Predictive modeling of risk factors and complications of cataract surgery

Adherence of studies involving artificial intelligence in the analysis of ophthalmology electronic medical records to AI-specific items from the CONSORT-AI guideline: a systematic review

PURPOSE

In the context of ophthalmologic practice, there has been a rapid increase in the amount of data collected using electronic health records (EHR). Artificial intelligence (AI) offers a promising means of centralizing data collection and analysis, but to date, most AI algorithms have only been applied to analyzing image data in ophthalmologic practice. In this review we aimed to characterize the use of AI in the analysis of EHR, and to critically appraise the adherence of each included study to the CONSORT-AI reporting guideline.

METHODS

A comprehensive search of three relevant databases (MEDLINE, EMBASE, and Cochrane Library) from January 2010 to February 2023 was conducted. The included studies were evaluated for reporting quality based on the AI-specific items from the CONSORT-AI reporting guideline.

RESULTS

Of the 4,968 articles identified by our search, 89 studies met all inclusion criteria and were included in this review. Most of the studies utilized AI for ocular disease prediction (n = 41, 46.1%), and diabetic retinopathy was the most studied ocular pathology (n = 19, 21.3%). The overall mean CONSORT-AI score across the 14 measured items was 12.1 (range 8-14, median 12). Categories with the lowest adherence rates were: describing handling of poor quality data (48.3%), specifying participant inclusion and exclusion criteria (56.2%), and detailing access to the AI intervention or its code, including any restrictions (62.9%).

CONCLUSIONS

In conclusion, we have identified that AI is prominently being used for disease prediction in ophthalmology clinics, however these algorithms are limited by their lack of generalizability and cross-center reproducibility. A standardized framework for AI reporting should be developed, to improve AI applications in the management of ocular disease and ophthalmology decision making.

Risk factors for complications in resident-performed cataract surgery: A systematic review

We assessed risk factors for complications associated with resident-performed cataract surgery. Using the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines, we searched 4databases in September, 2023. We included peer-reviewed, full-text, English-language articles assessing risk factors for complications in resident performed cataract surgery. We excluded studies describing cataract surgeries performed by fellows, combined surgeries, and studies with insufficient information. Our initial search yielded 6244 articles; 15 articles were included after title/abstract and full-text review. Patient-related risk factors included older age, hypertension, prior vitrectomy, zonular pathology, pseudoexfoliation, poor preoperative visual acuity, small pupils, and selected types of cataracts. Surgeon-related risk factors included resident postgraduate year and surgeon right-handedness. Other risk factors included absence of supervision, long phacoemulsification time, and phacoemulsification with high power and torsion. The quality of the studies was assessed using the Grading of Recommendations, Assessment, Development, and Evaluation; most studies graded as moderate, primarily due to risk of bias. When assigning cases to residents, graduate medical educators should consider general and resident-specific risk factors to facilitate teaching and preserve patient safety.

Predicting Glaucoma Surgical Outcomes Using Neural Networks and Machine Learning on Electronic Health Records

Purpose

To develop machine learning (ML) and deep learning (DL) models to predict glaucoma surgical outcomes, including postoperative intraocular pressure, use of ocular antihypertensive medications, and need for repeat surgery.

Methods

We identified glaucoma surgeries performed at Stanford from 2013-2024, with two or more postoperative visits with intraocular pressure (IOP) measurement. Patient features were identified from the electronic health record (EHR), including demographics, prior diagnosis and procedure codes, medications and eye exam findings. Classical ML and DL models were developed to predict which glaucoma surgeries would result in surgical failure, defined as (1) IOP not reduced by more than 20% of preoperative baseline on two consecutive postoperative visits, (2) increased classes of glaucoma medications, and (3) need for additional glaucoma surgery or revision of original surgery.

Results

A total of 2398 glaucoma surgeries of 1571 patients were included, of which 1677 surgeries met failure criteria. Random forest performed best for prediction of overall surgical failure, with accuracy of 75.5% and area under the receiver operator curve (AUROC) of 76.7%, similar to the deep learning model (accuracy 75.5%, AUROC 76.6%). Across all models, prediction performance was better for IOP outcomes (AUROC 86%) than need for an additional surgery (AUROC 76%) or need for additional glaucoma medication (AUC 70%).

Conclusions

ML and DL algorithms can predict glaucoma surgery outcomes using structured data inputs from EHRs.

Translational Relevance

Models that predict outcomes of glaucoma surgery may one day provide the basis for clinical decision support tools supporting surgeons in personalizing glaucoma treatment plans.

LncRNA MEG3 regulates ferroptosis of lens epithelial cells via PTBP1/GPX4 axis to participate in age-related cataract

Age-related cataract (ARC) is regarded as the principal cause of vision impairment among the aged. The regulatory role of long noncoding RNAs (LncRNAs) in ARC remains unclear. The lncRNA maternally expressed gene 3 (MEG3) has been reported to promote ARC progression, and the underlying mechanism was further investigated in this study. Lens epithelium samples were collected to verify the expression of MEG3. Lens epithelial cells (LECs) were treated with H2O2 to mimic microenvironment of ARC in vitro. Cell viability, reactive oxygen species, and ferroptosis were evaluated during the in viro experiments. In the present work, lncRNA MEG3 was highly expressed in ARC group, compared with normal group. MEG3 was induced, cell viability and glutathione peroxidase 4 (GPX4) level were inhibited, and ferroptosis was promoted in H2O2 treated LECs. LncRNA MEG3 silence reversed the effects of H2O2 on viability and ferroptosis in LECs. Thereafter, lncRNA MEG3 was found to bind to PTBP1 for GPX4 degradation. Silencing of GPX4 reversed the regulation of lncRNA MEG3 inhibition in H2O2-treated LECs. To sum up, lncRNA MEG3 exhibited high expression in ARC. In H2O2-induced LECs, inhibition of lncRNA MEG3 accelerated cell viability and repressed ferroptosis by interaction with PTBP1 for GPX4 messenger RNA decay. Targeting lncRNA MEG3 may be a novel treatment of ARC.

Demographic and Systemic Risk Factors for Persistent Corneal Edema Following Cataract Surgery in Patients With and Without Diabetes

PURPOSE

To identify risk factors associated with the development of corneal edema (CE) and the need for corneal transplantation following cataract surgery.

DESIGN

Retrospective cohort study.

METHODS SETTING

Nation-wide sample of Medicare beneficiaries from 2011-2015.

STUDY POPULATION

Medicare beneficiaries aged over 65 years who received cataract surgery between 2011-2014 with at least 1 year of continuous follow-up. Data was retrieved from the Denominator and Physician Supplier Part B file from the Center for Medicare and Medicaid Services.

MAIN OUTCOME(S) AND MEASURE(S)

The main outcome was the association between demographic characteristics (e.g., age, sex, race/ethnicity) and systemic factors including diabetes status, hypertension, and tobacco use on the incidence of CE and the subsequent need for corneal transplantation following cataract surgery.

