John Eric Jelovsek

Aims: Men with lower urinary tract symptoms (LUTS) represent a heterogeneous group, and treatment decisions are often based on severity of symptoms and physical examination findings. Identification of clinically meaningful subtypes could allow for more personalized care. This study advances phenotyping efforts from the Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) by adding data domains to previous phenotyping using urologic symptoms alone.Methods: Two‐hundred‐seventeen LUTS, demographics, medical history, and physical examination datapoints from the LURN Observational Cohort study were assessed among 519 men with at least one bothersome LUTS, using weighted Tanimoto indices, semi‐supervised learning, and resampling‐based consensus clustering to identify distinct clusters of participants. Differentially abundant serum proteins of 220 men were compared across identified clusters.Results: Five refined male clusters (RM1–RM5) were identified. Two clusters reported mild LUTS (RM1: n = 66; RM2: n = 84). RM1 was older than RM2 (70.3 vs. 56.1 years), had more comorbidities (functional comorbidity index 2.4 vs. 1.5) and erectile dysfunction. Two benign prostatic hyperplasia‐like symptom clusters were identified (RM3: n = 64; RM4: n = 188). RM3 has the largest postvoid residual volume (275 mL); RM4 reported more urinary frequency, urgency, urinary incontinence, pain, and psychosocial symptoms. RM5 (n = 119) was characterized by urgency urinary incontinence, frequency, and significant comorbidities and psychosocial symptoms. Fifteen (RM2) to 87 (RM1) differentially abundant proteins were identified within each cluster. Minimal overlap was observed between affected proteins and pathways across clusters.Conclusions: Protein signatures across newly discovered subgroups suggest identified subtypes are biochemically distinct. Findings should be validated, but may represent populations with separate pathophysiology and therapeutic needs.Clinical Trial Registration: The LURN ClinicalTrials.gov Identifier is NCT02485808.

Aims: The Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) is undertaking a new cohort study in LURN II including cases and controls.Methods: This new cohort was enrolled to specifically study urinary urgency and urgency urinary incontinence, lower urinary tract symptoms (LUTSs) that are often difficult to treat due to a lack of understanding of their phenotypes and pathophysiologies.Results: This paper will focus on the motivation for the second iteration of LURN and highlight the new research techniques and plans for more thorough phenotyping of this population.Conclusions: This paper will outline the gaps in understanding in treating LUTSs, specifically urinary urgency.

Objective: The complexity and rapid pace of development of algorithmic technologies pose challenges for their regulation and oversight in healthcare settings. We sought to improve our institution’s approach to evaluation and governance of algorithmic technologies used in clinical care and operations by creating an Implementation Guide that standardizes evaluation criteria so that local oversight is performed in an objective fashion. Materials and Methods: Building on a framework that applies key ethical and quality principles (clinical value and safety, fairness and equity, usability and adoption, transparency and accountability, and regulatory compliance), we created concrete guidelines for evaluating algorithmic technologies at our institution. Results: An Implementation Guide articulates evaluation criteria used during review of algorithmic technologies and details what evidence supports the implementation of ethical and quality principles for trustworthy health AI. Application of the processes described in the Implementation Guide can lead to algorithms that are safer as well as more effective, fair, and equitable upon implementation, as illustrated through 4 examples of technologies at different phases of the algorithmic lifecycle that underwent evaluation at our academic medical center. Discussion: By providing clear descriptions/definitions of evaluation criteria and embedding them within standardized processes, we streamlined oversight processes and educated communities using and developing algorithmic technologies within our institution. Conclusions: We developed a scalable, adaptable framework for translating principles into evaluation criteria and specific requirements that support trustworthy implementation of algorithmic technologies in patient care and healthcare operations.

Objective: To develop a model for predicting postpartum readmission for hypertension and pre‐eclampsia at delivery discharge and assess external validation or model transportability across clinical sites.Design: Prediction model using data available in the electronic health record from two clinical sites.Setting: Two tertiary care health systems from the Southern (2014–2015) and Northeastern USA (2017–2019).Population: A total of 28 201 postpartum individuals: 10 100 in the South and 18 101 in the Northeast.Methods: An internal‐external cross validation (IECV) approach was used to assess external validation or model transportability across the two sites. In IECV, data from each health system were first used to develop and internally validate a prediction model; each model was then externally validated using the other health system. Models were fit using penalised logistic regression, and accuracy was estimated using discrimination (concordance index), calibration curves and decision curves. Internal validation was performed using bootstrapping with bias‐corrected performance measures. Decision curve analysis was used to display potential cut points where the model provided net benefit for clinical decision‐making.Main outcome measures: The outcome was postpartum readmission for either hypertension or pre‐eclampsia c‐statistic South: 0.88; 95% confidence interval [CI] 0.87–0.89; Northeast: 0.74; 95% CI 0.74–0.74). In IECV, discrimination was inconsistent across sites, with improved discrimination for the Northeastern model on the Southern cohort (c‐statistic 0.61 and 0.86, respectively), but calibration was not adequate. Next, model updating was performed using the combined dataset to develop a new model. This final model had adequate discrimination (c‐statistic: 0.80, 95% CI 0.80–0.80), moderate calibration (intercept −0.153, slope 0.960, E_max 0.042) and provided superior net benefit at clinical decision‐making thresholds between 1% and 7% for interventions preventing readmission. An online calculator is provided here.Conclusions: Postpartum readmission for hypertension and pre‐eclampsia may be accurately predicted but further model validation is needed. Model updating using data from multiple sites will be needed before use across clinical settings.

