Mariel S. Lavieri

Improving Screening for Hepatocellular Carcinoma by Incorporating Data on Levels of α-Fetoprotein, Over Time

BACKGROUND & AIMS Current screening algorithms for hepatocellular carcinoma (HCC) view each testing interval independently, without considering prior test results. We investigated whether measurements of α-fetoprotein (AFP), over time, can be used to identify patients most likely to develop HCC. METHODS We performed a nested case-control study using data from subjects in the Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis trial; 82 patients with HCC were matched 1:3 to individuals without HCC (controls), using bootstrap methods to ensure similar follow-up times between groups. We assessed the independent association between development of HCC and the following: (1) most recent level of AFP, (2) standard deviation in level of AFP, and (3) rate of increase in AFP using a multiple logistic regression that included patient-specific risk factors such as age, platelet count, and smoking status. RESULTS In bivariable analysis, all 3 AFP metrics were associated with HCC development; the most strongly associated was the standard deviation of AFP (odds ratio, 1.03 per unit increase in standard deviation; P < .001). Incorporating the standard deviation of AFP and rate of AFP increase, along with patient-specific risk factors, improved the prognostic accuracy to an area under the receiver-operating characteristic curve of 0.81, compared with 0.76 when only the most recent AFP level was used. CONCLUSIONS Patterns of AFP test results can more accurately identify patients with hepatitis C and advanced fibrosis or cirrhosis most likely to develop HCC, compared with most recent AFP test results.

Sharpening the focus on causes and timing of readmission after radical cystectomy for bladder cancer

Readmissions after radical cystectomy are common, burdensome, and poorly understood. For these reasons, the authors conducted a population‐based study that focused on the causes of and time to readmission after radical cystectomy.

Understanding hospital readmission intensity after radical cystectomy.

PURPOSE Hospital readmissions after radical cystectomy vary with respect to intensity in terms of impact on patients and health care systems. Therefore, we conducted a population based study to examine factors associated with increasing readmission intensity after radical cystectomy for bladder cancer. MATERIALS AND METHODS Using SEER (Surveillance, Epidemiology, and End Results)-Medicare data we identified 1,782 patients who underwent radical cystectomy from 2003 to 2009. We defined readmission intensity in terms of length of stay (days) divided into quartiles of less than 3 (lowest), 3 to 4, 5 to 7 and more than 7 (highest). We used logistic regression to examine factors associated with readmission intensity. RESULTS More than half of the patients with the highest intensity readmissions were readmitted within the first week and 77% were readmitted within 2 weeks of discharge. Patients with the highest intensity readmissions were similar in age, gender, race, socioeconomic status, pathological stage, comorbidity, neoadjuvant chemotherapy use and urinary diversion type compared to patients with the lowest intensity readmissions. After multivariable adjustment, complications during the index cystectomy admission (p <0.001), readmission week (p=0.04), and the interaction between index length of stay and discharge to a skilled nursing facility (p=0.04) were associated with the highest readmission intensity. CONCLUSIONS Readmission intensity differs widely after discharge following radical cystectomy. As postoperative efforts to minimize the readmission burden increase, a better understanding of the factors that contribute to the highest intensity readmissions will help direct limited resources (eg telephone calls, office visits) toward high yield areas.

When to treat prostate cancer patients based on their PSA dynamics

This paper provides an innovative approach to help clinicians decide when to start radiation therapy in prostate cancer patients. The decision is based on predictions of the time when the patients prostate specific antigen (PSA) level reaches its lowest point (nadir). These predictions are based on a log quadratic model for how the PSA level changes over time. The distribution of the time of the PSA nadir (which might be linked to maximal tumor regression) is derived from an approximation to the ratio of two correlated normal random variables. Using a dynamic Kalman filter model, the parameter estimates are updated as new patient information becomes available. Clustering is incorporated to improve prior estimates of the curve parameters. The model balances the risk of beginning radiation therapy too soon so that hormone therapy has not achieved its maximum effect vs. waiting too long to start therapy so that there is an increased risk of tumor cells becoming resistant to the treatment. A comparison of clinically implementable policies (cumulative probability policy and threshold probability policy) based on this new approach is applied to a cohort of prostate cancer patients. It shows that our approach outperforms the current protocol.

Dynamic Forecasting and Control Algorithms of Glaucoma Progression for Clinician Decision Support

In managing chronic diseases such as glaucoma, the timing of periodic examinations is crucial, as it may significantly impact patients’ outcomes. We address the question of when to monitor a glaucoma patient by integrating a dynamic, stochastic state space system model of disease evolution with novel optimization approaches to predict the likelihood of progression at any future time. Information about each patient’s disease state is learned sequentially through a series of noisy medical tests. This information is used to determine the best time to next test based on each patient’s individual disease trajectory as well as population information. We develop closed-form solutions and study structural properties of our algorithm. While some have proposed that fixed-interval monitoring can be improved upon, our methodology validates a sophisticated model-based approach to doing so. Based on data from two large-scale, 10+ years clinical trials, we show that our methods significantly outperform fixed-interval sc...