RESULTS

Among 187,746 beneficiaries, 67,734 had diabetes and 120,012 did not. Beneficiaries with diabetes were more likely to develop CE compared to those without (Odds ratio [OR] 1.19, 95% Confidence Interval [CI] [1.02-1.40]). Compared to those aged 65-74, beneficiaries aged 75-84 and over 85 were more likely to develop CE (OR 1.29 [1.09-1.52]) and OR 1.96 [1.55-2.46], respectively). Asian (OR 2.42 [1.66-3.40]), Hispanic (OR 2.60 [1.73-3.74]), and North American Native (OR 3.59 [1.78-6.39]) race was associated with increased likelihood of developing CE. North American Native beneficiaries had higher risk of requiring corneal transplantation compared to White beneficiaries (OR 9.30 [2.26-25.31]). Female sex decreased likelihood of requiring corneal transplantation post-operatively (OR 0.56 [0.36-0.87]). Amongst those with diabetes, the presence of proliferative diabetic retinopathy increased the likelihood of developing CE (OR 1.94 [1.05-3.39]).

CONCLUSION

Older age, diabetes, and non-White race elevate the risk of CE following cataract surgery, with race incurring the highest risk. Further research is needed to understand the factors underlying the significantly increased risk of CE in racial and ethnic minorities within the United States.

Disease progression and clinical outcomes in latent osteoarthritis phenotypes: Data from the Osteoarthritis Initiative

The prevalence of knee osteoarthritis (OA) is widespread and the heterogeneous patient factors and clinical symptoms in OA patients impede developing personalized treatments for OA patients. In this study, we used unsupervised and supervised machine learning to organize the heterogeneity in knee OA patients and predict disease progression in individuals from the Osteoarthritis Initiative (OAI) dataset. We identified four distinct knee OA phenotypes using unsupervised learning that were defined by nutrition, disability, stiffness, and pain (knee and back) and were strongly related to disease fate. Interestingly, the absence of supplemental vitamins from an individual's diet was protective from disease progression. Moreover, we established a phenotyping tool and prognostic model from 5 variables (WOMAC disability score of the right knee, WOMAC total score of the right knee, WOMAC total score of the left knee, supplemental vitamins and minerals frequency, and antioxidant combination multivitamins frequency) that can be utilized in clinical practice to determine the risk of knee OA progression in individual patients. We also developed a prognostic model to estimate the risk for total knee replacement and provide suggestions for modifiable variables to improve long-term knee health. This combination of unsupervised and supervised data-driven tools provides a framework to identify knee OA phenotype in a clinical scenario and personalize treatment strategies.

Are there sex-based disparities in cataract surgery?

AIM

To investigate sex-based differences in the occurrence of intra-operative and post-operative complications and associated visual outcomes following cataract surgery.

METHODS

This was a retrospective study of patients who had phacoemulsification cataract surgery at the University of Colorado School of Medicine. Data collected included the patient's health history, ocular comorbidities, operative and post-operative complications, and the post-operative best corrected visual acuity (BCVA). The data were analyzed using univariate and multivariable logistic regression with generalized estimating equations to account for the correlation of some patients having two eyes included in the study.

RESULTS

A total of 11 977 eyes from 7253 patients were included in the study. Ocular comorbidities differed by sex, with males having significantly higher percentages of traumatic cataracts (males 0.7% vs females 0.1%), prior ocular surgery (6.7% vs 5.5%), and mature cataracts (2.8% vs 1.9%). Conversely, females had significantly higher rates of pseudoexfoliation (2.0% vs 3.2%). In unadjusted analysis, males had higher rates of posterior capsular rupture (0.8% vs 0.4%) and vitreous loss (1.0% vs 0.6%), but this difference was not significant after adjustment for confounders. Males had a significantly increased risk of post-operative retinal detachment, but in multivariable analysis this was no longer significant. Males were significantly less likely to undergo post-operative neodymium-doped yttrium aluminum garnet (Nd:YAG) laser capsulotomy for posterior capsule opacification (OR=0.8, 95%CI=0.7-0.9, P=0.0005). The BCVA was slightly worse for males pre-operatively; but post-operatively, both sexes exhibited similar visual acuity of Snellen equivalent 20/25.

CONCLUSION

The study finds that in a cohort of patients presenting for cataract surgery, sex differences exist in pre-operative comorbidities and surgical characteristics that contribute to higher rates of some complications for males. However, observed surgical complication rates exhibit almost no difference by sex after adjusting for pre-operative differences and post-operative BCVA is similar between sexes.

Disease progression and clinical outcomes in latent osteoarthritis phenotypes: Data from the Osteoarthritis Initiative

The prevalence of knee osteoarthritis (OA) is widespread and the heterogeneous patient factors and clinical symptoms in OA patients impede developing personalized treatments for OA patients. In this study, we used unsupervised and supervised machine learning to organize the heterogeneity in knee OA patients and predict disease progression in individuals from the Osteoarthritis Initiative (OAI) dataset. We identified four distinct knee OA phenotypes using unsupervised learning that were defined by nutrition, disability, stiffness, and pain (knee and back) and were strongly related to disease fate. Interestingly, the absence of supplemental vitamins from an individual's diet was protective from disease progression. Moreover, we established a phenotyping tool and prognostic model from 5 variables (WOMAC disability score of the right knee, WOMAC total score of the right knee, WOMAC total score of the left knee, supplemental vitamins and minerals frequency, and antioxidant combination multivitamins frequency) that can be utilized in clinical practice to determine the risk of knee OA progression in individual patients. We also developed a prognostic model to estimate the risk for total knee replacement and provide suggestions for modifiable variables to improve long-term knee health. This combination of unsupervised and supervised data-driven tools provides a framework to identify knee OA phenotype in a clinical scenario and personalize treatment strategies.

Outcomes and Risk Factors for Complications in Cataract Patients with Hepatitis C Virus Infection

PURPOSE

To describe outcomes of patients with hepatitis C virus (HCV) seropositivity undergoing cataract surgery, as well as investigate risk factors for surgical complications.

METHODS

This is a retrospective cohort study of all consecutive patients who underwent cataract surgery at a tertiary care hospital in the United States between 2014 and 2019. The exposure of interest was HCV seropositivity and outcomes included surgical complications and associated risk factors, visual acuity, and post-operative complications.

RESULTS

A total of 11,276 eyes of 6,858 patients were included in the study, of which 122 patients (1.78%) and 210 eyes (1.86%) were HCV positive. Average age at surgery was 63.4 (8.4) years for HCV positive patients and 69.1 (10.6) years for HCV negative patients. Patients with HCV were more likely to suffer a complication during cataract surgery, 2.9% versus 1.2% (OR 2.27, 95% CI 1.03 to 5.01, p = .0415). Postoperative best corrected visual acuity was excellent: median and range 0.00 (-0.13, 3.00) logMAR for HCV positive eyes versus 0.00 (-0.30, 3.00) logMAR for HCV negative eyes. Among HCV positive patients, elevated alanine transaminase (>52 U/L) was associated with a higher intraoperative complication rate (10.0% vs 1.8%, OR 5.53, 95% CI 1.05 to 29.2, p = .044).