Purpose: To develop a tool to predict a woman's treatment pattern for bothersome urinary urgency (UU) and/or UU incontinence over 1 year after presenting for care at urology or urogynecology clinics.Methods: The Symptoms of Lower Urinary Tract Dysfunction Research Network observational cohort study enrolled adult women with bothersome UU and/or UU incontinence using the lower urinary tract symptoms (LUTS) Tool who were seeking care for LUTS. Treatments for UU and/or urgency incontinence were ordered from least to most invasive. Ordinal logistic and Cox proportional hazard regression models were fit to predict the most invasive level of treatment during follow‐up and overactive bladder (OAB) medication discontinuation, respectively. Binary logistic regression was performed to predict sling treatment during the study follow‐up. Clinical tools were then created using the models listed above to predict treatment pattern over 12 months.Results: Among 349 women, 281 reported UU incontinence, and 68 reported UU at baseline. The highest level of treatment during the study was as follows: 20% no treatment, 24% behavioral treatments, 23% physical therapy, 26% OAB medication, 1% percutaneous tibial nerve stimulation, 3% onabotulinumtoxin A, and 3% sacral neuromodulation. Slings were placed in 10% (n = 36) of participants before baseline and in 11% (n = 40) during study follow‐up. Baseline factors associated with predicting the most invasive level of treatment included baseline level of treatment, hypertension, UU incontinence severity, stress urinary incontinence (SUI) severity, and anticholinergic burden score. Less severe baseline depression and less severe UU incontinence were associated with OAB medication discontinuation. UU and SUI severity were associated with sling placement during the study period. Three tools are available to predict: (1) highest level of treatment; (2) OAB medication discontinuation; and (3) sling placement.Conclusions: OAB treatment prediction tools developed in this study can help providers individualize treatment plans and identify not only patients at risk for treatment discontinuation but also patients who may not be escalated to potentially beneficial OAB treatments, with the goal to improve clinical outcomes for patients suffering from this chronic and often debilitating condition.

Background: Limited epidemiological data exist describing how patients engage with various treatments for overactive bladder (OAB). To improve care for patients with OAB, it is essential to gain a better understanding of how patients interface with OAB treatments longitudinally, that is, how often patients change treatments and the pattern of this treatment change in terms of escalation and de‐escalation.Objectives: To describe treatment patterns for women with bothersome urinary urgency (UU) and/or urgency urinary incontinence (UUI) presenting to specialty care over 1 year.Study Design: The Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) study enrolled adult women with bothersome UU and/or UUI seeking care for lower urinary tract symptoms (LUTS) between January 2015 and September 2016. An ordinal logistic regression model was fitted to describe the probabilities of escalating or de‐escalating level of treatment during 1‐year follow‐up.Results: Among 349 women, 281 reported UUI and 68 reported UU at baseline. At the end of 1 year of treatment by a urologist or urogynecologist, the highest level of treatment received by participants was 5% expectant management, 36% behavioral treatments (BT), 26% physical therapy (PT), 26% OAB medications, 1% percutaneous tibial nerve stimulation, 3% intradetrusor onabotulinum toxin A injection, and 3% sacral neuromodulation. Participants using BT or PT at baseline were more likely to be de‐escalated to no treatment than participants on OAB medications at baseline, who tended to stay on medications. Predictors of the highest level of treatment included starting level of treatment, hypertension, UUI severity, stress urinary incontinence, and anticholinergic burden score.Conclusions: Treatment patterns for UU and UUI are diverse. Even for patients with significant bother from OAB presenting to specialty clinics, further treatment often only involves conservative or medical therapies. This study highlights the need for improved treatment algorithms to escalate patients with persistent symptoms, or to adjust care in those who have been unsuccessfully treated.

Objective: To develop and validate a model to predict obstetric anal sphincter injuries (OASIS) using only information available at the time of admission for labour.Design: A clinical predictive model using a retrospective cohort.Setting: A US health system containing one community and one tertiary hospital.Sample: A total of 22 873 pregnancy episodes with in‐hospital delivery at or beyond 21 weeks of gestation.Methods: Thirty antepartum risk factors were identified as candidate variables, and a prediction model was built using logistic regression predicting OASIS versus no OASIS. Models were fit using the overall study population and separately using hospital‐specific cohorts. Bootstrapping was used for internal validation and external cross‐validation was performed between the two hospital cohorts.Main outcome measures: Model performance was estimated using the bias‐corrected concordance index (c‐index), calibration plots and decision curves.Results: Fifteen risk factors were retained in the final model. Decreasing parity, previous caesarean birth and cardiovascular disease increased risk of OASIS, whereas tobacco use and black race decreased risk. The final model from the total study population had good discrimination (c‐index 0.77, 95% confidence interval [CI] 0.75–0.78) and was able to accurately predict risks between 0 and 35%, where average risk for OASIS was 3%. The site‐specific model fit using patients only from the tertiary hospital had c‐stat 0.74 (95% CI 0.72–0.77) on community hospital patients, and the community hospital model was 0.77 (95%CI 0.76–0.80) on the tertiary hospital patients.Conclusions: OASIS can be accurately predicted based on variables known at the time of admission for labour. These predictions could be useful for selectively implementing OASIS prevention strategies.

Artificial intelligence/machine learning models are being rapidly developed and used in clinical practice. However, many models are deployed without a clear understanding of clinical or operational impact and frequently lack monitoring plans that can detect potential safety signals. There is a lack of consensus in establishing governance to deploy, pilot, and monitor algorithms within operational healthcare delivery workflows. Here, we describe a governance framework that combines current regulatory best practices and lifecycle management of predictive models being used for clinical care. Since January 2021, we have successfully added models to our governance portfolio and are currently managing 52 models.