Projections in donor organs available for liver transplantation in the United States: 2014-2025.

With the aging US population, demographic shifts, and obesity epidemic, there is potential for further exacerbation of the current liver donor shortage. We aimed to project the availability of liver grafts in the United States. We performed a secondary analysis of the Organ Procurement and Transplantation Network database of all adult donors from 2000 to 2012 and calculated the total number of donors available and transplanted donor livers stratified by age, race, and body mass index (BMI) group per year. We used National Health and Nutrition Examination Survey and Centers for Disease Control and Prevention historical data to stratify the general population by age, sex, race, and BMI. We then used US population age and race projections provided by the US Census Bureau and the Weldon Cooper Center for Public Service and made national and regional projections of available donors and donor liver utilization from 2014 to 2025. We performed sensitivity analyses and varied the rate of the rise in obesity, proportion of Hispanics, population growth, liver utilization rate, and donation after cardiac death (DCD) utilization. The projected adult population growth in the United States from 2014 to 2025 will be 7.1%. However, we project that there will be a 6.1% increase in the number of used liver grafts. There is marked regional heterogeneity in liver donor growth. Projections were significantly affected by changes in BMI, DCD utilization, and liver utilization rates but not by changes in the Hispanic proportion of the US population or changes in the overall population growth. Overall population growth will outpace the growth of available donor organs and thus potentially exacerbate the existing liver graft shortage. The projected growth in organs is highly heterogeneous across different United Network for Organ Sharing regions. Focused strategies to increase the liver donor pool are warranted. Liver Transpl 21:855‐863, 2015.

Predicting Emergency Department Volume Using Forecasting Methods to Create a ''Surge Response'' for Noncrisis Events

OBJECTIVES This study investigated whether emergency department (ED) variables could be used in mathematical models to predict a future surge in ED volume based on recent levels of use of physician capacity. The models may be used to guide decisions related to on-call staffing in non-crisis-related surges of patient volume. METHODS A retrospective analysis was conducted using information spanning July 2009 through June 2010 from a large urban teaching hospital with a Level I trauma center. A comparison of significance was used to assess the impact of multiple patient-specific variables on the state of the ED. Physician capacity was modeled based on historical physician treatment capacity and productivity. Binary logistic regression analysis was used to determine the probability that the available physician capacity would be sufficient to treat all patients forecasted to arrive in the next time period. The prediction horizons used were 15 minutes, 30 minutes, 1 hour, 2 hours, 4 hours, 8 hours, and 12 hours. Five consecutive months of patient data from July 2010 through November 2010, similar to the data used to generate the models, was used to validate the models. Positive predictive values, Type I and Type II errors, and real-time accuracy in predicting noncrisis surge events were used to evaluate the forecast accuracy of the models. RESULTS The ratio of new patients requiring treatment over total physician capacity (termed the care utilization ratio [CUR]) was deemed a robust predictor of the state of the ED (with a CUR greater than 1 indicating that the physician capacity would not be sufficient to treat all patients forecasted to arrive). Prediction intervals of 30 minutes, 8 hours, and 12 hours performed best of all models analyzed, with deviances of 1.000, 0.951, and 0.864, respectively. A 95% significance was used to validate the models against the July 2010 through November 2010 data set. Positive predictive values ranged from 0.738 to 0.872, true positives ranged from 74% to 94%, and true negatives ranged from 70% to 90% depending on the threshold used to determine the state of the ED with the 30-minute prediction model. CONCLUSIONS The CUR is a new and robust indicator of an ED systems performance. The study was able to model the tradeoff of longer time to response versus shorter but more accurate predictions, by investigating different prediction intervals. Current practice would have been improved by using the proposed models and would have identified the surge in patient volume earlier on noncrisis days.