CONCLUSION

While patients with HCV are more likely to have complications during cataract surgery, final best corrected visual acuity was excellent regardless of HCV status. Patients with HCV are more likely to undergo cataract surgery at a younger age, and those with elevated alanine transaminase are at highest risk for complications.

Experimental study comparing 2 different phacoemulsification systems with intraocular pressure control during steady-state flow and occlusion break surge events

PURPOSE

To compare peak surge and surge duration time after occlusion break, incision leakage compensation, and passive vacuum for 2 phacoemulsification systems.

SETTING

Carl Zeiss Meditec AG, Oberkochen, Germany.

DESIGN

Laboratory study.

METHODS

A spring-eye model was used to test Alcon Centurion Vision and Zeiss Quatera 700 systems. Peak surge and duration was measured after an occlusion break. Quatera tested in flow and vacuum priority modes. Vacuum limits ranged from 300 to 700 mm Hg with intraocular pressure (IOP) set at 30 mm Hg, 55 mm Hg, and 80 mm Hg. IOP vs incision leakage rates of 0 to 15 cc/min and passive vacuum were measured.

RESULTS

At 30 mm Hg IOP set point and vacuum limits ranging 300 to 700 mm Hg, the surge duration time after occlusion break ranged 419 to 1740 milliseconds (ms) for Centurion, 284 to 408 ms for Quatera in the flow mode, and 282 to 354 ms for Quatera in the vacuum mode. At 55 mm Hg, values ranged 268 to 1590 ms for Centurion, 258 to 471 ms for Quatera in the flow mode, and 239 to 284 ms for Quatera in the vacuum mode. At 80 mm Hg, values were 243 to 1520 ms for Centurion, 238 to 314 ms for Quatera in the flow mode, and 221 to 279 ms in the vacuum mode. Centurion exhibited slightly less peak surge than the Quatera. At 55 mm Hg: incision leakage rates 0 to 15 cc/min, Quatera held the IOP within ±2 mm Hg of target; Centurion was unable to hold IOP target allowing a 11.7 mm Hg decrease with 32% higher passive vacuum.

CONCLUSIONS

Quatera demonstrated slightly higher surge peak values and notably shorter surge duration times after occlusion break than Centurion. Quatera demonstrated better incision leakage compensation and lower passive vacuum than Centurion.

Socioeconomic status, mental health, and nutrition are the principal traits for low back pain phenotyping: Data from the osteoarthritis initiative

Background

Low back pain (LBP) is a heterogeneous disease with biological, physical, and psychosocial etiologies. Models for predicting LBP severity and chronicity have not made a clinical impact, perhaps due to difficulty deciphering multidimensional phenotypes. In this study, our objective was to develop a computational framework to comprehensively screen metrics related to LBP severity and chronicity and identify the most influential.

Methods

We identified individuals from the observational, longitudinal Osteoarthritis Initiative cohort (N = 4796) who reported LBP at enrollment (N = 215). OAI descriptor variables (N = 1190) were used to cluster individuals via unsupervised learning and uncover latent LBP phenotypes. We also developed a dimensionality reduction algorithm to visualize clusters/phenotypes using Uniform Manifold Approximation and Projection (UMAP). Next, to predict chronicity, we identified those with acute LBP (N = 40) and persistent LBP over 8 years of follow-up (N = 66) and built logistic regression and supervised machine learning models.

Results

We identified three LBP phenotypes: a "high socioeconomic status, low pain severity group", a "low socioeconomic status, high pain severity group", and an intermediate group. Mental health and nutrition were also key clustering variables, while traditional biomedical factors (e.g., age, sex, BMI) were not. Those who developed chronic LBP were differentiated by higher pain interference and lower alcohol consumption (a correlate to poor physical fitness and lower soceioeconomic status). All models for predicting chronicity had satisfactory performance (accuracy 76%-78%).

Conclusions

We developed a computational pipeline capable of screening hundreds of variables and visualizing LBP cohorts. We found that socioeconomic status, mental health, nutrition, and pain interference were more influential in LBP than traditional biomedical descriptors like age, sex, and BMI.

Ophthalmology Operation Note Encoding with Open-Source Machine Learning and Natural Language Processing

INTRODUCTION

Accurate assignment of procedural codes has important medico-legal, academic, and economic purposes for healthcare providers. Procedural coding requires accurate documentation and exhaustive manual labour to interpret complex operation notes. Ophthalmology operation notes are highly specialised making the process time-consuming and challenging to implement. This study aimed to develop natural language processing (NLP) models trained by medical professionals to assign procedural codes based on the surgical report. The automation and accuracy of these models can reduce burden on healthcare providers and generate reimbursements that reflect the operation performed.

METHODS

A retrospective analysis of ophthalmological operation notes from two metropolitan hospitals over a 12-month period was conducted. Procedural codes according to the Medicare Benefits Schedule (MBS) were applied. XGBoost, decision tree, Bidirectional Encoder Representations from Transformers (BERT) and logistic regression models were developed for classification experiments. Experiments involved both multi-label and binary classification, and the best performing model was used on the holdout test dataset.

RESULTS

There were 1,000 operation notes included in the study. Following manual review, the five most common procedures were cataract surgery (374 cases), vitrectomy (298 cases), laser therapy (149 cases), trabeculectomy (56 cases), and intravitreal injections (49 cases). Across the entire dataset, current coding was correct in 53.9% of cases. The BERT model had the highest classification accuracy (88.0%) in the multi-label classification on these five procedures. The total reimbursement achieved by the machine learning algorithm was $184,689.45 ($923.45 per case) compared with the gold standard of $214,527.50 ($1,072.64 per case).

CONCLUSION

Our study demonstrates accurate classification of ophthalmic operation notes into MBS coding categories with NLP technology. Combining human and machine-led approaches involves using NLP to screen operation notes to code procedures, with human review for further scrutiny. This technology can allow the assignment of correct MBS codes with greater accuracy. Further research and application in this area can facilitate accurate logging of unit activity, leading to reimbursements for healthcare providers. Increased accuracy of procedural coding can play an important role in training and education, study of disease epidemiology and improve research ways to optimise patient outcomes.

When should cataract surgeons seek assistance from experienced colleagues?

PURPOSE

To assess which cases should be assorted exclusively to high-volume surgeons and identify when should a cataract surgeon seek assistance from a senior colleague.

METHODS

The medical records of 2853 patients with age-related cataract were reviewed. Preoperative risk factors were documented for each case, and they were divided into surgeons who had more (> 400 surgeries/year) or less experience (< 400 surgeries/year). Ophthalmology residents were excluded from this review. The cases that involved posterior capsule rupture, dropped nucleus, zonular dehiscence and anterior capsular tear with or without vitreous loss were defined as "complicated".

