

The proportion of eyes that experienced poor visual outcomes within 2 years of treatment initiation and its baseline predictors.

RESULTS

The proportion of eyes with sustained VA>10 letter loss was 14% at 2 years while 16% of eyes had VA< 55 letters 2 years after starting intravitreal therapy. Initial treatment with intravitreal corticosteroid was independently associated with higher incidence of >10 letter loss was (Hazard ratio [HR], 3.21; 95% confidence interval [CI], 1.60-6.44; P< 0.01). No improvement in VA 3 months after starting treatment was associated with >10 letter loss (HR, 6.81; 95% CI, 4.11-11.27; P <0.01) and VA<55 letters at 2 years (HR, 4.28; 95% CI, 2.66- 6.89; P <0.01). The other factors related to higher risk of VA<55 letters were older age (HR, 1.02 per year; 95% CI, 1-1.04; P = 0.04) and poor baseline VA (HR, 0.68 per 5 letters; 95% CI, 0.65- 0.72, P <0.001).

CONCLUSION

Fourteen percent of eyes managed with intravitreal therapy in routine clinical care experienced >10 letter loss and 16% had VA<55 letters 2 years after starting treatment for DME. Identification of the incidence and predictors of poor outcomes provides more accurate assessment of the potential benefit from intravitreal therapy.

Predictors of progression in untreated keratoconus: a Save Sight Keratoconus Registry study

AIMS

We set out to identify risk factors for progression in untreated keratoconus patients from 34 centres across Australia, New Zealand, Spain and Italy.

METHODS

Patients were divided into 'progressors' and 'stable' patients for each clinical parameter: visual acuity (VA), steepest keratometry (maximum keratometry (Max-K)) and thinnest corneal thickness (TCT). Primary outcomes were the proportion of eyes with sustained progression in VA, Max-K or TCT within 3 years. Secondary outcomes included predictors of progression.

RESULTS

There were 3994 untreated eyes from 2283 patients. The proportion of eyes with VA, Max-K and TCT progression at 1 year were 3.2%, 6.6% and 3.1% respectively. Factors associated with VA loss were higher baseline VA (HR 1.15 per logMAR line increase in VA; p<0.001) and steeper baseline Max-K (HR 1.07 per 1D increase; p<0.001). Younger baseline age was associated with Max-K steepening (HR 0.96 per year older; p=0.001). Thicker baseline TCT, steeper baseline Max-K and younger baseline age were associated with TCT thinning: (HR 1.08 per 10 µm increase in TCT; p<0.001), (HR 1.03 per 1D increase; p=0.02) and (HR 0.98 per year younger; p=0.01), respectively.

CONCLUSIONS

Steeper Max-K and younger age were the most clinically useful baseline predictors of progression as they were associated with worsening of two clinical parameters. Every 1D steeper Max-K was associated with a 7% and 3% greater risk of worsening VA and thinning TCT, respectively. Each 1 year younger was associated with a 4% and 2% greater risk of steepening Max-K and thinning TCT, respectively.

Big Data in Ophthalmology

Big data is the fuel of mankind's fourth industrial revolution. Coupled with new technology such as artificial intelligence and deep learning, the potential of big data is poised to be harnessed to its maximal in years to come. In ophthalmology, given the data-intensive nature of this specialty, big data will similarly play an important role. Electronic medical records, administrative and health insurance databases, mega national biobanks, crowd source data from mobile applications and social media, and international epidemiology consortia are emerging forms of "big data" in ophthalmology. In this review, we discuss the characteristics of big data, its potential applications in ophthalmology, and the challenges in leveraging and using these data. Importantly, in the next phase of work, it will be pertinent to further translate "big data" findings into real-world applications, to improve quality of eye care, and cost-effectiveness and efficiency of health services in ophthalmology.

Japan-Retinal Detachment Registry Report I: preoperative findings in eyes with primary retinal detachment

PURPOSE

To report the demographics and clinical characteristics of patients with a primary retinal detachment (RD).

DESIGN

Prospective cohort study by a registry design.

PARTICIPANTS

Patients with RD treated at vitreoretinal sub-specialty institutions in Japan from February 2016 to March 2017.

METHODS

Descriptive statistics for the primary RD, and multivariable ordered logistic regression and multiple linear regression analyses were performed.

RESULTS

3178 eyes of 3178 cases were analyzed. The interval from onset to surgery was significantly shorter in patients in the 40-year age group than in other age groups except for the 50-year age group (P<0.05, Steel-Dwass test). The proportion of complex cases was significantly higher in the 10-year, 70-year, and 80+ year age groups than in the 40 and 50-year age groups (P<0.05, Steel-Dwass test). The size of RD was significantly associated with the male sex (odds ratio, 1.29; 95% confidence interval [CI], 1.07 to 1.56; P=0.0085) and the interval from onset to surgery (odds ratio, 1.03 95% CI, 1.01 to 1.04; P=0.0014). Low IOPs in eyes with RD were significantly associated with an older age (-0.24 mmHg/10 years, 95% CI, -0.32 to -0.16], P<0.0001) and larger RD area (-0.91 mmHg/quadrant, 95% CI, [-1.06 to -0.76], P <0.0001).

CONCLUSION

Profile and clinical characteristics of patients with a primary RD were not exactly the same as previous reports. A preoperative low IOP was associated with several ocular factors while the area of RD was associated not only with ocular but with social factors as well.

Deep Anterior Lamellar Keratoplasty for Keratoconus: Multisurgeon Results

PURPOSE

To examine clinical outcomes in deep anterior lamellar keratoplasty (DALK) for keratoconus using contemporary techniques in a multisurgeon public healthcare setting.

DESIGN

Consecutive, retrospective case series.

METHODS

Setting: Moorfields Eye Hospital, London, United Kingdom.

STUDY POPULATION

Consecutive cases of keratoconus treated with non-laser assisted DALK from September 1, 2012, to September 31, 2016.

OBSERVATION PROCEDURE

Data on preoperative status, operative details, intraoperative and postoperative complications, secondary interventions, and visual outcomes were archived for analysis.

MAIN OUTCOME MEASURES

Graft failure rate and percentage of patients with corrected distance visual acuity (CDVA) ≥20/40 within 1 year of surgery and at final review after suture removal.

RESULTS

Three hundred fifty-seven eyes of 338 patients undergoing DALK (91.3% big-bubble technique attempted) were analyzed. A total of 4.2% (95% confidence interval [CI] 2.4%-6.8%) of corneal transplants had failed within the follow-up period (21.8 ± 11.4 months), and 75.9% of eyes had CDVA ≥20/40 within 1 year of surgery, rising to 81% after suture removal. Forty-two primary surgeons (31 trainees) participated. Intraoperative perforation of Descemet membrane occurred in 45.4% of eyes. A total of 24.1% were converted to penetrating keratoplasty (PK) intraoperatively. Conversion to PK increased the risk of transplant rejection (P = .026; odds ratio [OR] 1.94; 95% CI 1.1-3.5) and secondary glaucoma (P = .016; OR 4.0; 95% CI 1.3-12.4). Transplant rejection increased the risk of graft failure both overall (P = .017; OR = 3.9; 95% CI 1.4-11.0) and when cases converted to PK were excluded (P = .028; OR = 3.35; 95% CI 1.1-9.9).

CONCLUSION

DALK for keratoconus achieves early results similar to those published for PK in a multisurgeon setting. Conservative management of intraoperative Descemet membrane perforation, where possible, may be safer than conversion to PK.