Projected increase in obesity and non-alcoholic steatohepatitis-related liver transplantation waitlist additions in the United States

Nonalcoholic steatohepatitis (NASH) cirrhosis is the fastest growing indication for liver transplantation (LT) in the United States. We aimed to determine the temporal trend behind the rise in obesity and NASH‐related additions to the LT waitlist in the United States and make projections for future NASH burden on the LT waitlist. We used data from the Organ Procurement and Transplantation Network database from 2000 to 2014 to obtain the number of NASH‐related LT waitlist additions. The obese population in the United States from 2000 to 2014 was estimated using data from the U.S. Census Bureau and the National Health and Nutrition Examination Survey. Based on obesity trends, we established a time lag between obesity prevalence and NASH‐related waitlist additions. We used data from the U.S. Census Bureau on population projections from 2016 to 2030 to forecast obesity estimates and NASH‐related LT waitlist additions. From 2000 to 2014, the proportion of obese individuals significantly increased 44.9% and the number of NASH‐related annual waitlist additions increased from 391 to 1,605. Increase in obesity prevalence was strongly associated with LT waitlist additions 9 years later in derivation and validation cohorts (R2 = 0.9). Based on these data, annual NASH‐related waitlist additions are anticipated to increase by 55.4% (1,354‐2,104) between 2016 and 2030. There is significant regional variation in obesity rates and in the anticipated increase in NASH‐related waitlist additions (P < 0.01). Conclusion: We project a marked increase in demand for LT for NASH given population obesity trends. Continued public health efforts to curb obesity prevalence are needed to reduce the projected future burden of NASH. (Hepatology 2017).

A model to optimize followup care and reduce hospital readmissions after radical cystectomy

PURPOSE Radical cystectomy has one of the highest readmission rates across all surgical procedures at approximately 25%. We developed a mathematical model to optimize outpatient followup regimens for radical cystectomy. MATERIALS AND METHODS We used delay-time analysis, a systems engineering approach, to maximize the probability of detecting patients susceptible to readmission through office visits and telephone calls. Our data source includes patients readmitted after radical cystectomy from the Healthcare Cost and Utilization Project State Inpatient Databases in 2009 and 2010 as well as from our institutional bladder cancer database from 2007 to 2011. We measured the interval from hospital discharge to the point when a patient first exhibits concerning symptoms. Our primary end point is 30-day hospital readmission. Our model optimized the timing and sequence of followup care after radical cystectomy. RESULTS The timing of office visits and telephone calls is more important in detecting a patient at risk for readmission than the sequence of these encounters. Patients are most likely to exhibit concerning symptoms between 4 and 5 days after discharge home. An optimally scheduled office visit can detect up to 16% of potential readmissions, which can be increased to 36% with 1 office visit followed by 4 telephone calls. CONCLUSIONS Our model improves the detection of concerning symptoms after radical cystectomy by optimizing the timing and number of outpatient encounters. By understanding how to design better outpatient followup care for patients treated with radical cystectomy we can help reduce the readmission burden for this population.

Maximizing the Value of Mobile Health Monitoring by Avoiding Redundant Patient Reports: Prediction of Depression-Related Symptoms and Adherence Problems in Automated Health Assessment Services

Background Interactive voice response (IVR) calls enhance health systems’ ability to identify health risk factors, thereby enabling targeted clinical follow-up. However, redundant assessments may increase patient dropout and represent a lost opportunity to collect more clinically useful data. Objective We determined the extent to which previous IVR assessments predicted subsequent responses among patients with depression diagnoses, potentially obviating the need to repeatedly collect the same information. We also evaluated whether frequent (ie, weekly) IVR assessment attempts were significantly more predictive of patients’ subsequent reports than information collected biweekly or monthly. Methods Using data from 1050 IVR assessments for 208 patients with depression diagnoses, we examined the predictability of four IVR-reported outcomes: moderate/severe depressive symptoms (score ≥10 on the PHQ-9), fair/poor general health, poor antidepressant adherence, and days in bed due to poor mental health. We used logistic models with training and test samples to predict patients’ IVR responses based on their five most recent weekly, biweekly, and monthly assessment attempts. The marginal benefit of more frequent assessments was evaluated based on Receiver Operator Characteristic (ROC) curves and statistical comparisons of the area under the curves (AUC). Results Patients’ reports about their depressive symptoms and perceived health status were highly predictable based on prior assessment responses. For models predicting moderate/severe depression, the AUC was 0.91 (95% CI 0.89-0.93) when assuming weekly assessment attempts and only slightly less when assuming biweekly assessments (AUC: 0.89; CI 0.87-0.91) or monthly attempts (AUC: 0.89; CI 0.86-0.91). The AUC for models predicting reports of fair/poor health status was similar when weekly assessments were compared with those occurring biweekly (P value for the difference=.11) or monthly (P=.81). Reports of medication adherence problems and days in bed were somewhat less predictable but also showed small differences between assessments attempted weekly, biweekly, and monthly. Conclusions The technical feasibility of gathering high frequency health data via IVR may in some instances exceed the clinical benefit of doing so. Predictive analytics could make data gathering more efficient with negligible loss in effectiveness. In particular, weekly or biweekly depressive symptom reports may provide little marginal information regarding how the person is doing relative to collecting that information monthly. The next generation of automated health assessment services should use data mining techniques to avoid redundant assessments and should gather data at the frequency that maximizes the value of the information collected.