RESULTS

From the 3247 cataract extraction surgeries that were reviewed, we were unable to identify any statistically significant difference in the complication rates between the two surgeon groups. In the stepwise regression analysis, both groups supported advanced age (> 85) and mature cataracts with up to fourfold odds ratios (OR). Low-volume surgeons had a fivefold OR in the presence of phacodonesis and a fourfold OR in the case of posterior polar cataract. Finally, the low- and high-volume groups had their highest complication rates in the cumulative four and five risk factors, respectively.

CONCLUSION

In the presence of advanced age, mature cataracts, phacodonesis and posterior polar cataract, the complication rates appear to be higher for the less experienced surgeons. Meticulous preoperative assessment with detailed documentation of each patient's risk factors can result in fewer complications. The medical complexity of each case can be used as indicator of whether a more experienced surgeon should perform the surgery or not.

Predictors of visual acuity improvement after phacoemulsification cataract surgery

Purpose

This study aimed to assess preoperative predictors of visual outcome after phacoemulsification cataract surgery in Jordan, a Middle Eastern country.

Methods

This was a retrospective longitudinal study of adult patients who underwent phacoemulsification cataract surgery from January 2019 to July 2021. For each patient, we included only the first operated eye. We obtained pre-operative ocular history, cataract surgery complication risk based on a predesigned score, visual acuity, best correction, and best corrected visual acuity. We recorded intraoperative complications. We also obtained postoperative best corrected visual acuity and refractive error for correction after 1–3 months.

Results

A total of 1,370 patients were included in this study, with a mean age of 66.39 (± 9.48). 48.4% of patients achieved visual acuity ≥ 0.8, and 72.7% achieved visual acuity ≥ 0.5. The mean visual acuity improvement after phacoemulsification cataract surgery was 0.33 (95% CI 0.31–0.35). In the regression model, significant predictors that affected visual acuity improvement included the presence of diabetic retinopathy, glaucoma, and complication risk factors (i.e., high-risk surgery).

Conclusion

Predictors of visual acuity improvement vary between studies. This study was conducted in a developing country; we defined predictors of visual acuity improvement. We also provided a new preoperative phacoemulsification cataract surgery complication risk score.

Automated Identification of Clinical Procedures in Free-Text Electronic Clinical Records with a Low-Code Named Entity Recognition Workflow

INTRODUCTION

 Clinical procedures are often performed in outpatient clinics without prior scheduling at the administrative level, and documentation of the procedure often occurs solely in free-text clinical electronic notes. Natural language processing (NLP), particularly named entity recognition (NER), may provide a solution to extracting procedure data from free-text electronic notes.

METHODS

 Free-text notes from outpatient ophthalmology visits were collected from the electronic clinical records at a single institution over 3 months. The Prodigy low-code annotation tool was used to create an annotation dataset and train a custom NER model for clinical procedures. Clinical procedures were extracted from the entire set of clinical notes.

RESULTS

 There were a total of 5,098 clinic notes extracted for the study period; 1,923 clinic notes were used to build the NER model, which included a total of 231 manual annotations. The NER model achieved an F-score of 0.767, a precision of 0.810, and a recall of 0.729. The most common procedures performed included intravitreal injections of therapeutic substances, removal of corneal foreign bodies, and epithelial debridement of corneal ulcers.

CONCLUSIONS

 The use of a low-code annotation software tool allows the rapid creation of a custom annotation dataset to train a NER model to identify clinical procedures stored in free-text electronic clinical notes. This enables clinicians to rapidly gather previously unidentified procedural data for quality improvement and auditing purposes. Low-code annotation tools may reduce time and coding barriers to clinician participation in NLP research.

Applications of natural language processing in ophthalmology: present and future

Advances in technology, including novel ophthalmic imaging devices and adoption of the electronic health record (EHR), have resulted in significantly increased data available for both clinical use and research in ophthalmology. While artificial intelligence (AI) algorithms have the potential to utilize these data to transform clinical care, current applications of AI in ophthalmology have focused mostly on image-based deep learning. Unstructured free-text in the EHR represents a tremendous amount of underutilized data in big data analyses and predictive AI. Natural language processing (NLP) is a type of AI involved in processing human language that can be used to develop automated algorithms using these vast quantities of available text data. The purpose of this review was to introduce ophthalmologists to NLP by (1) reviewing current applications of NLP in ophthalmology and (2) exploring potential applications of NLP. We reviewed current literature published in Pubmed and Google Scholar for articles related to NLP and ophthalmology, and used ancestor search to expand our references. Overall, we found 19 published studies of NLP in ophthalmology. The majority of these publications (16) focused on extracting specific text such as visual acuity from free-text notes for the purposes of quantitative analysis. Other applications included: domain embedding, predictive modeling, and topic modeling. Future ophthalmic applications of NLP may also focus on developing search engines for data within free-text notes, cleaning notes, automated question-answering, and translating ophthalmology notes for other specialties or for patients, especially with a growing interest in open notes. As medicine becomes more data-oriented, NLP offers increasing opportunities to augment our ability to harness free-text data and drive innovations in healthcare delivery and treatment of ophthalmic conditions.

New meaning for NLP: the trials and tribulations of natural language processing with GPT-3 in ophthalmology

Natural language processing (NLP) is a subfield of machine intelligence focused on the interaction of human language with computer systems. NLP has recently been discussed in the mainstream media and the literature with the advent of Generative Pre-trained Transformer 3 (GPT-3), a language model capable of producing human-like text. The release of GPT-3 has also sparked renewed interest on the applicability of NLP to contemporary healthcare problems. This article provides an overview of NLP models, with a focus on GPT-3, as well as discussion of applications specific to ophthalmology. We also outline the limitations of GPT-3 and the challenges with its integration into routine ophthalmic care.

Novel technical and privacy-preserving technology for artificial intelligence in ophthalmology

PURPOSE OF REVIEW

The application of artificial intelligence (AI) in medicine and ophthalmology has experienced exponential breakthroughs in recent years in diagnosis, prognosis, and aiding clinical decision-making. The use of digital data has also heralded the need for privacy-preserving technology to protect patient confidentiality and to guard against threats such as adversarial attacks. Hence, this review aims to outline novel AI-based systems for ophthalmology use, privacy-preserving measures, potential challenges, and future directions of each.

RECENT FINDINGS

Several key AI algorithms used to improve disease detection and outcomes include: Data-driven, imagedriven, natural language processing (NLP)-driven, genomics-driven, and multimodality algorithms. However, deep learning systems are susceptible to adversarial attacks, and use of data for training models is associated with privacy concerns. Several data protection methods address these concerns in the form of blockchain technology, federated learning, and generative adversarial networks.

SUMMARY

AI-applications have vast potential to meet many eyecare needs, consequently reducing burden on scarce healthcare resources. A pertinent challenge would be to maintain data privacy and confidentiality while supporting AI endeavors, where data protection methods would need to rapidly evolve with AI technology needs. Ultimately, for AI to succeed in medicine and ophthalmology, a balance would need to be found between innovation and privacy.

The contribution of wet labs in the education of ophthalmologists

Background

The need for pre-training in experimental eye surgery is considered necessary. It is an essential way to assess trainees in ophthalmology based on their instrument and tissue handling and skills. This article aims to underline this necessity and demonstrate the ocular health professionals' opinion on this issue.

Methods

74 participants (45 females and 29 males) were included in the study. Ophthalmology residents, ophthalmologists participated in the wet lab session. The evaluation of the contribution of the wet labs were provided by filling a new questionnaire form. In this way, an interactive questionnaire was developed.

Results

Regarding trainees' grading of wet labs' significance as a first step for guiding their surgical career, it was positively correlated with their subjective view of labs' utility to both improve their surgical skills (p = 0.001) and maintain pre-existing ones (p < 0.001). We should also note that all of them (100%) stated that wet labs were necessary during residency, especially in repeated sessions, and that they would recommend them to their colleagues.

Conclusion

The surgical skills improved significantly after participation in a wet lab, according to participants, who rated the experience as highly educational. Wet labs can reduce the learning curve of difficult surgical techniques, accelerate the rate for trainees to achieve surgical competency, and treat patients safely and effectively.

•

Pre-training in experimental eye surgery is considered to be necessary.

•

Health professionals evaluated the role of wet labs by filling a new questionnaire.

•

Participants had a positive opinion on wet labs' role, especially during residency.

•

Wet labs can reduce the learning curve of difficult surgical techniques.

•

Safe and effective treatment of ophthalmology patients is more feasible.

Anti-cataract effects of coconut water in vivo and in vitro

OBJECTIVE

To determine the anti-cataract effects of coconut water (CW) in vivo and in vitro, and to explore the potential pathogenic mechanism.

METHODS

In this study, 48 male Sprague-Dawley rats were randomly divided into 4 groups: control (CO), diabetic (DM), diabetic treated with CW (DM + CW), and diabetic treated with Glibenclamide (DM + Gli). Except for the CO group, in the other three groups, intraperitoneal injection of STZ (60 mg/kg) was conducted to establish diabetic models. The experiment was conducted for 20 weeks. The slit-lamp examination was undertaken during the period of experiment (20 weeks), and then, all rats were sacrificed. The levels of superoxide dismutase (SOD), malondialdehyde (MDA), and glutathione peroxidase (GSH-Px) in the left lens were measured by using biochemical assays. The right lens was used for pathological analysis. The rat lens epithelial cells (LECs) were cultured in vitro and the subcultured cell were divided into four groups, namely the normal glucose group (5 mmol /L glucose, Group I), the high glucose group (40 mmol/L glucose, Group II), high glucose +5% CW group (Group III), and high glucose +10% CW group (Group IV). LECs were cultured under the conditions as described above for 48 h. Cell proliferation and the morphological changes were observed with interted phase contrast microscope.The level of cell apoptosis was determined by flow cytometry. the level of SOD, MDA and GSH-Px were also detected.

RESULTS

The lens opacity index decreased in diabetic rats, and LECs apoptosis ratio also decreased in high glucose environments that received CW. Under treatment with CW, reduced MDA level and elevated activities of SOD and GSH-Px were detected, both in vivo and in vitro experiments. The increased severity of cataract and LECs apoptosis were noted in diabetic rats that received normal water, while CW markedly mitigated the enhanced cataract severity and the reduction of LECs induced by diabetes mellitus.

CONCLUSION

CW is a functional food that can protect the lens from diabetic cataract. The possible underlying mechanism may be partly explained via the decreased oxidative stress in lens. However, further research needs to be conducted to indicate the pathogenic mechanism of anti-diabetic effects of CW.

Deep learning-based natural language processing in ophthalmology: applications, challenges and future directions

PURPOSE OF REVIEW

Artificial intelligence (AI) is the fourth industrial revolution in mankind's history. Natural language processing (NLP) is a type of AI that transforms human language, to one that computers can interpret and process. NLP is still in the formative stages of development in healthcare, with promising applications and potential challenges in its applications. This review provides an overview of AI-based NLP, its applications in healthcare and ophthalmology, next-generation use case, as well as potential challenges in deployment.

RECENT FINDINGS

The integration of AI-based NLP systems into existing clinical care shows considerable promise in disease screening, risk stratification, and treatment monitoring, amongst others. Stakeholder collaboration, greater public acceptance, and advancing technologies will continue to shape the NLP landscape in healthcare and ophthalmology.

SUMMARY

Healthcare has always endeavored to be patient centric and personalized. For AI-based NLP systems to become an eventual reality in larger-scale applications, it is pertinent for key stakeholders to collaborate and address potential challenges in application. Ultimately, these would enable more equitable and generalizable use of NLP systems for the betterment of healthcare and society.

Predisposing Factors for Severe Complications after Cataract Surgery: A Nationwide Population-Based Study

We conducted a retrospective group study to evaluate the potential systemic risk factors for major postoperative complications of cataract surgery. Individuals diagnosed with (n = 2046) and without (n = 8184) serious complications after cataract surgery were matched 1:4 for age, sex, and index date obtained using Taiwan’s National Health Insurance Research Database. The outcome was defined as at least one new inpatient or outpatient diagnosis of systemic disease one year before the index date. The effect of demographic data on postoperative complications was also analyzed in the multivariable model. Data were analyzed using univariate and multivariate conditional logistic regression models to calculate odds ratios (ORs) and 95% confidence intervals of the risk of developing serious complications. After the entire study interval, the major postoperative complications of cataract surgery were associated with the following systemic diseases: hypertension (adjusted OR (aOR) = 2.329, p < 0.001), diabetes mellitus (aOR = 2.818, p < 0.001), hyperlipidemia (aOR = 1.702, p < 0.001), congestive heart failure (aOR = 2.891, p < 0.001), rheumatic disease (aOR = 1.965, p < 0.001), and kidney disease needing hemodialysis (aOR = 2.942, p < 0.001). Additionally, demographic data including old age, higher urbanization level, higher level of care, and more frequent inpatient department visits were associated with a higher rate of postoperative complications. In conclusion, metabolic syndrome, chronic heart failure, end-stage renal disease, rheumatic disease, older age, and frequent inpatient department visits are correlated with the development of severe postoperative complications of cataract surgery. Therefore, cataract surgery patients should be informed about a higher possibility of postoperative complications.

Post-injection endophthalmitis rates with reduced povidone-iodine prophylaxis in patients with self-reported iodine sensitivity

BACKGROUND

Our objectives were (1) to report the post-injection endophthalmitis rate over 18 months, and (2) to determine any difference in the incidence of endophthalmitis in patients treated with reduced or no 5% povidone-iodine (PI) due to self-reported PI sensitivity.

METHODS

We performed a retrospective cohort study of all patients who received intravitreal injections (IVIs) from January 1st, 2018 to June 26th, 2019. Information on patients' age, gender visual acuities, the number of injections, drug administered, self-reported iodine sensitivity and injection protocols were obtained from electronic and paper records. For endophthalmitis cases, vitreous culture results and treatment were also noted. Patients were divided into three cohorts based on the injection protocol used for statistical analysis.

RESULTS

During the study period 22,046 IVIs were administered to 3332 eyes of 2709 patients. Intolerance to PI was reported by 2.4% of patients. The incidence of endophthalmitis was 0.02% (4/21,185) with the standard 5% PI protocol, 0.78% (6/769) with a reduced PI protocol involving fewer drops of 5% PI and chlorohexidine 0.05% for periorbital skin cleansing, and 1.09% (1/92) without any PI use. Receiving the standard PI protocol was associated with significantly lower rates of endophthalmitis compared to both the reduced PI and no PI protocols (p < 0.0001).

CONCLUSIONS

Patients who opt for less or no PI use are likely at significantly increased risk of developing post-IVI endophthalmitis. It is imperative to educate, counsel and consent these patients accordingly while exploring alternative antiseptic solutions.

Cataract Surgery Outcomes in Human Immunodeficiency Virus Positive Patients at a Tertiary Care Academic Medical Center in the United States

Purpose: To compare cataract surgery complications and visual outcomes in patients with and without human immunodeficiency virus (HIV).Methods: A retrospective cohort study was conducted on eyes undergoing phacoemulsification cataract surgery at an academic eye center from 1/1/2014 to 8/31/18. Outcomes included best corrected distance visual acuity (CDVA), intraoperative complications, cystoid macular edema (CME), and persistent anterior uveitis (PAU). Binary outcomes were analyzed using logistic regressions with generalized estimating equations. Visual outcomes were analyzed using a linear mixed model.Results: 9756 eyes from 5988 patients were included in the analysis. Of these, 66 eyes from 39 patients were HIV positive (HIV+). HIV+ patients were significantly younger at the time of surgery than HIV negative patients (p < .0001). Among HIV+ patients with available lab data, the mean CD4 count was 697.3 (SD = 335.7), and 48.7% of subjects had an undetectable viral load. Five eyes from three HIV+ patients had a history of cytomegalovirus retinitis (CMVR). Positive HIV status was not associated with increased risk of intraoperative complications. Post-operative CDVA was better in the HIV negative group compared to the HIV+ group but not significantly different (about 20/24 vs. 20/28, p = .0829). Eyes from HIV+ patients were at increased risk of developing PAU after surgery (adjusted OR = 6.04, 95% CI: 2.42-15.1, p = .0001), as well as CME (adjusted OR = 3.25, 95% CI: 1.02-10.4, p = .0470).Conclusions: Eyes from HIV+ patients were at greater risk of developing PAU and clinically significant CME; however, HIV+ patients had similar CDVA after cataract surgery compared to HIV negative patients.

Comparing cataract surgery complication rates in veterans receiving VA and community care

OBJECTIVES

To compare 90-day postoperative complication rates between Veterans receiving cataract surgery in VA vs Community Care (CC) during the first year of implementation of the Veterans Choice Act.

DATA SOURCES

Fiscal Year (FY) 2015 VA and CC outpatient data from VA's Corporate Data Warehouse (CDW) 10/01/14-9/30/15). FY14 data were used to obtain baseline clinical information prior to surgery.

STUDY DESIGN

Retrospective one-year study using secondary data to compare 90-day complication rates following cataract surgery (measured using National Quality Forum (NQF) criteria) in VA vs CC. NQF defines major complications from a specified list of Current Procedural Terminology (CPT) codes. We ran a series of logistic regression models to predict 90-day complication rates, adjusting for Veterans' sociodemographic characteristics, comorbidities, preoperative ocular conditions, eye risk group, and type of cataract surgery (classified as routine vs complex).

DATA COLLECTION

We linked VA and CC users through patient identifiers obtained from the CDW files. Our sample included all enrolled Veterans who received outpatient cataract surgery either in the VA or through CC during FY15. Cataract surgeries were identified through CPT codes 66 984 (routine) and 66 982 (complex).

PRINCIPAL FINDINGS

Of the 83,879 cataract surgeries performed in FY15, 31 percent occurred through CC. Undergoing complex surgery and having a high-risk eye (based on preoperative ocular conditions) were the strongest clinical predictors of 90-day postoperative complications. Overall, we found low complication rates, ranging from 1.1 percent in low-risk eyes to 3.6 percent in high-risk eyes. After adjustment for important confounders (eg, race, rurality, and preoperative ocular conditions), there were no statistically significant differences in 90-day complication rates between Veterans receiving cataract surgery in VA vs CC.

CONCLUSIONS

As more Veterans seek care through CC, future studies should continue to monitor quality of care across the two care settings to help inform VA's "make vs buy decisions."

Applications of Artificial Intelligence to Electronic Health Record Data in Ophthalmology

Widespread adoption of electronic health records (EHRs) has resulted in the collection of massive amounts of clinical data. In ophthalmology in particular, the volume range of data captured in EHR systems has been growing rapidly. Yet making effective secondary use of this EHR data for improving patient care and facilitating clinical decision-making has remained challenging due to the complexity and heterogeneity of these data. Artificial intelligence (AI) techniques present a promising way to analyze these multimodal data sets. While AI techniques have been extensively applied to imaging data, there are a limited number of studies employing AI techniques with clinical data from the EHR. The objective of this review is to provide an overview of different AI methods applied to EHR data in the field of ophthalmology. This literature review highlights that the secondary use of EHR data has focused on glaucoma, diabetic retinopathy, age-related macular degeneration, and cataracts with the use of AI techniques. These techniques have been used to improve ocular disease diagnosis, risk assessment, and progression prediction. Techniques such as supervised machine learning, deep learning, and natural language processing were most commonly used in the articles reviewed.

Auckland Cataract Study IV: Practical application of NZCRS cataract risk stratification to reduce phacoemulsification complications

IMPORTANCE

Reduction of intraoperative complications in phacoemulsification cataract surgery.

BACKGROUND

To assess practicability of a risk stratification system, the New Zealand Cataract Risk Stratification (NZCRS) system, in a major teaching hospital service, without investigator oversight, to ascertain whether benefits identified in research studies are maintained in busy clinical practice.

DESIGN

Prospective cohort study in a major public teaching hospital.

PARTICIPANTS

Five hundred cases of phacoemulsification cataract surgery.

METHODS

NZCRS system inserted into 621 consecutive preoperative cataract patient files. Recommendation to allocate higher-risk cases to experienced surgeons.

MAIN OUTCOME MEASURES

NZCRS system uptake and adherence, appropriate identification of high risk cases and intraoperative complication rates.

RESULTS

NZCRS scores calculated in 500 of 621 (80.5%) cases and 98 (19.6%) scored as "high risk." Cataract surgery (N = 500) performed by: 12 Registrars (20%), 4 Fellows (7.2%), 26 Consultants (72.8%). Risk scores adhered to in 99%. Overall intraoperative complications (3.0%) included iris prolapse 1.6% and posterior capsule tear 0.8%. No statistical difference in complication rates identified between surgeon grades. Mean best-corrected visual acuity was 6/10 (20/32). Postoperatively, cystoid macular oedema occurred in 3.2%. Rescoring by an experienced investigator noted a greater number of "high risk scores" (31.6% vs 19.6%) related to differences in subjective scoring of anterior chamber depth and cataract density.

CONCLUSIONS AND RELEVANCE

Practical uptake of cataract risk stratification was promising in this study with NZCRS calculated in 80.5% with 99% adherence to scoring recommendations. Compared to baseline studies, in the day-to-day clinical setting, a continued, decreasing trend in frequency and severity of intraoperative complications was noted. Subjective variability of risk scoring may be further improved by better, objective, standardization.

Natural Language Processing of Clinical Notes on Chronic Diseases: Systematic Review

Background

Novel approaches that complement and go beyond evidence-based medicine are required in the domain of chronic diseases, given the growing incidence of such conditions on the worldwide population. A promising avenue is the secondary use of electronic health records (EHRs), where patient data are analyzed to conduct clinical and translational research. Methods based on machine learning to process EHRs are resulting in improved understanding of patient clinical trajectories and chronic disease risk prediction, creating a unique opportunity to derive previously unknown clinical insights. However, a wealth of clinical histories remains locked behind clinical narratives in free-form text. Consequently, unlocking the full potential of EHR data is contingent on the development of natural language processing (NLP) methods to automatically transform clinical text into structured clinical data that can guide clinical decisions and potentially delay or prevent disease onset.

Objective

The goal of the research was to provide a comprehensive overview of the development and uptake of NLP methods applied to free-text clinical notes related to chronic diseases, including the investigation of challenges faced by NLP methodologies in understanding clinical narratives.

Methods

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed and searches were conducted in 5 databases using “clinical notes,” “natural language processing,” and “chronic disease” and their variations as keywords to maximize coverage of the articles.

Results

Of the 2652 articles considered, 106 met the inclusion criteria. Review of the included papers resulted in identification of 43 chronic diseases, which were then further classified into 10 disease categories using the International Classification of Diseases, 10th Revision. The majority of studies focused on diseases of the circulatory system (n=38) while endocrine and metabolic diseases were fewest (n=14). This was due to the structure of clinical records related to metabolic diseases, which typically contain much more structured data, compared with medical records for diseases of the circulatory system, which focus more on unstructured data and consequently have seen a stronger focus of NLP. The review has shown that there is a significant increase in the use of machine learning methods compared to rule-based approaches; however, deep learning methods remain emergent (n=3). Consequently, the majority of works focus on classification of disease phenotype with only a handful of papers addressing extraction of comorbidities from the free text or integration of clinical notes with structured data. There is a notable use of relatively simple methods, such as shallow classifiers (or combination with rule-based methods), due to the interpretability of predictions, which still represents a significant issue for more complex methods. Finally, scarcity of publicly available data may also have contributed to insufficient development of more advanced methods, such as extraction of word embeddings from clinical notes.

Conclusions

Efforts are still required to improve (1) progression of clinical NLP methods from extraction toward understanding; (2) recognition of relations among entities rather than entities in isolation; (3) temporal extraction to understand past, current, and future clinical events; (4) exploitation of alternative sources of clinical knowledge; and (5) availability of large-scale, de-identified clinical corpora.

Differences in Cataract Surgery Rates Based on Dementia Status

BACKGROUND

Cataract surgery substantially improves patient quality of life. Despite the rising prevalence of dementia in the US, little is known about use of cataract surgery among this group.

OBJECTIVE

To evaluate the relationship between dementia status and cataract surgery.

METHODS

Using administrative insurance claims for a representative sample of 1,125,387 US Medicare beneficiaries who received eye care between 2006 and 2015, we compared cataract surgery rates between patients with and without dementia via multivariable regression models to adjust for patient characteristics. Main outcome measures were annual rates of cataract surgery and hazard ratio and 95% confidence interval (CI) for receiving cataract surgery.

RESULTS

Cataract surgery was performed in 457,128 patients, 23,331 with a prior diagnosis of dementia. 16.7% of dementia patients underwent cataract surgery, compared to 43.8% of patients without dementia. 59 cataract surgeries were performed per 1000 dementia patients annually, versus 105 surgeries per 1000 nondementia patients. After adjusting for patient characteristics, dementia patients were approximately half as likely to receive cataract surgery compared to nondementia patients (adjusted HR = 0.53, 95% CI 0.53-0.54). Among the subset of patients who received a first cataract surgery, those with dementia were also less likely to receive second-eye cataract surgery (adjusted HR = 0.87, 95% CI 0.86-0.88).

CONCLUSION

US Medicare patients with dementia are less likely to undergo cataract surgery than those without dementia. This finding has implications for quality of care and dementia progression. More information is necessary to understand why rates of cataract surgery are lower for these patients, and to identify conditions where benefits of surgery may outweigh risks.

Auckland Cataract Study III: Refining Preoperative Assessment With Cataract Risk Stratification to Reduce Intraoperative Complications

PURPOSE

To assess intraoperative complications of phacoemulsification surgery in public teaching hospital settings using modified preoperative risk stratification systems.

DESIGN

Prospective cohort study.

METHODS

Preoperative risk stratification of 500 consecutive cataract cases using the New Zealand Cataract Risk Stratification (NZCRS) scoring system. Recommended allocation of higher-risk phacoemulsification procedures to experienced surgeons in public teaching hospital setting.

MAIN OUTCOME MEASURE

Intraoperative complications relative to adherence to stratification recommendations.

RESULTS

NZCRS classified 192 cases (38%) as high-risk, recommended for fellows or consultants (attendings). Primary surgeons were residents (n = 142, 28%), fellows (n = 88, 18%), and consultants (n = 270, 54%). Overall rate (N = 500) of any intraoperative complication was 5.0%. Where NZCRS scoring recommendations were observed (n = 448) the intraoperative complication rate was 4.5% but in "nonadherence" cases (n = 52 residents operating on higher-risk cases) this nearly doubled (9.6%). Postoperative complications occurred in 5.2%, primarily cystoid macular edema (3.7%). Postoperatively, mean unaided visual acuity was 6/12 (20/40) and best-corrected visual acuity improved from 6/20 (20/63) preoperatively to 6/10 (20/32) postoperatively (P < .05).

CONCLUSIONS

The NZCRS system aids identification of higher-risk cataract cases and appropriate case-to-surgeon allocation and may increase surgeon awareness of risk factors. Compared to 2 previous studies under similar conditions in the same institution, the NZCRS system was associated with a 40% reduction in intraoperative complications (8.4% to 5%). The rate of posterior capsular tear was 0.6% (P = .035) compared to 2.6% in baseline phase and 1.4% in a prior risk stratification phase. Risk stratification seems to reduce intraoperative phacoemulsification complications in public teaching hospital settings.

Surgical timing and presence of a vitreoretinal fellow on postoperative adverse events following pars plana vitrectomy

INTRODUCTION

To evaluate the adverse event rate following pars plana vitrectomy as a function of surgical start time and the presence of a vitreoretinal fellow.

METHODS

Single-institution retrospective cohort study of patients undergoing pars plana vitrectomy from 1 January 2016 to 31 December 2016 at Stanford University School of Medicine (Palo Alto, CA, USA). Records were reviewed for surgical start time, the presence of vitreoretinal fellow, and postoperative adverse events defined as any finding deviating from the expected postoperative course requiring observation or intervention.

RESULTS

A total of 310 pars plana vitrectomies were performed. There was no statistical difference in the rate of any adverse event when comparing cases starting after 16:01 (9/13, 69.2%) and after 12:01 (42/99, 42.4%) to a morning start time (69/198, 34.9%, adjusted p = 0.083). There was a statistically significant increase in the risk of postoperative vitreous hemorrhage with afternoon and evening cases as compared to morning cases (adjusted p = 0.021). In addition, there was no difference in any adverse event with a fellow present (93/244, 38.1%) compared to without (27/66, 40.9%, adjusted p = 0.163). There was a higher risk of postoperative hypotony when a fellow was involved (6.6% vs 0%, p = 0.028), though this difference disappeared after adjusting for confounders (adjusted p = 0.252). There was no difference in the length of surgery with and without a fellow (49 vs 54 min, respectively; p = 0.990).

DISCUSSION

Afternoon start time and the presence of a fellow were not independent risk factors for postoperative adverse events.

Analysis of a large data set to identify predictors of blood transfusion in primary total hip and knee arthroplasty

BACKGROUND

The aim of this study was to identify the predictors of need for allogenic blood transfusion (ALBT) in primary lower limb total joint arthroplasty (TJA).

STUDY DESIGN AND METHODS

This study utilized a large dataset of 15,187 patients undergoing primary unilateral TJA. Risk factors and demographic information were extracted from the electronic health record. A predictive model was developed by both a random forest (RF) algorithm and logistic regression (LR). The area under the receiver operating characteristic curve (AUC-ROC) was used to compare the accuracy of the two methods.

RESULTS

The rate of ALBT was 18.9% in total. Patient-related factors associated with higher risk of an ALBT included female sex, American Society of Anesthesiologists (ASA) II, ASA III, and ASA IV. Surgery-related risk factors for ALBT were operative time, drain use, and amount of intraoperative blood loss. Higher preoperative hemoglobin and tranexamic acid use were associated with decreased risk for ALBT. The RF model had a better predictive accuracy (area under the curve [AUC] 0.84) than the LR model (AUC, 0.77; p < 0.001).

CONCLUSION

The risk factors identified in the current study can provide specific, personalized perioperative ALBT risk assessment for a patient considering lower limb TJA. Furthermore, the predictive accuracy of the RF algorithm was significantly higher than that of LR, making it a potential tool for future personalized preoperative prediction of risk for perioperative ALBT.

Predictive modeling of inpatient mortality in departments of internal medicine

Despite overwhelming data on predictors of inpatient mortality, it is unclear which variables are the most instructive in predicting mortality of patients in departments of internal medicine. This study aims to identify the most informative predictors of inpatient mortality, and builds a prediction model on an individual level, given a constellation of patient characteristics. We use a penalized method for developing the prediction model by applying the least-absolute-shrinkage and selection-operator regression. We utilize a cohort of adult patients admitted to any of 5 departments of internal medicine during 3.5 years. We integrated data from electronic health records that included clinical, epidemiological, administrative, and laboratory variables. The prediction model was evaluated using the validation sample. Of 10,788 patients hospitalized during the study period, 874 (8.1%) died during admission. We find that the strongest predictors of inpatient mortality are prior admission within 3 months, malignant morbidity, serum creatinine levels, and hypoalbuminemia at hospital admission, and an admitting diagnosis of sepsis, pneumonia, malignant neoplastic disease, or cerebrovascular disease. The C-statistic of the risk prediction model is 89.4% (95% CI 88.4-90.4%). The predictive performance of this model is better than a multivariate stepwise logistic regression model. By utilizing the prediction model, the AUC for the independent (validation) data set is 85.7% (95% CI 84.1-87.3%). Using penalized regression, this prediction model identifies the most informative predictors of inpatient mortality. The model illustrates the potential value and feasibility of a tool that can aid physicians in decision-making.

Measuring outcomes in cataract surgery

PURPOSE OF REVIEW

The article presents a review of recently published studies reporting postcataract surgery outcomes and the use of electronic systems to track them.

RECENT FINDINGS

Current publications report several parameters to measure cataract outcomes such as visual acuity, patient-reported visual function, contrast sensitivity, reading speed, residual refractive errors and complications (intraoperative and postoperative).

SUMMARY

Cataracts currently afflict an estimated 94 million people worldwide, and surgical removal is the only effective therapy known. Tracking outcomes through registry databases has been shown to be a powerful tool for improving patient outcomes, understanding and adopting best clinical practices, reducing costs and increasing value delivered. Large datasets present in electronic registry systems are valuable resources for evaluating the quality of care by allowing researchers and healthcare providers to analyze, understand and adjust to 'real-world' best practices and adverse events.

The Auckland Cataract Study II: Reducing Complications by Preoperative Risk Stratification and Case Allocation in a Teaching Hospital

PURPOSE

To assess the effect of preoperative risk stratification for phacoemulsification surgery on intraoperative complications in a teaching hospital.

DESIGN

Prospective cohort study.

METHODS

Prospective assessment of consecutive phacoemulsification cases (N = 500) enabled calculation of a risk score (M-score of 0-8) using a risk stratification system. M-scores of >3 were allocated to senior surgeons. All surgeries were performed in a public teaching hospital setting, Auckland, New Zealand, in early 2016. Postoperatively, data were reviewed for complications and corrected distance visual acuity (CDVA). Results were compared to a prospective study (N = 500, phase 1) performed prior to formal introduction of risk stratification.

RESULTS

Intraoperative complications increased with increasing M-scores (P = .044). Median M-score for complicated cases was higher (P = .022). Odds ratio (OR) for a complication increased 1.269 per unit increase in M-score (95% confidence interval [CI] 1.007-1.599, P = .043). Overall rate of any intraoperative complication was 5.0%. Intraoperative complication rates decreased from 8.4% to 5.0% (OR = 0.576, P = .043) comparing phase 1 and phase 2 (formal introduction of risk stratification). The severity of complications also reduced. A significant decrease in complications for M = 0 (ie, minimal risk cases) was also identified comparing the current study (3.1%) to phase 1 (7.2%), P = .034. There was no change in postoperative complication risks (OR 0.812, P = .434) or in mean postoperative CDVA (20/30, P = .484) comparing current with phase 1 outcomes.

CONCLUSION

A simple preoperative risk stratification system, based on standard patient information gathered at preoperative consultation, appears to reduce intraoperative complications and support safer surgical training by appropriate allocation of higher-risk cases.